summary refs log tree commit diff
path: root/sysdeps/x86_64
diff options
context:
space:
mode:
Diffstat (limited to 'sysdeps/x86_64')
-rw-r--r--sysdeps/x86_64/fpu/Makeconfig1
-rw-r--r--sysdeps/x86_64/fpu/Versions2
-rw-r--r--sysdeps/x86_64/fpu/libm-test-ulps20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core-sse2.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core.c27
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core_sse4.S418
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core-sse.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core.c27
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core_avx2.S429
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core-avx2.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core.c27
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core_avx512.S287
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core-avx2.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core.c28
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core_avx512.S269
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core-sse2.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core.c28
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core_sse4.S311
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core-sse.S20
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core.c28
-rw-r--r--sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core_avx2.S331
-rw-r--r--sysdeps/x86_64/fpu/svml_d_exp102_core.S29
-rw-r--r--sysdeps/x86_64/fpu/svml_d_exp104_core.S29
-rw-r--r--sysdeps/x86_64/fpu/svml_d_exp104_core_avx.S25
-rw-r--r--sysdeps/x86_64/fpu/svml_d_exp108_core.S25
-rw-r--r--sysdeps/x86_64/fpu/svml_s_exp10f16_core.S25
-rw-r--r--sysdeps/x86_64/fpu/svml_s_exp10f4_core.S29
-rw-r--r--sysdeps/x86_64/fpu/svml_s_exp10f8_core.S29
-rw-r--r--sysdeps/x86_64/fpu/svml_s_exp10f8_core_avx.S25
-rw-r--r--sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx2.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx512f.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-libmvec-exp10.c3
-rw-r--r--sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx2.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx512f.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-libmvec-exp10f.c3
-rw-r--r--sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c1
-rw-r--r--sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c1
45 files changed, 2589 insertions, 0 deletions
diff --git a/sysdeps/x86_64/fpu/Makeconfig b/sysdeps/x86_64/fpu/Makeconfig
index 03b2364417..fd0a9da439 100644
--- a/sysdeps/x86_64/fpu/Makeconfig
+++ b/sysdeps/x86_64/fpu/Makeconfig
@@ -27,6 +27,7 @@ libmvec-funcs = \
   atan \
   cos \
   exp \
+  exp10 \
   exp2 \
   hypot \
   log \
diff --git a/sysdeps/x86_64/fpu/Versions b/sysdeps/x86_64/fpu/Versions
index 12b7ad1830..f29cfa4cbf 100644
--- a/sysdeps/x86_64/fpu/Versions
+++ b/sysdeps/x86_64/fpu/Versions
@@ -17,11 +17,13 @@ libmvec {
     _ZGVbN2v_acos; _ZGVcN4v_acos; _ZGVdN4v_acos; _ZGVeN8v_acos;
     _ZGVbN2v_asin; _ZGVcN4v_asin; _ZGVdN4v_asin; _ZGVeN8v_asin;
     _ZGVbN2v_atan; _ZGVcN4v_atan; _ZGVdN4v_atan; _ZGVeN8v_atan;
+    _ZGVbN2v_exp10; _ZGVcN4v_exp10; _ZGVdN4v_exp10; _ZGVeN8v_exp10;
     _ZGVbN2v_exp2; _ZGVcN4v_exp2; _ZGVdN4v_exp2; _ZGVeN8v_exp2;
     _ZGVbN2vv_hypot; _ZGVcN4vv_hypot; _ZGVdN4vv_hypot; _ZGVeN8vv_hypot;
     _ZGVbN4v_acosf; _ZGVcN8v_acosf; _ZGVdN8v_acosf; _ZGVeN16v_acosf;
     _ZGVbN4v_asinf; _ZGVcN8v_asinf; _ZGVdN8v_asinf; _ZGVeN16v_asinf;
     _ZGVbN4v_atanf; _ZGVcN8v_atanf; _ZGVdN8v_atanf; _ZGVeN16v_atanf;
+    _ZGVbN4v_exp10f; _ZGVcN8v_exp10f; _ZGVdN8v_exp10f; _ZGVeN16v_exp10f;
     _ZGVbN4v_exp2f; _ZGVcN8v_exp2f; _ZGVdN8v_exp2f; _ZGVeN16v_exp2f;
     _ZGVbN4vv_hypotf; _ZGVcN8vv_hypotf; _ZGVdN8vv_hypotf; _ZGVeN16vv_hypotf;
   }
diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps
index bc4479ad39..45f2e4bb53 100644
--- a/sysdeps/x86_64/fpu/libm-test-ulps
+++ b/sysdeps/x86_64/fpu/libm-test-ulps
@@ -1252,6 +1252,26 @@ float: 1
 float128: 3
 ldouble: 2
 
+Function: "exp10_vlen16":
+float: 3
+
+Function: "exp10_vlen2":
+double: 1
+
+Function: "exp10_vlen4":
+double: 1
+float: 1
+
+Function: "exp10_vlen4_avx2":
+double: 1
+
+Function: "exp10_vlen8":
+double: 1
+float: 1
+
+Function: "exp10_vlen8_avx2":
+float: 1
+
 Function: "exp2":
 double: 1
 float: 1
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core-sse2.S
new file mode 100644
index 0000000000..ab615c0323
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core-sse2.S
@@ -0,0 +1,20 @@
+/* SSE2 version of vectorized exp10, vector length is 2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVbN2v_exp10 _ZGVbN2v_exp10_sse2
+#include "../svml_d_exp102_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core.c
new file mode 100644
index 0000000000..5c5625b278
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core.c
@@ -0,0 +1,27 @@
+/* Multiple versions of vectorized exp10, vector length is 2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVbN2v_exp10
+#include "ifunc-mathvec-sse4_1.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVbN2v_exp10, __GI__ZGVbN2v_exp10, __redirect__ZGVbN2v_exp10)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core_sse4.S
new file mode 100644
index 0000000000..7c6e5de3e0
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp102_core_sse4.S
@@ -0,0 +1,418 @@
+/* Function exp10 vectorized with SSE4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *
+ *   exp10(x)  = 2^x/log10(2) = 2^n * (1 + T[j]) * (1 + P(y))
+ *   where
+ *        x = m*log10(2)/K + y,  y in [-log10(2)/K..log10(2)/K]
+ *        m = n*K + j,           m,n,j - signed integer, j in [-K/2..K/2]
+ *
+ *        values of 2^j/K are tabulated
+ *
+ *        P(y) is a minimax polynomial approximation of exp10(x)-1
+ *        on small interval [-log10(2)/K..log10(2)/K]
+ *
+ *  Special cases:
+ *
+ *   exp10(NaN)  = NaN
+ *   exp10(+INF) = +INF
+ *   exp10(-INF) = 0
+ *   exp10(x)    = 1 for subnormals
+ *   For IEEE double
+ *     if x >  3.39782712893383973096e+02 then exp10(x) overflow
+ *     if x < -3.45133219101941108420e+02 then exp10(x) underflow
+ *
+ */
+
+/* Offsets for data table __svml_dexp10_data_internal
+ */
+#define _dbT                          	0
+#define _dbLg2_10                     	1024
+#define _dbShifter                    	1040
+#define _dbInvLg2_10hi                	1056
+#define _dbInvLg2_10lo                	1072
+#define _dPC1                         	1088
+#define _dPC2                         	1104
+#define _dPC3                         	1120
+#define _dPC4                         	1136
+#define _dPC5                         	1152
+#define _lExpMask                     	1168
+#define _iIndexMask                   	1184
+#define _iAbsMask                     	1200
+#define _iDomainRange                 	1216
+
+#include <sysdep.h>
+
+        .text
+	.section .text.sse4,"ax",@progbits
+ENTRY(_ZGVbN2v_exp10_sse4)
+        subq      $72, %rsp
+        cfi_def_cfa_offset(80)
+
+/*  R  */
+        movaps    %xmm0, %xmm12
+
+/*  Load arument  */
+        movups    _dbLg2_10+__svml_dexp10_data_internal(%rip), %xmm13
+        lea       __svml_dexp10_data_internal(%rip), %rsi
+        mulpd     %xmm0, %xmm13
+        movups    _dbShifter+__svml_dexp10_data_internal(%rip), %xmm1
+        addpd     %xmm1, %xmm13
+        movaps    %xmm13, %xmm9
+        subpd     %xmm1, %xmm9
+        movups    _dbInvLg2_10hi+__svml_dexp10_data_internal(%rip), %xmm8
+        mulpd     %xmm9, %xmm8
+        movups    _dbInvLg2_10lo+__svml_dexp10_data_internal(%rip), %xmm10
+        mulpd     %xmm9, %xmm10
+        subpd     %xmm8, %xmm12
+        subpd     %xmm10, %xmm12
+
+/*
+ *  Polynomial
+ * poly(dN) = a1*dR+...+a5*dR^5
+ */
+        movups    _dPC5+__svml_dexp10_data_internal(%rip), %xmm11
+        mulpd     %xmm12, %xmm11
+        addpd     _dPC4+__svml_dexp10_data_internal(%rip), %xmm11
+        mulpd     %xmm12, %xmm11
+        addpd     _dPC3+__svml_dexp10_data_internal(%rip), %xmm11
+        mulpd     %xmm12, %xmm11
+        addpd     _dPC2+__svml_dexp10_data_internal(%rip), %xmm11
+
+/* a1+...+a5*dR^4 ! */
+        mulpd     %xmm12, %xmm11
+        addpd     _dPC1+__svml_dexp10_data_internal(%rip), %xmm11
+        movq      _iIndexMask+__svml_dexp10_data_internal(%rip), %xmm5
+
+/*  Index and lookup  */
+        pshufd    $136, %xmm13, %xmm6
+
+/*  2^N  */
+        psllq     $45, %xmm13
+        pand      %xmm5, %xmm6
+
+/* iIndex*=sizeof(D); */
+        pslld     $3, %xmm6
+        movd      %xmm6, %eax
+        pshufd    $1, %xmm6, %xmm7
+        movq      _iAbsMask+__svml_dexp10_data_internal(%rip), %xmm2
+
+/* a1*dR+...+a5*dR^5 */
+        mulpd     %xmm11, %xmm12
+        movd      %xmm7, %ecx
+
+/* Check for overflow\underflow  */
+        pshufd    $221, %xmm0, %xmm4
+        movq      _iDomainRange+__svml_dexp10_data_internal(%rip), %xmm3
+        pand      %xmm2, %xmm4
+        movslq    %eax, %rax
+        pcmpgtd   %xmm3, %xmm4
+        movslq    %ecx, %rcx
+        movmskps  %xmm4, %edx
+
+/* lM==EXP(2^N) */
+        pand      _lExpMask+__svml_dexp10_data_internal(%rip), %xmm13
+        movsd     (%rsi,%rax), %xmm1
+        movhpd    (%rsi,%rcx), %xmm1
+
+/* Tj*poly */
+        mulpd     %xmm1, %xmm12
+        addpd     %xmm12, %xmm1
+
+/* quick 2^N */
+        paddq     %xmm13, %xmm1
+        andl      $3, %edx
+
+/*  Finish   */
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx rbp r12 r13 r14 r15 edx xmm0 xmm1
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        movaps    %xmm1, %xmm0
+        addq      $72, %rsp
+        cfi_def_cfa_offset(8)
+        ret
+        cfi_def_cfa_offset(80)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        movups    %xmm0, 32(%rsp)
+        movups    %xmm1, 48(%rsp)
+                                # LOE rbx rbp r12 r13 r14 r15 edx xmm1
+
+        xorl      %eax, %eax
+        movq      %r12, 16(%rsp)
+        cfi_offset(12, -64)
+        movl      %eax, %r12d
+        movq      %r13, 8(%rsp)
+        cfi_offset(13, -72)
+        movl      %edx, %r13d
+        movq      %r14, (%rsp)
+        cfi_offset(14, -80)
+                                # LOE rbx rbp r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx rbp r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $2, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx rbp r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        movups    48(%rsp), %xmm1
+
+/* Go to exit */
+        jmp       L(EXIT)
+        cfi_offset(12, -64)
+        cfi_offset(13, -72)
+        cfi_offset(14, -80)
+                                # LOE rbx rbp r12 r13 r14 r15 xmm1
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movsd     32(%rsp,%r14,8), %xmm0
+        call      exp10@PLT
+                                # LOE rbx rbp r14 r15 r12d r13d xmm0
+
+        movsd     %xmm0, 48(%rsp,%r14,8)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx rbp r15 r12d r13d
+END(_ZGVbN2v_exp10_sse4)
+
+        .section .rodata, "a"
+        .align 16
+
+#ifdef __svml_dexp10_data_internal_typedef
+typedef unsigned int VUINT32;
+typedef struct
+{
+        __declspec(align(16)) VUINT32 _dbT[(1<<7)][2];
+        __declspec(align(16)) VUINT32 _dbLg2_10[2][2];
+        __declspec(align(16)) VUINT32 _dbShifter[2][2];
+        __declspec(align(16)) VUINT32 _dbInvLg2_10hi[2][2];
+        __declspec(align(16)) VUINT32 _dbInvLg2_10lo[2][2];
+        __declspec(align(16)) VUINT32 _dPC1[2][2];
+        __declspec(align(16)) VUINT32 _dPC2[2][2];
+        __declspec(align(16)) VUINT32 _dPC3[2][2];
+        __declspec(align(16)) VUINT32 _dPC4[2][2];
+        __declspec(align(16)) VUINT32 _dPC5[2][2];
+        __declspec(align(16)) VUINT32 _lExpMask[2][2];
+        __declspec(align(16)) VUINT32 _iIndexMask[4][1];
+        __declspec(align(16)) VUINT32 _iAbsMask[4][1];
+        __declspec(align(16)) VUINT32 _iDomainRange[4][1];
+} __svml_dexp10_data_internal;
+#endif
+__svml_dexp10_data_internal:
+        /*== _dbT ==*/
+        .quad 0x3ff0000000000000    /*2^( 0 /128)*/
+        .quad 0x3ff0163da9fb3335    /*2^( 1 /128)*/
+        .quad 0x3ff02c9a3e778061    /*2^( 2 /128)*/
+        .quad 0x3ff04315e86e7f85    /*2^( 3 /128)*/
+        .quad 0x3ff059b0d3158574    /*2^( 4 /128)*/
+        .quad 0x3ff0706b29ddf6de    /*2^( 5 /128)*/
+        .quad 0x3ff0874518759bc8    /*2^( 6 /128)*/
+        .quad 0x3ff09e3ecac6f383    /*2^( 7 /128)*/
+        .quad 0x3ff0b5586cf9890f    /*2^( 8 /128)*/
+        .quad 0x3ff0cc922b7247f7    /*2^( 9 /128)*/
+        .quad 0x3ff0e3ec32d3d1a2    /*2^( 10 /128)*/
+        .quad 0x3ff0fb66affed31b    /*2^( 11 /128)*/
+        .quad 0x3ff11301d0125b51    /*2^( 12 /128)*/
+        .quad 0x3ff12abdc06c31cc    /*2^( 13 /128)*/
+        .quad 0x3ff1429aaea92de0    /*2^( 14 /128)*/
+        .quad 0x3ff15a98c8a58e51    /*2^( 15 /128)*/
+        .quad 0x3ff172b83c7d517b    /*2^( 16 /128)*/
+        .quad 0x3ff18af9388c8dea    /*2^( 17 /128)*/
+        .quad 0x3ff1a35beb6fcb75    /*2^( 18 /128)*/
+        .quad 0x3ff1bbe084045cd4    /*2^( 19 /128)*/
+        .quad 0x3ff1d4873168b9aa    /*2^( 20 /128)*/
+        .quad 0x3ff1ed5022fcd91d    /*2^( 21 /128)*/
+        .quad 0x3ff2063b88628cd6    /*2^( 22 /128)*/
+        .quad 0x3ff21f49917ddc96    /*2^( 23 /128)*/
+        .quad 0x3ff2387a6e756238    /*2^( 24 /128)*/
+        .quad 0x3ff251ce4fb2a63f    /*2^( 25 /128)*/
+        .quad 0x3ff26b4565e27cdd    /*2^( 26 /128)*/
+        .quad 0x3ff284dfe1f56381    /*2^( 27 /128)*/
+        .quad 0x3ff29e9df51fdee1    /*2^( 28 /128)*/
+        .quad 0x3ff2b87fd0dad990    /*2^( 29 /128)*/
+        .quad 0x3ff2d285a6e4030b    /*2^( 30 /128)*/
+        .quad 0x3ff2ecafa93e2f56    /*2^( 31 /128)*/
+        .quad 0x3ff306fe0a31b715    /*2^( 32 /128)*/
+        .quad 0x3ff32170fc4cd831    /*2^( 33 /128)*/
+        .quad 0x3ff33c08b26416ff    /*2^( 34 /128)*/
+        .quad 0x3ff356c55f929ff1    /*2^( 35 /128)*/
+        .quad 0x3ff371a7373aa9cb    /*2^( 36 /128)*/
+        .quad 0x3ff38cae6d05d866    /*2^( 37 /128)*/
+        .quad 0x3ff3a7db34e59ff7    /*2^( 38 /128)*/
+        .quad 0x3ff3c32dc313a8e5    /*2^( 39 /128)*/
+        .quad 0x3ff3dea64c123422    /*2^( 40 /128)*/
+        .quad 0x3ff3fa4504ac801c    /*2^( 41 /128)*/
+        .quad 0x3ff4160a21f72e2a    /*2^( 42 /128)*/
+        .quad 0x3ff431f5d950a897    /*2^( 43 /128)*/
+        .quad 0x3ff44e086061892d    /*2^( 44 /128)*/
+        .quad 0x3ff46a41ed1d0057    /*2^( 45 /128)*/
+        .quad 0x3ff486a2b5c13cd0    /*2^( 46 /128)*/
+        .quad 0x3ff4a32af0d7d3de    /*2^( 47 /128)*/
+        .quad 0x3ff4bfdad5362a27    /*2^( 48 /128)*/
+        .quad 0x3ff4dcb299fddd0d    /*2^( 49 /128)*/
+        .quad 0x3ff4f9b2769d2ca7    /*2^( 50 /128)*/
+        .quad 0x3ff516daa2cf6642    /*2^( 51 /128)*/
+        .quad 0x3ff5342b569d4f82    /*2^( 52 /128)*/
+        .quad 0x3ff551a4ca5d920f    /*2^( 53 /128)*/
+        .quad 0x3ff56f4736b527da    /*2^( 54 /128)*/
+        .quad 0x3ff58d12d497c7fd    /*2^( 55 /128)*/
+        .quad 0x3ff5ab07dd485429    /*2^( 56 /128)*/
+        .quad 0x3ff5c9268a5946b7    /*2^( 57 /128)*/
+        .quad 0x3ff5e76f15ad2148    /*2^( 58 /128)*/
+        .quad 0x3ff605e1b976dc09    /*2^( 59 /128)*/
+        .quad 0x3ff6247eb03a5585    /*2^( 60 /128)*/
+        .quad 0x3ff6434634ccc320    /*2^( 61 /128)*/
+        .quad 0x3ff6623882552225    /*2^( 62 /128)*/
+        .quad 0x3ff68155d44ca973    /*2^( 63 /128)*/
+        .quad 0x3ff6a09e667f3bcd    /*2^( 64 /128)*/
+        .quad 0x3ff6c012750bdabf    /*2^( 65 /128)*/
+        .quad 0x3ff6dfb23c651a2f    /*2^( 66 /128)*/
+        .quad 0x3ff6ff7df9519484    /*2^( 67 /128)*/
+        .quad 0x3ff71f75e8ec5f74    /*2^( 68 /128)*/
+        .quad 0x3ff73f9a48a58174    /*2^( 69 /128)*/
+        .quad 0x3ff75feb564267c9    /*2^( 70 /128)*/
+        .quad 0x3ff780694fde5d3f    /*2^( 71 /128)*/
+        .quad 0x3ff7a11473eb0187    /*2^( 72 /128)*/
+        .quad 0x3ff7c1ed0130c132    /*2^( 73 /128)*/
+        .quad 0x3ff7e2f336cf4e62    /*2^( 74 /128)*/
+        .quad 0x3ff80427543e1a12    /*2^( 75 /128)*/
+        .quad 0x3ff82589994cce13    /*2^( 76 /128)*/
+        .quad 0x3ff8471a4623c7ad    /*2^( 77 /128)*/
+        .quad 0x3ff868d99b4492ed    /*2^( 78 /128)*/
+        .quad 0x3ff88ac7d98a6699    /*2^( 79 /128)*/
+        .quad 0x3ff8ace5422aa0db    /*2^( 80 /128)*/
+        .quad 0x3ff8cf3216b5448c    /*2^( 81 /128)*/
+        .quad 0x3ff8f1ae99157736    /*2^( 82 /128)*/
+        .quad 0x3ff9145b0b91ffc6    /*2^( 83 /128)*/
+        .quad 0x3ff93737b0cdc5e5    /*2^( 84 /128)*/
+        .quad 0x3ff95a44cbc8520f    /*2^( 85 /128)*/
+        .quad 0x3ff97d829fde4e50    /*2^( 86 /128)*/
+        .quad 0x3ff9a0f170ca07ba    /*2^( 87 /128)*/
+        .quad 0x3ff9c49182a3f090    /*2^( 88 /128)*/
+        .quad 0x3ff9e86319e32323    /*2^( 89 /128)*/
+        .quad 0x3ffa0c667b5de565    /*2^( 90 /128)*/
+        .quad 0x3ffa309bec4a2d33    /*2^( 91 /128)*/
+        .quad 0x3ffa5503b23e255d    /*2^( 92 /128)*/
+        .quad 0x3ffa799e1330b358    /*2^( 93 /128)*/
+        .quad 0x3ffa9e6b5579fdbf    /*2^( 94 /128)*/
+        .quad 0x3ffac36bbfd3f37a    /*2^( 95 /128)*/
+        .quad 0x3ffae89f995ad3ad    /*2^( 96 /128)*/
+        .quad 0x3ffb0e07298db666    /*2^( 97 /128)*/
+        .quad 0x3ffb33a2b84f15fb    /*2^( 98 /128)*/
+        .quad 0x3ffb59728de5593a    /*2^( 99 /128)*/
+        .quad 0x3ffb7f76f2fb5e47    /*2^( 100 /128)*/
+        .quad 0x3ffba5b030a1064a    /*2^( 101 /128)*/
+        .quad 0x3ffbcc1e904bc1d2    /*2^( 102 /128)*/
+        .quad 0x3ffbf2c25bd71e09    /*2^( 103 /128)*/
+        .quad 0x3ffc199bdd85529c    /*2^( 104 /128)*/
+        .quad 0x3ffc40ab5fffd07a    /*2^( 105 /128)*/
+        .quad 0x3ffc67f12e57d14b    /*2^( 106 /128)*/
+        .quad 0x3ffc8f6d9406e7b5    /*2^( 107 /128)*/
+        .quad 0x3ffcb720dcef9069    /*2^( 108 /128)*/
+        .quad 0x3ffcdf0b555dc3fa    /*2^( 109 /128)*/
+        .quad 0x3ffd072d4a07897c    /*2^( 110 /128)*/
+        .quad 0x3ffd2f87080d89f2    /*2^( 111 /128)*/
+        .quad 0x3ffd5818dcfba487    /*2^( 112 /128)*/
+        .quad 0x3ffd80e316c98398    /*2^( 113 /128)*/
+        .quad 0x3ffda9e603db3285    /*2^( 114 /128)*/
+        .quad 0x3ffdd321f301b460    /*2^( 115 /128)*/
+        .quad 0x3ffdfc97337b9b5f    /*2^( 116 /128)*/
+        .quad 0x3ffe264614f5a129    /*2^( 117 /128)*/
+        .quad 0x3ffe502ee78b3ff6    /*2^( 118 /128)*/
+        .quad 0x3ffe7a51fbc74c83    /*2^( 119 /128)*/
+        .quad 0x3ffea4afa2a490da    /*2^( 120 /128)*/
+        .quad 0x3ffecf482d8e67f1    /*2^( 121 /128)*/
+        .quad 0x3ffefa1bee615a27    /*2^( 122 /128)*/
+        .quad 0x3fff252b376bba97    /*2^( 123 /128)*/
+        .quad 0x3fff50765b6e4540    /*2^( 124 /128)*/
+        .quad 0x3fff7bfdad9cbe14    /*2^( 125 /128)*/
+        .quad 0x3fffa7c1819e90d8    /*2^( 126 /128)*/
+        .quad 0x3fffd3c22b8f71f1     /*2^( 127 /128)*/
+        .align 16
+        .quad 0x407a934f0979a371, 0x407a934f0979a371  /* _dbLg2_10*2^K */
+        .align 16
+        .quad 0x4338800000000000, 0x4338800000000000  /* _dbShifter */
+        .align 16
+        .quad 0x3f63441350a00000, 0x3f63441350a00000  /* _dbInvLg2_10hi/2^K 53-11-K bits*/
+        .align 16
+        .quad 0xbd10c0219dc1da99, 0xbd10c0219dc1da99  /* _dbInvLg2_10lo/2^K */
+        //PC0 = 1.0
+        .align 16
+        .quad 0x40026bb1bbb55516, 0x40026bb1bbb55516  /* _dPC1 */
+        .align 16
+        .quad 0x40053524c73ce8e3, 0x40053524c73ce8e3  /* _dPC2 */
+        .align 16
+        .quad 0x4000470591ccea8b, 0x4000470591ccea8b  /* _dPC3 */
+        .align 16
+        .quad 0x3ff2bd767584db59, 0x3ff2bd767584db59  /* _dPC4 */
+        .align 16
+        .quad 0x3fe144c03efafb54, 0x3fe144c03efafb54  /* _dPC5 */
+        .align 16
+        .quad 0xfff0000000000000, 0xfff0000000000000  /* _lExpMask */
+        .align 16
+        .long 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f          /* _iIndexMask =(2^K-1)*/
+        //common
+        .align 16
+        .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff          /* _iAbsMask */
+        .align 16
+        .long 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70 /* _iDomainRange */
+        .align 16
+        .type	__svml_dexp10_data_internal,@object
+        .size	__svml_dexp10_data_internal,.-__svml_dexp10_data_internal
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core-sse.S
new file mode 100644
index 0000000000..260c052143
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core-sse.S
@@ -0,0 +1,20 @@
+/* SSE version of vectorized exp10, vector length is 4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVdN4v_exp10 _ZGVdN4v_exp10_sse_wrapper
+#include "../svml_d_exp104_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core.c
new file mode 100644
index 0000000000..e3e302be72
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core.c
@@ -0,0 +1,27 @@
+/* Multiple versions of vectorized exp10, vector length is 4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVdN4v_exp10
+#include "ifunc-mathvec-avx2.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVdN4v_exp10, __GI__ZGVdN4v_exp10, __redirect__ZGVdN4v_exp10)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core_avx2.S
new file mode 100644
index 0000000000..1a53f43c9e
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp104_core_avx2.S
@@ -0,0 +1,429 @@
+/* Function exp10 vectorized with AVX2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *
+ *   exp10(x)  = 2^x/log10(2) = 2^n * (1 + T[j]) * (1 + P(y))
+ *   where
+ *        x = m*log10(2)/K + y,  y in [-log10(2)/K..log10(2)/K]
+ *        m = n*K + j,           m,n,j - signed integer, j in [-K/2..K/2]
+ *
+ *        values of 2^j/K are tabulated
+ *
+ *        P(y) is a minimax polynomial approximation of exp10(x)-1
+ *        on small interval [-log10(2)/K..log10(2)/K]
+ *
+ *  Special cases:
+ *
+ *   exp10(NaN)  = NaN
+ *   exp10(+INF) = +INF
+ *   exp10(-INF) = 0
+ *   exp10(x)    = 1 for subnormals
+ *   For IEEE double
+ *     if x >  3.39782712893383973096e+02 then exp10(x) overflow
+ *     if x < -3.45133219101941108420e+02 then exp10(x) underflow
+ *
+ */
+
+/* Offsets for data table __svml_dexp10_data_internal
+ */
+#define _dbT                          	0
+#define _dbLg2_10                     	1024
+#define _dbShifter                    	1056
+#define _dbInvLg2_10hi                	1088
+#define _dbInvLg2_10lo                	1120
+#define _dPC1                         	1152
+#define _dPC2                         	1184
+#define _dPC3                         	1216
+#define _dPC4                         	1248
+#define _dPC5                         	1280
+#define _lExpMask                     	1312
+#define _iIndexMask                   	1344
+#define _iAbsMask                     	1376
+#define _iDomainRange                 	1408
+
+#include <sysdep.h>
+
+        .text
+	.section .text.avx2,"ax",@progbits
+ENTRY(_ZGVdN4v_exp10_avx2)
+        pushq     %rbp
+        cfi_def_cfa_offset(16)
+        movq      %rsp, %rbp
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+        andq      $-32, %rsp
+        subq      $96, %rsp
+        lea       __svml_dexp10_data_internal(%rip), %r8
+        vmovapd   %ymm0, %ymm2
+        vmovupd   _dbShifter+__svml_dexp10_data_internal(%rip), %ymm3
+
+/*  Load arument  */
+        vmovupd   _dbLg2_10+__svml_dexp10_data_internal(%rip), %ymm0
+        vfmadd213pd %ymm3, %ymm2, %ymm0
+        vsubpd    %ymm3, %ymm0, %ymm1
+
+/*  R  */
+        vmovupd   _dbInvLg2_10hi+__svml_dexp10_data_internal(%rip), %ymm3
+        vfnmadd213pd %ymm2, %ymm1, %ymm3
+
+/* Check for overflow\underflow  */
+        vextractf128 $1, %ymm2, %xmm4
+        vfnmadd132pd _dbInvLg2_10lo+__svml_dexp10_data_internal(%rip), %ymm3, %ymm1
+        vshufps   $221, %xmm4, %xmm2, %xmm5
+        vandps    _iAbsMask+__svml_dexp10_data_internal(%rip), %xmm5, %xmm6
+        vpcmpgtd  _iDomainRange+__svml_dexp10_data_internal(%rip), %xmm6, %xmm7
+
+/*
+ *  Polynomial
+ * poly(dN) = a1*dR+...+a5*dR^5
+ */
+        vmovupd   _dPC5+__svml_dexp10_data_internal(%rip), %ymm4
+        vmovmskps %xmm7, %eax
+        vfmadd213pd _dPC4+__svml_dexp10_data_internal(%rip), %ymm1, %ymm4
+        vfmadd213pd _dPC3+__svml_dexp10_data_internal(%rip), %ymm1, %ymm4
+        vfmadd213pd _dPC2+__svml_dexp10_data_internal(%rip), %ymm1, %ymm4
+
+/* a1+...+a5*dR^4 ! */
+        vfmadd213pd _dPC1+__svml_dexp10_data_internal(%rip), %ymm1, %ymm4
+
+/* a1*dR+...+a5*dR^5 */
+        vmulpd    %ymm4, %ymm1, %ymm1
+
+/*  Index and lookup  */
+        vextractf128 $1, %ymm0, %xmm8
+        vshufps   $136, %xmm8, %xmm0, %xmm9
+        vandps    _iIndexMask+__svml_dexp10_data_internal(%rip), %xmm9, %xmm10
+
+/* iIndex*=sizeof(D); */
+        vpslld    $3, %xmm10, %xmm13
+        vmovd     %xmm13, %edx
+
+/*  2^N  */
+        vpsllq    $45, %ymm0, %ymm0
+        vpextrd   $2, %xmm13, %esi
+        movslq    %edx, %rdx
+        vpextrd   $1, %xmm13, %ecx
+        movslq    %esi, %rsi
+        vpextrd   $3, %xmm13, %edi
+        movslq    %ecx, %rcx
+        movslq    %edi, %rdi
+        vmovsd    (%r8,%rdx), %xmm11
+        vmovsd    (%r8,%rsi), %xmm14
+        vmovhpd   (%r8,%rcx), %xmm11, %xmm12
+        vmovhpd   (%r8,%rdi), %xmm14, %xmm15
+
+/* lM==EXP(2^N) */
+        vpand     _lExpMask+__svml_dexp10_data_internal(%rip), %ymm0, %ymm6
+        vinsertf128 $1, %xmm15, %ymm12, %ymm5
+
+/* Tj*poly */
+        vfmadd213pd %ymm5, %ymm5, %ymm1
+
+/* quick 2^N */
+        vpaddq    %ymm6, %ymm1, %ymm0
+
+/*  Finish   */
+        testl     %eax, %eax
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx r12 r13 r14 r15 eax ymm0 ymm2
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        movq      %rbp, %rsp
+        popq      %rbp
+        cfi_def_cfa(7, 8)
+        cfi_restore(6)
+        ret
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        vmovupd   %ymm2, 32(%rsp)
+        vmovupd   %ymm0, 64(%rsp)
+                                # LOE rbx r12 r13 r14 r15 eax ymm0
+
+        xorl      %edx, %edx
+                                # LOE rbx r12 r13 r14 r15 eax edx
+
+        vzeroupper
+        movq      %r12, 16(%rsp)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 0xff, 0xff, 0xff, 0x22
+        movl      %edx, %r12d
+        movq      %r13, 8(%rsp)
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22
+        movl      %eax, %r13d
+        movq      %r14, (%rsp)
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $4, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        vmovupd   64(%rsp), %ymm0
+
+/* Go to exit */
+        jmp       L(EXIT)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r12 r13 r14 r15 ymm0
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movsd     32(%rsp,%r14,8), %xmm0
+        call      exp10@PLT
+                                # LOE rbx r14 r15 r12d r13d xmm0
+
+        movsd     %xmm0, 64(%rsp,%r14,8)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx r15 r12d r13d
+END(_ZGVdN4v_exp10_avx2)
+
+        .section .rodata, "a"
+        .align 32
+
+#ifdef __svml_dexp10_data_internal_typedef
+typedef unsigned int VUINT32;
+typedef struct
+{
+        __declspec(align(32)) VUINT32 _dbT[(1<<7)][2];
+        __declspec(align(32)) VUINT32 _dbLg2_10[4][2];
+        __declspec(align(32)) VUINT32 _dbShifter[4][2];
+        __declspec(align(32)) VUINT32 _dbInvLg2_10hi[4][2];
+        __declspec(align(32)) VUINT32 _dbInvLg2_10lo[4][2];
+        __declspec(align(32)) VUINT32 _dPC1[4][2];
+        __declspec(align(32)) VUINT32 _dPC2[4][2];
+        __declspec(align(32)) VUINT32 _dPC3[4][2];
+        __declspec(align(32)) VUINT32 _dPC4[4][2];
+        __declspec(align(32)) VUINT32 _dPC5[4][2];
+        __declspec(align(32)) VUINT32 _lExpMask[4][2];
+        __declspec(align(32)) VUINT32 _iIndexMask[8][1];
+        __declspec(align(32)) VUINT32 _iAbsMask[8][1];
+        __declspec(align(32)) VUINT32 _iDomainRange[8][1];
+} __svml_dexp10_data_internal;
+#endif
+__svml_dexp10_data_internal:
+        /*== _dbT ==*/
+        .quad 0x3ff0000000000000    /*2^( 0 /128)*/
+        .quad 0x3ff0163da9fb3335    /*2^( 1 /128)*/
+        .quad 0x3ff02c9a3e778061    /*2^( 2 /128)*/
+        .quad 0x3ff04315e86e7f85    /*2^( 3 /128)*/
+        .quad 0x3ff059b0d3158574    /*2^( 4 /128)*/
+        .quad 0x3ff0706b29ddf6de    /*2^( 5 /128)*/
+        .quad 0x3ff0874518759bc8    /*2^( 6 /128)*/
+        .quad 0x3ff09e3ecac6f383    /*2^( 7 /128)*/
+        .quad 0x3ff0b5586cf9890f    /*2^( 8 /128)*/
+        .quad 0x3ff0cc922b7247f7    /*2^( 9 /128)*/
+        .quad 0x3ff0e3ec32d3d1a2    /*2^( 10 /128)*/
+        .quad 0x3ff0fb66affed31b    /*2^( 11 /128)*/
+        .quad 0x3ff11301d0125b51    /*2^( 12 /128)*/
+        .quad 0x3ff12abdc06c31cc    /*2^( 13 /128)*/
+        .quad 0x3ff1429aaea92de0    /*2^( 14 /128)*/
+        .quad 0x3ff15a98c8a58e51    /*2^( 15 /128)*/
+        .quad 0x3ff172b83c7d517b    /*2^( 16 /128)*/
+        .quad 0x3ff18af9388c8dea    /*2^( 17 /128)*/
+        .quad 0x3ff1a35beb6fcb75    /*2^( 18 /128)*/
+        .quad 0x3ff1bbe084045cd4    /*2^( 19 /128)*/
+        .quad 0x3ff1d4873168b9aa    /*2^( 20 /128)*/
+        .quad 0x3ff1ed5022fcd91d    /*2^( 21 /128)*/
+        .quad 0x3ff2063b88628cd6    /*2^( 22 /128)*/
+        .quad 0x3ff21f49917ddc96    /*2^( 23 /128)*/
+        .quad 0x3ff2387a6e756238    /*2^( 24 /128)*/
+        .quad 0x3ff251ce4fb2a63f    /*2^( 25 /128)*/
+        .quad 0x3ff26b4565e27cdd    /*2^( 26 /128)*/
+        .quad 0x3ff284dfe1f56381    /*2^( 27 /128)*/
+        .quad 0x3ff29e9df51fdee1    /*2^( 28 /128)*/
+        .quad 0x3ff2b87fd0dad990    /*2^( 29 /128)*/
+        .quad 0x3ff2d285a6e4030b    /*2^( 30 /128)*/
+        .quad 0x3ff2ecafa93e2f56    /*2^( 31 /128)*/
+        .quad 0x3ff306fe0a31b715    /*2^( 32 /128)*/
+        .quad 0x3ff32170fc4cd831    /*2^( 33 /128)*/
+        .quad 0x3ff33c08b26416ff    /*2^( 34 /128)*/
+        .quad 0x3ff356c55f929ff1    /*2^( 35 /128)*/
+        .quad 0x3ff371a7373aa9cb    /*2^( 36 /128)*/
+        .quad 0x3ff38cae6d05d866    /*2^( 37 /128)*/
+        .quad 0x3ff3a7db34e59ff7    /*2^( 38 /128)*/
+        .quad 0x3ff3c32dc313a8e5    /*2^( 39 /128)*/
+        .quad 0x3ff3dea64c123422    /*2^( 40 /128)*/
+        .quad 0x3ff3fa4504ac801c    /*2^( 41 /128)*/
+        .quad 0x3ff4160a21f72e2a    /*2^( 42 /128)*/
+        .quad 0x3ff431f5d950a897    /*2^( 43 /128)*/
+        .quad 0x3ff44e086061892d    /*2^( 44 /128)*/
+        .quad 0x3ff46a41ed1d0057    /*2^( 45 /128)*/
+        .quad 0x3ff486a2b5c13cd0    /*2^( 46 /128)*/
+        .quad 0x3ff4a32af0d7d3de    /*2^( 47 /128)*/
+        .quad 0x3ff4bfdad5362a27    /*2^( 48 /128)*/
+        .quad 0x3ff4dcb299fddd0d    /*2^( 49 /128)*/
+        .quad 0x3ff4f9b2769d2ca7    /*2^( 50 /128)*/
+        .quad 0x3ff516daa2cf6642    /*2^( 51 /128)*/
+        .quad 0x3ff5342b569d4f82    /*2^( 52 /128)*/
+        .quad 0x3ff551a4ca5d920f    /*2^( 53 /128)*/
+        .quad 0x3ff56f4736b527da    /*2^( 54 /128)*/
+        .quad 0x3ff58d12d497c7fd    /*2^( 55 /128)*/
+        .quad 0x3ff5ab07dd485429    /*2^( 56 /128)*/
+        .quad 0x3ff5c9268a5946b7    /*2^( 57 /128)*/
+        .quad 0x3ff5e76f15ad2148    /*2^( 58 /128)*/
+        .quad 0x3ff605e1b976dc09    /*2^( 59 /128)*/
+        .quad 0x3ff6247eb03a5585    /*2^( 60 /128)*/
+        .quad 0x3ff6434634ccc320    /*2^( 61 /128)*/
+        .quad 0x3ff6623882552225    /*2^( 62 /128)*/
+        .quad 0x3ff68155d44ca973    /*2^( 63 /128)*/
+        .quad 0x3ff6a09e667f3bcd    /*2^( 64 /128)*/
+        .quad 0x3ff6c012750bdabf    /*2^( 65 /128)*/
+        .quad 0x3ff6dfb23c651a2f    /*2^( 66 /128)*/
+        .quad 0x3ff6ff7df9519484    /*2^( 67 /128)*/
+        .quad 0x3ff71f75e8ec5f74    /*2^( 68 /128)*/
+        .quad 0x3ff73f9a48a58174    /*2^( 69 /128)*/
+        .quad 0x3ff75feb564267c9    /*2^( 70 /128)*/
+        .quad 0x3ff780694fde5d3f    /*2^( 71 /128)*/
+        .quad 0x3ff7a11473eb0187    /*2^( 72 /128)*/
+        .quad 0x3ff7c1ed0130c132    /*2^( 73 /128)*/
+        .quad 0x3ff7e2f336cf4e62    /*2^( 74 /128)*/
+        .quad 0x3ff80427543e1a12    /*2^( 75 /128)*/
+        .quad 0x3ff82589994cce13    /*2^( 76 /128)*/
+        .quad 0x3ff8471a4623c7ad    /*2^( 77 /128)*/
+        .quad 0x3ff868d99b4492ed    /*2^( 78 /128)*/
+        .quad 0x3ff88ac7d98a6699    /*2^( 79 /128)*/
+        .quad 0x3ff8ace5422aa0db    /*2^( 80 /128)*/
+        .quad 0x3ff8cf3216b5448c    /*2^( 81 /128)*/
+        .quad 0x3ff8f1ae99157736    /*2^( 82 /128)*/
+        .quad 0x3ff9145b0b91ffc6    /*2^( 83 /128)*/
+        .quad 0x3ff93737b0cdc5e5    /*2^( 84 /128)*/
+        .quad 0x3ff95a44cbc8520f    /*2^( 85 /128)*/
+        .quad 0x3ff97d829fde4e50    /*2^( 86 /128)*/
+        .quad 0x3ff9a0f170ca07ba    /*2^( 87 /128)*/
+        .quad 0x3ff9c49182a3f090    /*2^( 88 /128)*/
+        .quad 0x3ff9e86319e32323    /*2^( 89 /128)*/
+        .quad 0x3ffa0c667b5de565    /*2^( 90 /128)*/
+        .quad 0x3ffa309bec4a2d33    /*2^( 91 /128)*/
+        .quad 0x3ffa5503b23e255d    /*2^( 92 /128)*/
+        .quad 0x3ffa799e1330b358    /*2^( 93 /128)*/
+        .quad 0x3ffa9e6b5579fdbf    /*2^( 94 /128)*/
+        .quad 0x3ffac36bbfd3f37a    /*2^( 95 /128)*/
+        .quad 0x3ffae89f995ad3ad    /*2^( 96 /128)*/
+        .quad 0x3ffb0e07298db666    /*2^( 97 /128)*/
+        .quad 0x3ffb33a2b84f15fb    /*2^( 98 /128)*/
+        .quad 0x3ffb59728de5593a    /*2^( 99 /128)*/
+        .quad 0x3ffb7f76f2fb5e47    /*2^( 100 /128)*/
+        .quad 0x3ffba5b030a1064a    /*2^( 101 /128)*/
+        .quad 0x3ffbcc1e904bc1d2    /*2^( 102 /128)*/
+        .quad 0x3ffbf2c25bd71e09    /*2^( 103 /128)*/
+        .quad 0x3ffc199bdd85529c    /*2^( 104 /128)*/
+        .quad 0x3ffc40ab5fffd07a    /*2^( 105 /128)*/
+        .quad 0x3ffc67f12e57d14b    /*2^( 106 /128)*/
+        .quad 0x3ffc8f6d9406e7b5    /*2^( 107 /128)*/
+        .quad 0x3ffcb720dcef9069    /*2^( 108 /128)*/
+        .quad 0x3ffcdf0b555dc3fa    /*2^( 109 /128)*/
+        .quad 0x3ffd072d4a07897c    /*2^( 110 /128)*/
+        .quad 0x3ffd2f87080d89f2    /*2^( 111 /128)*/
+        .quad 0x3ffd5818dcfba487    /*2^( 112 /128)*/
+        .quad 0x3ffd80e316c98398    /*2^( 113 /128)*/
+        .quad 0x3ffda9e603db3285    /*2^( 114 /128)*/
+        .quad 0x3ffdd321f301b460    /*2^( 115 /128)*/
+        .quad 0x3ffdfc97337b9b5f    /*2^( 116 /128)*/
+        .quad 0x3ffe264614f5a129    /*2^( 117 /128)*/
+        .quad 0x3ffe502ee78b3ff6    /*2^( 118 /128)*/
+        .quad 0x3ffe7a51fbc74c83    /*2^( 119 /128)*/
+        .quad 0x3ffea4afa2a490da    /*2^( 120 /128)*/
+        .quad 0x3ffecf482d8e67f1    /*2^( 121 /128)*/
+        .quad 0x3ffefa1bee615a27    /*2^( 122 /128)*/
+        .quad 0x3fff252b376bba97    /*2^( 123 /128)*/
+        .quad 0x3fff50765b6e4540    /*2^( 124 /128)*/
+        .quad 0x3fff7bfdad9cbe14    /*2^( 125 /128)*/
+        .quad 0x3fffa7c1819e90d8    /*2^( 126 /128)*/
+        .quad 0x3fffd3c22b8f71f1     /*2^( 127 /128)*/
+        .align 32
+        .quad 0x407a934f0979a371, 0x407a934f0979a371, 0x407a934f0979a371, 0x407a934f0979a371  /* _dbLg2_10*2^K */
+        .align 32
+        .quad 0x4338800000000000, 0x4338800000000000, 0x4338800000000000, 0x4338800000000000  /* _dbShifter */
+        .align 32
+        .quad 0x3f63441350a00000, 0x3f63441350a00000, 0x3f63441350a00000, 0x3f63441350a00000  /* _dbInvLg2_10hi/2^K 53-11-K bits*/
+        .align 32
+        .quad 0xbd10c0219dc1da99, 0xbd10c0219dc1da99, 0xbd10c0219dc1da99, 0xbd10c0219dc1da99  /* _dbInvLg2_10lo/2^K */
+        //PC0 = 1.0
+        .align 32
+        .quad 0x40026bb1bbb55516, 0x40026bb1bbb55516, 0x40026bb1bbb55516, 0x40026bb1bbb55516  /* _dPC1 */
+        .align 32
+        .quad 0x40053524c73ce8e3, 0x40053524c73ce8e3, 0x40053524c73ce8e3, 0x40053524c73ce8e3  /* _dPC2 */
+        .align 32
+        .quad 0x4000470591ccea8b, 0x4000470591ccea8b, 0x4000470591ccea8b, 0x4000470591ccea8b  /* _dPC3 */
+        .align 32
+        .quad 0x3ff2bd767584db59, 0x3ff2bd767584db59, 0x3ff2bd767584db59, 0x3ff2bd767584db59  /* _dPC4 */
+        .align 32
+        .quad 0x3fe144c03efafb54, 0x3fe144c03efafb54, 0x3fe144c03efafb54, 0x3fe144c03efafb54  /* _dPC5 */
+        .align 32
+        .quad 0xfff0000000000000, 0xfff0000000000000, 0xfff0000000000000, 0xfff0000000000000  /* _lExpMask */
+        .align 32
+        .long 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f, 0x0000007f          /* _iIndexMask =(2^K-1)*/
+        //common
+        .align 32
+        .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff          /* _iAbsMask */
+        .align 32
+        .long 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70, 0x40733a70 /* _iDomainRange */
+        .align 32
+        .type	__svml_dexp10_data_internal,@object
+        .size	__svml_dexp10_data_internal,.-__svml_dexp10_data_internal
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core-avx2.S
new file mode 100644
index 0000000000..3aff9446d3
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core-avx2.S
@@ -0,0 +1,20 @@
+/* AVX2 version of vectorized exp10, vector length is 8.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVeN8v_exp10 _ZGVeN8v_exp10_avx2_wrapper
+#include "../svml_d_exp108_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core.c
new file mode 100644
index 0000000000..d592663169
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core.c
@@ -0,0 +1,27 @@
+/* Multiple versions of vectorized exp10, vector length is 8.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVeN8v_exp10
+#include "ifunc-mathvec-avx512-skx.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVeN8v_exp10, __GI__ZGVeN8v_exp10, __redirect__ZGVeN8v_exp10)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core_avx512.S
new file mode 100644
index 0000000000..953cb5bc1a
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp108_core_avx512.S
@@ -0,0 +1,287 @@
+/* Function exp10 vectorized with AVX-512.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *   Typical exp10() implementation, except that:
+ *    - tables are small (16 elements), allowing for fast gathers
+ *    - all arguments processed in the main path
+ *        - final VSCALEF assists branch-free design (correct overflow/underflow and special case responses)
+ *        - a VAND is used to ensure the reduced argument |R|<2, even for large inputs
+ *        - RZ mode used to avoid oveflow to +/-Inf for x*log2(e); helps with special case handling
+ *        - SAE used to avoid spurious flag settings
+ *
+ */
+
+/* Offsets for data table __svml_dexp10_data_internal_avx512
+ */
+#define Exp_tbl_H                     	0
+#define L2E                           	128
+#define Shifter                       	192
+#define L2H                           	256
+#define L2L                           	320
+#define EMask                         	384
+#define poly_coeff6                   	448
+#define poly_coeff5                   	512
+#define poly_coeff4                   	576
+#define poly_coeff3                   	640
+#define poly_coeff2                   	704
+#define poly_coeff1                   	768
+#define AbsMask                       	832
+#define Threshold                     	896
+
+#include <sysdep.h>
+
+        .text
+	.section .text.evex512,"ax",@progbits
+ENTRY(_ZGVeN8v_exp10_skx)
+        pushq     %rbp
+        cfi_def_cfa_offset(16)
+        movq      %rsp, %rbp
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+        andq      $-64, %rsp
+        subq      $192, %rsp
+        vmovups   L2E+__svml_dexp10_data_internal_avx512(%rip), %zmm4
+        vmovups   Shifter+__svml_dexp10_data_internal_avx512(%rip), %zmm2
+        vmovups   L2H+__svml_dexp10_data_internal_avx512(%rip), %zmm5
+        vmovups   L2L+__svml_dexp10_data_internal_avx512(%rip), %zmm3
+
+/* polynomial */
+        vmovups   poly_coeff6+__svml_dexp10_data_internal_avx512(%rip), %zmm6
+        vmovups   poly_coeff4+__svml_dexp10_data_internal_avx512(%rip), %zmm7
+        vmovups   poly_coeff3+__svml_dexp10_data_internal_avx512(%rip), %zmm9
+        vmovups   poly_coeff2+__svml_dexp10_data_internal_avx512(%rip), %zmm8
+        vmovups   poly_coeff1+__svml_dexp10_data_internal_avx512(%rip), %zmm11
+        vmovups   Threshold+__svml_dexp10_data_internal_avx512(%rip), %zmm14
+        vmovaps   %zmm0, %zmm1
+
+/* 2^(52-4)*1.5 + x * log2(e) */
+        vfmadd213pd {rz-sae}, %zmm2, %zmm1, %zmm4
+        vandpd    AbsMask+__svml_dexp10_data_internal_avx512(%rip), %zmm1, %zmm13
+
+/* Z0 ~ x*log2(e), rounded down to 4 fractional bits */
+        vsubpd    {rn-sae}, %zmm2, %zmm4, %zmm0
+
+/* Table lookup: Th */
+        vmovups   __svml_dexp10_data_internal_avx512(%rip), %zmm2
+        vcmppd    $29, {sae}, %zmm14, %zmm13, %k0
+
+/* R = x - Z0*log(2) */
+        vfnmadd213pd {rn-sae}, %zmm1, %zmm0, %zmm5
+        vpermt2pd Exp_tbl_H+64+__svml_dexp10_data_internal_avx512(%rip), %zmm4, %zmm2
+        kmovw     %k0, %edx
+        vfnmadd231pd {rn-sae}, %zmm0, %zmm3, %zmm5
+        vmovups   poly_coeff5+__svml_dexp10_data_internal_avx512(%rip), %zmm3
+
+/* ensure |R|<2 even for special cases */
+        vandpd    EMask+__svml_dexp10_data_internal_avx512(%rip), %zmm5, %zmm12
+        vmulpd    {rn-sae}, %zmm12, %zmm12, %zmm10
+        vmulpd    {rn-sae}, %zmm12, %zmm2, %zmm15
+        vfmadd231pd {rn-sae}, %zmm12, %zmm6, %zmm3
+        vfmadd231pd {rn-sae}, %zmm12, %zmm7, %zmm9
+        vfmadd231pd {rn-sae}, %zmm12, %zmm8, %zmm11
+        vfmadd213pd {rn-sae}, %zmm9, %zmm10, %zmm3
+        vfmadd213pd {rn-sae}, %zmm11, %zmm10, %zmm3
+        vfmadd213pd {rn-sae}, %zmm2, %zmm15, %zmm3
+        vscalefpd {rn-sae}, %zmm0, %zmm3, %zmm0
+        testl     %edx, %edx
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        movq      %rbp, %rsp
+        popq      %rbp
+        cfi_def_cfa(7, 8)
+        cfi_restore(6)
+        ret
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        vmovups   %zmm1, 64(%rsp)
+        vmovups   %zmm0, 128(%rsp)
+                                # LOE rbx r12 r13 r14 r15 edx zmm0
+
+        xorl      %eax, %eax
+                                # LOE rbx r12 r13 r14 r15 eax edx
+
+        vzeroupper
+        movq      %r12, 16(%rsp)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -176; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 0xff, 0xff, 0xff, 0x22
+        movl      %eax, %r12d
+        movq      %r13, 8(%rsp)
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+        movl      %edx, %r13d
+        movq      %r14, (%rsp)
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $8, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        vmovups   128(%rsp), %zmm0
+
+/* Go to exit */
+        jmp       L(EXIT)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -176; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r12 r13 r14 r15 zmm0
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movsd     64(%rsp,%r14,8), %xmm0
+        call      exp10@PLT
+                                # LOE rbx r14 r15 r12d r13d xmm0
+
+        movsd     %xmm0, 128(%rsp,%r14,8)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx r15 r12d r13d
+END(_ZGVeN8v_exp10_skx)
+
+        .section .rodata, "a"
+        .align 64
+
+#ifdef __svml_dexp10_data_internal_avx512_typedef
+typedef unsigned int VUINT32;
+typedef struct {
+        __declspec(align(64)) VUINT32 Exp_tbl_H[16][2];
+        __declspec(align(64)) VUINT32 L2E[8][2];
+        __declspec(align(64)) VUINT32 Shifter[8][2];
+        __declspec(align(64)) VUINT32 L2H[8][2];
+        __declspec(align(64)) VUINT32 L2L[8][2];
+        __declspec(align(64)) VUINT32 EMask[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff6[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff5[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff4[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff3[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff2[8][2];
+        __declspec(align(64)) VUINT32 poly_coeff1[8][2];
+        __declspec(align(64)) VUINT32 AbsMask[8][2];
+        __declspec(align(64)) VUINT32 Threshold[8][2];
+    } __svml_dexp10_data_internal_avx512;
+#endif
+__svml_dexp10_data_internal_avx512:
+        /*== Exp_tbl_H ==*/
+        .quad 0x3ff0000000000000
+        .quad 0x3ff0b5586cf9890f
+        .quad 0x3ff172b83c7d517b
+        .quad 0x3ff2387a6e756238
+        .quad 0x3ff306fe0a31b715
+        .quad 0x3ff3dea64c123422
+        .quad 0x3ff4bfdad5362a27
+        .quad 0x3ff5ab07dd485429
+        .quad 0x3ff6a09e667f3bcd
+        .quad 0x3ff7a11473eb0187
+        .quad 0x3ff8ace5422aa0db
+        .quad 0x3ff9c49182a3f090
+        .quad 0x3ffae89f995ad3ad
+        .quad 0x3ffc199bdd85529c
+        .quad 0x3ffd5818dcfba487
+        .quad 0x3ffea4afa2a490da
+        /*== log2(e) ==*/
+        .align 64
+        .quad 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371, 0x400A934F0979A371
+        /*== Shifter=2^(52-4)*1.5 ==*/
+        .align 64
+        .quad 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0, 0x42f8000000003ff0
+        /*== L2H = log(2)_high ==*/
+        .align 64
+        .quad 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff, 0x3fd34413509f79ff
+        /*== L2L = log(2)_low ==*/
+        .align 64
+        .quad 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21, 0xbc49dc1da994fd21
+        /*== EMask ==*/
+        .align 64
+        .quad 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff, 0xbfffffffffffffff
+        /*== poly_coeff6 ==*/
+        .align 64
+        .quad 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020, 0x3fcb137ed8ac2020
+        /*== poly_coeff5 ==*/
+        .align 64
+        .quad 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424, 0x3fe141a8e24f9424
+        /*== poly_coeff4 ==*/
+        .align 64
+        .quad 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d, 0x3ff2bd77a0926c9d
+        /*== poly_coeff3 ==*/
+        .align 64
+        .quad 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8, 0x40004705908704c8
+        /*== poly_coeff2 ==*/
+        .align 64
+        .quad 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25, 0x40053524c73dfe25
+        /*== poly_coeff1 ==*/
+        .align 64
+        .quad 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2, 0x40026bb1bbb554c2
+        /*== AbsMask ==*/
+        .align 64
+        .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff
+        /*== Threshold ==*/
+        .align 64
+        .quad 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41, 0x40733A7146F72A41
+        .align 64
+        .type	__svml_dexp10_data_internal_avx512,@object
+        .size	__svml_dexp10_data_internal_avx512,.-__svml_dexp10_data_internal_avx512
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core-avx2.S
new file mode 100644
index 0000000000..dda41c9c8f
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core-avx2.S
@@ -0,0 +1,20 @@
+/* AVX2 version of vectorized exp10f.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVeN16v_exp10f _ZGVeN16v_exp10f_avx2_wrapper
+#include "../svml_s_exp10f16_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core.c
new file mode 100644
index 0000000000..8176a5912b
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core.c
@@ -0,0 +1,28 @@
+/* Multiple versions of vectorized exp10f, vector length is 16.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVeN16v_exp10f
+#include "ifunc-mathvec-avx512-skx.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVeN16v_exp10f, __GI__ZGVeN16v_exp10f,
+	       __redirect__ZGVeN16v_exp10f)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core_avx512.S
new file mode 100644
index 0000000000..fc9309c90f
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f16_core_avx512.S
@@ -0,0 +1,269 @@
+/* Function exp10f vectorized with AVX-512.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *   Typical exp10() implementation, except that:
+ *    - tables are small (16 elements), allowing for fast gathers
+ *    - all arguments processed in the main path
+ *        - final VSCALEF assists branch-free design (correct overflow/underflow and special case responses)
+ *        - a VAND is used to ensure the reduced argument |R|<2, even for large inputs
+ *        - RZ mode used to avoid oveflow to +/-Inf for x*log2(e); helps with special case handling
+ *        - SAE used to avoid spurious flag settings
+ *
+ */
+
+/* Offsets for data table __svml_sexp10_data_internal_avx512
+ */
+#define Exp_tbl_L                     	0
+#define Exp_tbl_H                     	128
+#define L2E                           	256
+#define Shifter                       	320
+#define L2H                           	384
+#define L2L                           	448
+#define EMask                         	512
+#define AbsMask                       	576
+#define Threshold                     	640
+#define poly_coeff2                   	704
+#define poly_coeff1                   	768
+
+#include <sysdep.h>
+
+        .text
+	.section .text.exex512,"ax",@progbits
+ENTRY(_ZGVeN16v_exp10f_skx)
+        pushq     %rbp
+        cfi_def_cfa_offset(16)
+        movq      %rsp, %rbp
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+        andq      $-64, %rsp
+        subq      $192, %rsp
+        vmovups   L2E+__svml_sexp10_data_internal_avx512(%rip), %zmm2
+        vmovups   Shifter+__svml_sexp10_data_internal_avx512(%rip), %zmm1
+        vmovups   L2H+__svml_sexp10_data_internal_avx512(%rip), %zmm5
+        vmovups   L2L+__svml_sexp10_data_internal_avx512(%rip), %zmm4
+
+/* ensure |R|<2 even for special cases */
+        vmovups   EMask+__svml_sexp10_data_internal_avx512(%rip), %zmm6
+        vmovups   poly_coeff2+__svml_sexp10_data_internal_avx512(%rip), %zmm9
+
+/* 2^(52-4)*1.5 + x * log2(e) */
+        vfmadd213ps {rz-sae}, %zmm1, %zmm0, %zmm2
+        vmovups   poly_coeff1+__svml_sexp10_data_internal_avx512(%rip), %zmm10
+        vmovups   __svml_sexp10_data_internal_avx512(%rip), %zmm8
+        vmovups   Exp_tbl_H+__svml_sexp10_data_internal_avx512(%rip), %zmm15
+        vmovups   Threshold+__svml_sexp10_data_internal_avx512(%rip), %zmm13
+        vpsrld    $5, %zmm2, %zmm3
+
+/* Z0 ~ x*log2(e), rounded down to 6 fractional bits */
+        vsubps    {rn-sae}, %zmm1, %zmm2, %zmm1
+        vpermt2ps Exp_tbl_L+64+__svml_sexp10_data_internal_avx512(%rip), %zmm2, %zmm8
+        vpermt2ps Exp_tbl_H+64+__svml_sexp10_data_internal_avx512(%rip), %zmm3, %zmm15
+        vandps    AbsMask+__svml_sexp10_data_internal_avx512(%rip), %zmm0, %zmm12
+
+/* R = x - Z0*log(2) */
+        vfnmadd213ps {rn-sae}, %zmm0, %zmm1, %zmm5
+        vcmpps    $29, {sae}, %zmm13, %zmm12, %k0
+        vfnmadd231ps {rn-sae}, %zmm1, %zmm4, %zmm5
+        kmovw     %k0, %edx
+        vrangeps  $2, {sae}, %zmm6, %zmm5, %zmm11
+        vfmadd231ps {rn-sae}, %zmm11, %zmm9, %zmm10
+        vmulps    {rn-sae}, %zmm11, %zmm10, %zmm14
+
+/* x!=0? */
+        vpxord    %zmm7, %zmm7, %zmm7
+        vcmpps    $4, {sae}, %zmm7, %zmm0, %k1
+
+/* Th*Tl */
+        vmulps    {rn-sae}, %zmm8, %zmm15, %zmm15{%k1}
+        vfmadd213ps {rn-sae}, %zmm15, %zmm14, %zmm15
+        vscalefps {rn-sae}, %zmm1, %zmm15, %zmm1
+        testl     %edx, %edx
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        vmovaps   %zmm1, %zmm0
+        movq      %rbp, %rsp
+        popq      %rbp
+        cfi_def_cfa(7, 8)
+        cfi_restore(6)
+        ret
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        vmovups   %zmm0, 64(%rsp)
+        vmovups   %zmm1, 128(%rsp)
+                                # LOE rbx r12 r13 r14 r15 edx zmm1
+
+        xorl      %eax, %eax
+                                # LOE rbx r12 r13 r14 r15 eax edx
+
+        vzeroupper
+        movq      %r12, 16(%rsp)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -176; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 0xff, 0xff, 0xff, 0x22
+        movl      %eax, %r12d
+        movq      %r13, 8(%rsp)
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+        movl      %edx, %r13d
+        movq      %r14, (%rsp)
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $16, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        vmovups   128(%rsp), %zmm1
+
+/* Go to exit */
+        jmp       L(EXIT)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -176; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r12 r13 r14 r15 zmm1
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movss     64(%rsp,%r14,4), %xmm0
+        call      exp10f@PLT
+                                # LOE rbx r14 r15 r12d r13d xmm0
+
+        movss     %xmm0, 128(%rsp,%r14,4)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx r15 r12d r13d
+END(_ZGVeN16v_exp10f_skx)
+
+        .section .rodata, "a"
+        .align 64
+
+#ifdef __svml_sexp10_data_internal_avx512_typedef
+typedef unsigned int VUINT32;
+typedef struct {
+        __declspec(align(64)) VUINT32 Exp_tbl_L[32][1];
+        __declspec(align(64)) VUINT32 Exp_tbl_H[32][1];
+        __declspec(align(64)) VUINT32 L2E[16][1];
+        __declspec(align(64)) VUINT32 Shifter[16][1];
+        __declspec(align(64)) VUINT32 L2H[16][1];
+        __declspec(align(64)) VUINT32 L2L[16][1];
+        __declspec(align(64)) VUINT32 EMask[16][1];
+        __declspec(align(64)) VUINT32 AbsMask[16][1];
+        __declspec(align(64)) VUINT32 Threshold[16][1];
+        __declspec(align(64)) VUINT32 poly_coeff2[16][1];
+        __declspec(align(64)) VUINT32 poly_coeff1[16][1];
+    } __svml_sexp10_data_internal_avx512;
+#endif
+__svml_sexp10_data_internal_avx512:
+        /*== Exp_tbl_L ==*/
+        .long 0x3f800001, 0x3f801631, 0x3f802c65, 0x3f80429d
+        .long 0x3f8058d9, 0x3f806f18, 0x3f80855c, 0x3f809ba3
+        .long 0x3f80b1ee, 0x3f80c83d, 0x3f80de90, 0x3f80f4e7
+        .long 0x3f810b42, 0x3f8121a0, 0x3f813803, 0x3f814e69
+        .long 0x3f8164d3, 0x3f817b41, 0x3f8191b3, 0x3f81a829
+        .long 0x3f81bea2, 0x3f81d520, 0x3f81eba2, 0x3f820227
+        .long 0x3f8218b0, 0x3f822f3d, 0x3f8245cf, 0x3f825c64
+        .long 0x3f8272fd, 0x3f828999, 0x3f82a03a, 0x3f82b6df
+        /*== Exp_tbl_H ==*/
+        .align 64
+        .long 0x3f800000, 0x3f82cd87, 0x3f85aac3, 0x3f88980f
+        .long 0x3f8b95c2, 0x3f8ea43a, 0x3f91c3d3, 0x3f94f4f0
+        .long 0x3f9837f0, 0x3f9b8d3a, 0x3f9ef532, 0x3fa27043
+        .long 0x3fa5fed7, 0x3fa9a15b, 0x3fad583f, 0x3fb123f6
+        .long 0x3fb504f3, 0x3fb8fbaf, 0x3fbd08a4, 0x3fc12c4d
+        .long 0x3fc5672a, 0x3fc9b9be, 0x3fce248c, 0x3fd2a81e
+        .long 0x3fd744fd, 0x3fdbfbb8, 0x3fe0ccdf, 0x3fe5b907
+        .long 0x3feac0c7, 0x3fefe4ba, 0x3ff5257d, 0x3ffa83b3
+        /*== log2(10) ==*/
+        .align 64
+        .long 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78, 0x40549A78
+        /*== Shifter=2^(23-10)*1.5 ==*/
+        .align 64
+        .long 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000, 0x46400000
+        /*== L2H = log(2)_high ==*/
+        .align 64
+        .long 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b, 0x3e9a209b
+        /*== L2L = log(2)_low ==*/
+        .align 64
+        .long 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860, 0xb2760860
+        /*== EMask ==*/
+        .align 64
+        .long 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000
+        /*== AbsMask ==*/
+        .align 64
+        .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff
+        /*== Threshold ==*/
+        .align 64
+        .long 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818, 0x4217B818
+        /*== poly_coeff2 ==*/
+        .align 64
+        .long 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA, 0x4029B7DA
+        /*== poly_coeff1 ==*/
+        .align 64
+        .long 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D, 0x40135D8D
+        .align 64
+        .type	__svml_sexp10_data_internal_avx512,@object
+        .size	__svml_sexp10_data_internal_avx512,.-__svml_sexp10_data_internal_avx512
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core-sse2.S
new file mode 100644
index 0000000000..460d01357d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core-sse2.S
@@ -0,0 +1,20 @@
+/* SSE2 version of vectorized exp10f, vector length is 4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVbN4v_exp10f _ZGVbN4v_exp10f_sse2
+#include "../svml_s_exp10f4_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core.c
new file mode 100644
index 0000000000..7ce90a9bae
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core.c
@@ -0,0 +1,28 @@
+/* Multiple versions of vectorized exp10f, vector length is 4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVbN4v_exp10f
+#include "ifunc-mathvec-sse4_1.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVbN4v_exp10f, __GI__ZGVbN4v_exp10f,
+	       __redirect__ZGVbN4v_exp10f)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core_sse4.S
new file mode 100644
index 0000000000..879592b789
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f4_core_sse4.S
@@ -0,0 +1,311 @@
+/* Function exp10f vectorized with SSE4.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *
+ *   exp10(x)  = 2^x/log10(2) = 2^n * (1 + T[j]) * (1 + P(y))
+ *   where
+ *        x = m*log10(2)/K + y,  y in [-log10(2)/K..log10(2)/K]
+ *        m = n*K + j,           m,n,j - signed integer, j in [-K/2..K/2]
+ *
+ *        values of 2^j/K are tabulated
+ *
+ *        P(y) is a minimax polynomial approximation of exp10(x)-1
+ *        on small interval [-log10(2)/K..log10(2)/K]
+ *
+ *  Special cases:
+ *
+ *   exp10(NaN)  = NaN
+ *   exp10(+INF) = +INF
+ *   exp10(-INF) = 0
+ *   exp10(x)    = 1 for subnormals
+ *   For IEEE float
+ *     if x >  38.5318412780761720 then exp10f(x) overflow
+ *     if x < -45.4555282592773440 then exp10f(x) underflow
+ *
+ */
+
+/* Offsets for data table __svml_sexp10_data_internal
+ */
+#define _sT                           	0
+#define _sLg2_10                      	128
+#define _sShifter                     	144
+#define _sInvLg2_10hi                 	160
+#define _sInvLg2_10lo                 	176
+#define _sPC0                         	192
+#define _sPC1                         	208
+#define _sPC2                         	224
+#define _iIndexMask                   	240
+#define _iAbsMask                     	256
+#define _iDomainRange                 	272
+
+#include <sysdep.h>
+
+        .text
+	.section .text.sse4,"ax",@progbits
+ENTRY(_ZGVbN4v_exp10f_sse4)
+        subq      $72, %rsp
+        cfi_def_cfa_offset(80)
+        movaps    %xmm0, %xmm4
+
+/*  Load arument  */
+        movups    _sLg2_10+__svml_sexp10_data_internal(%rip), %xmm2
+        lea       __svml_sexp10_data_internal(%rip), %r8
+        mulps     %xmm4, %xmm2
+        movups    _sShifter+__svml_sexp10_data_internal(%rip), %xmm5
+
+/*  R  */
+        movups    _sInvLg2_10hi+__svml_sexp10_data_internal(%rip), %xmm14
+        addps     %xmm5, %xmm2
+        movaps    %xmm2, %xmm1
+        movups    _sInvLg2_10lo+__svml_sexp10_data_internal(%rip), %xmm15
+        subps     %xmm5, %xmm1
+        mulps     %xmm1, %xmm14
+        movaps    %xmm4, %xmm5
+        mulps     %xmm1, %xmm15
+        subps     %xmm14, %xmm5
+
+/*
+ *  Polynomial
+ * exp10 = 2^N*(Tj+Tj*poly)
+ * poly(sN) = {1+later} a0+a1*sR
+ */
+        movups    _sPC2+__svml_sexp10_data_internal(%rip), %xmm1
+        subps     %xmm15, %xmm5
+        mulps     %xmm5, %xmm1
+        movdqu    _iIndexMask+__svml_sexp10_data_internal(%rip), %xmm3
+
+/*  Index and lookup  */
+        movdqa    %xmm3, %xmm10
+
+/* remove index bits */
+        pandn     %xmm2, %xmm3
+        pand      %xmm2, %xmm10
+
+/*  2^N  */
+        pslld     $18, %xmm3
+
+/* iIndex *= sizeof(S); */
+        pslld     $2, %xmm10
+        addps     _sPC1+__svml_sexp10_data_internal(%rip), %xmm1
+        movd      %xmm10, %edx
+        pshufd    $1, %xmm10, %xmm7
+        pshufd    $2, %xmm10, %xmm9
+        pshufd    $3, %xmm10, %xmm11
+        movd      %xmm7, %ecx
+        movd      %xmm9, %esi
+        movd      %xmm11, %edi
+
+/* Check for overflow\underflow  */
+        movdqu    _iAbsMask+__svml_sexp10_data_internal(%rip), %xmm6
+        pand      %xmm4, %xmm6
+        mulps     %xmm1, %xmm5
+        movslq    %edx, %rdx
+        addps     _sPC0+__svml_sexp10_data_internal(%rip), %xmm5
+        movslq    %ecx, %rcx
+        movslq    %esi, %rsi
+        movslq    %edi, %rdi
+        movd      (%r8,%rdx), %xmm0
+        movd      (%r8,%rcx), %xmm8
+        movd      (%r8,%rsi), %xmm13
+        movd      (%r8,%rdi), %xmm12
+        punpckldq %xmm8, %xmm0
+        punpckldq %xmm12, %xmm13
+        punpcklqdq %xmm13, %xmm0
+
+/* Tj_l+Tj_h*poly */
+        mulps     %xmm0, %xmm5
+        pcmpgtd   _iDomainRange+__svml_sexp10_data_internal(%rip), %xmm6
+        addps     %xmm5, %xmm0
+        movmskps  %xmm6, %eax
+
+/* quick mul 2^N */
+        paddd     %xmm3, %xmm0
+
+/*  Finish   */
+        testl     %eax, %eax
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx rbp r12 r13 r14 r15 eax xmm0 xmm4
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        addq      $72, %rsp
+        cfi_def_cfa_offset(8)
+        ret
+        cfi_def_cfa_offset(80)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        movups    %xmm4, 32(%rsp)
+        movups    %xmm0, 48(%rsp)
+                                # LOE rbx rbp r12 r13 r14 r15 eax
+
+        xorl      %edx, %edx
+        movq      %r12, 16(%rsp)
+        cfi_offset(12, -64)
+        movl      %edx, %r12d
+        movq      %r13, 8(%rsp)
+        cfi_offset(13, -72)
+        movl      %eax, %r13d
+        movq      %r14, (%rsp)
+        cfi_offset(14, -80)
+                                # LOE rbx rbp r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx rbp r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $4, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx rbp r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        movups    48(%rsp), %xmm0
+
+/* Go to exit */
+        jmp       L(EXIT)
+        cfi_offset(12, -64)
+        cfi_offset(13, -72)
+        cfi_offset(14, -80)
+                                # LOE rbx rbp r12 r13 r14 r15 xmm0
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movss     32(%rsp,%r14,4), %xmm0
+        call      exp10f@PLT
+                                # LOE rbx rbp r14 r15 r12d r13d xmm0
+
+        movss     %xmm0, 48(%rsp,%r14,4)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx rbp r15 r12d r13d
+END(_ZGVbN4v_exp10f_sse4)
+
+        .section .rodata, "a"
+        .align 16
+
+#ifdef __svml_sexp10_data_internal_typedef
+typedef unsigned int VUINT32;
+typedef struct
+{
+        __declspec(align(16)) VUINT32 _sT[(1<<5)][1];
+        __declspec(align(16)) VUINT32 _sLg2_10[4][1];
+        __declspec(align(16)) VUINT32 _sShifter[4][1];
+        __declspec(align(16)) VUINT32 _sInvLg2_10hi[4][1];
+        __declspec(align(16)) VUINT32 _sInvLg2_10lo[4][1];
+        __declspec(align(16)) VUINT32 _sPC0[4][1];
+        __declspec(align(16)) VUINT32 _sPC1[4][1];
+        __declspec(align(16)) VUINT32 _sPC2[4][1];
+        __declspec(align(16)) VUINT32 _iIndexMask[4][1];
+        __declspec(align(16)) VUINT32 _iAbsMask[4][1];
+        __declspec(align(16)) VUINT32 _iDomainRange[4][1];
+} __svml_sexp10_data_internal;
+#endif
+__svml_sexp10_data_internal:
+        /*== _sT ==*/
+        .long 0x3f800000  // 2^( 0 /32 )
+        .long 0x3f82cd87  // 2^( 1 /32 )
+        .long 0x3f85aac3  // 2^( 2 /32 )
+        .long 0x3f88980f  // 2^( 3 /32 )
+        .long 0x3f8b95c2  // 2^( 4 /32 )
+        .long 0x3f8ea43a  // 2^( 5 /32 )
+        .long 0x3f91c3d3  // 2^( 6 /32 )
+        .long 0x3f94f4f0  // 2^( 7 /32 )
+        .long 0x3f9837f0  // 2^( 8 /32 )
+        .long 0x3f9b8d3a  // 2^( 9 /32 )
+        .long 0x3f9ef532  // 2^( 10/32 )
+        .long 0x3fa27043  // 2^( 11/32 )
+        .long 0x3fa5fed7  // 2^( 12/32 )
+        .long 0x3fa9a15b  // 2^( 13/32 )
+        .long 0x3fad583f  // 2^( 14/32 )
+        .long 0x3fb123f6  // 2^( 15/32 )
+        .long 0x3fb504f3  // 2^( 16/32 )
+        .long 0x3fb8fbaf  // 2^( 17/32 )
+        .long 0x3fbd08a4  // 2^( 18/32 )
+        .long 0x3fc12c4d  // 2^( 19/32 )
+        .long 0x3fc5672a  // 2^( 20/32 )
+        .long 0x3fc9b9be  // 2^( 21/32 )
+        .long 0x3fce248c  // 2^( 22/32 )
+        .long 0x3fd2a81e  // 2^( 23/32 )
+        .long 0x3fd744fd  // 2^( 24/32 )
+        .long 0x3fdbfbb8  // 2^( 25/32 )
+        .long 0x3fe0ccdf  // 2^( 26/32 )
+        .long 0x3fe5b907  // 2^( 27/32 )
+        .long 0x3feac0c7  // 2^( 28/32 )
+        .long 0x3fefe4ba  // 2^( 29/32 )
+        .long 0x3ff5257d  // 2^( 30/32 )
+        .long 0x3ffa83b3  // 2^( 31/32 )
+        .align 16
+        .long 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78  /* _sLg2_10*2^K   */
+        .align 16
+        .long 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000  /* _sShifter) */
+        .align 16
+        .long 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000  /* _sInvLg2_10hi/2^K hi (24-K-7) bits*/
+        .align 16
+        .long 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc  /* _sInvLg2_10lo/2^K  lo bits */
+        // otherwise exp10(0) won't produce exact 1.0
+        .align 16
+        .long 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868  /* _sPC0 */
+        .align 16
+        .long 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b  /* _sPC1 */
+        .align 16
+        .long 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2  /* _sPC2 */
+        .align 16
+        .long 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f  /* _iIndexMask =(2^K-1)*/
+        //common
+        .align 16
+        .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff   /* _iAbsMask */
+        .align 16
+        .long 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818   /* _iDomainRange=-log10(max_denormal=0x007fffff) RZ */
+        .align 16
+        .type	__svml_sexp10_data_internal,@object
+        .size	__svml_sexp10_data_internal,.-__svml_sexp10_data_internal
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core-sse.S
new file mode 100644
index 0000000000..3f3fe252da
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core-sse.S
@@ -0,0 +1,20 @@
+/* SSE version of vectorized exp10f, vector length is 8.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+    Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define _ZGVdN8v_exp10f _ZGVdN8v_exp10f_sse_wrapper
+#include "../svml_s_exp10f8_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core.c
new file mode 100644
index 0000000000..1f5ed5a59d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core.c
@@ -0,0 +1,28 @@
+/* Multiple versions of vectorized exp10f, vector length is 8.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#define SYMBOL_NAME _ZGVdN8v_exp10f
+#include "ifunc-mathvec-avx2.h"
+
+libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ());
+
+#ifdef SHARED
+__hidden_ver1 (_ZGVdN8v_exp10f, __GI__ZGVdN8v_exp10f,
+	       __redirect__ZGVdN8v_exp10f)
+  __attribute__ ((visibility ("hidden")));
+#endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core_avx2.S
new file mode 100644
index 0000000000..b576412cf1
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp10f8_core_avx2.S
@@ -0,0 +1,331 @@
+/* Function exp10f vectorized with AVX2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   https://www.gnu.org/licenses/.  */
+
+/*
+ * ALGORITHM DESCRIPTION:
+ *
+ *   exp10(x)  = 2^x/log10(2) = 2^n * (1 + T[j]) * (1 + P(y))
+ *   where
+ *        x = m*log10(2)/K + y,  y in [-log10(2)/K..log10(2)/K]
+ *        m = n*K + j,           m,n,j - signed integer, j in [-K/2..K/2]
+ *
+ *        values of 2^j/K are tabulated
+ *
+ *        P(y) is a minimax polynomial approximation of exp10(x)-1
+ *        on small interval [-log10(2)/K..log10(2)/K]
+ *
+ *  Special cases:
+ *
+ *   exp10(NaN)  = NaN
+ *   exp10(+INF) = +INF
+ *   exp10(-INF) = 0
+ *   exp10(x)    = 1 for subnormals
+ *   For IEEE float
+ *     if x >  38.5318412780761720 then exp10f(x) overflow
+ *     if x < -45.4555282592773440 then exp10f(x) underflow
+ *
+ */
+
+/* Offsets for data table __svml_sexp10_data_internal
+ */
+#define _sT                           	0
+#define _sLg2_10                      	128
+#define _sShifter                     	160
+#define _sInvLg2_10hi                 	192
+#define _sInvLg2_10lo                 	224
+#define _sPC0                         	256
+#define _sPC1                         	288
+#define _sPC2                         	320
+#define _iIndexMask                   	352
+#define _iAbsMask                     	384
+#define _iDomainRange                 	416
+
+#include <sysdep.h>
+
+        .text
+	.section .text.avx2,"ax",@progbits
+ENTRY(_ZGVdN8v_exp10f_avx2)
+        pushq     %rbp
+        cfi_def_cfa_offset(16)
+        movq      %rsp, %rbp
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+        andq      $-32, %rsp
+        subq      $96, %rsp
+        lea       __svml_sexp10_data_internal(%rip), %rax
+        vmovups   _sShifter+__svml_sexp10_data_internal(%rip), %ymm4
+
+/*  Load arument  */
+        vmovups   _sLg2_10+__svml_sexp10_data_internal(%rip), %ymm1
+        vmovups   _iIndexMask+__svml_sexp10_data_internal(%rip), %ymm2
+        vmovaps   %ymm0, %ymm3
+        vfmadd213ps %ymm4, %ymm3, %ymm1
+
+/*  Index and lookup  */
+        vandps    %ymm2, %ymm1, %ymm7
+
+/* iIndex *= sizeof(S); */
+        vpslld    $2, %ymm7, %ymm10
+        vsubps    %ymm4, %ymm1, %ymm0
+
+/* Check for overflow\underflow  */
+        vandps    _iAbsMask+__svml_sexp10_data_internal(%rip), %ymm3, %ymm5
+        vpcmpgtd  _iDomainRange+__svml_sexp10_data_internal(%rip), %ymm5, %ymm6
+        vmovmskps %ymm6, %edx
+        vmovd     %xmm10, %ecx
+        vextractf128 $1, %ymm10, %xmm6
+        vpextrd   $1, %xmm10, %esi
+        vpextrd   $2, %xmm10, %edi
+        vpextrd   $3, %xmm10, %r8d
+        movslq    %ecx, %rcx
+        movslq    %esi, %rsi
+        movslq    %edi, %rdi
+        movslq    %r8d, %r8
+        vmovd     (%rax,%rcx), %xmm8
+        vmovd     (%rax,%rsi), %xmm9
+        vmovd     (%rax,%rdi), %xmm11
+        vmovd     (%rax,%r8), %xmm12
+        vpunpckldq %xmm9, %xmm8, %xmm13
+        vpunpckldq %xmm12, %xmm11, %xmm14
+        vpunpcklqdq %xmm14, %xmm13, %xmm15
+
+/*  R  */
+        vmovups   _sInvLg2_10hi+__svml_sexp10_data_internal(%rip), %ymm13
+        vmovd     %xmm6, %r9d
+        vfnmadd213ps %ymm3, %ymm0, %ymm13
+        vpextrd   $1, %xmm6, %r10d
+        movslq    %r9d, %r9
+        movslq    %r10d, %r10
+        vfnmadd132ps _sInvLg2_10lo+__svml_sexp10_data_internal(%rip), %ymm13, %ymm0
+        vmovd     (%rax,%r9), %xmm4
+        vmovd     (%rax,%r10), %xmm5
+        vpunpckldq %xmm5, %xmm4, %xmm9
+
+/*
+ *  Polynomial
+ * exp10 = 2^N*(Tj+Tj*poly)
+ * poly(sN) = {1+later} a0+a1*sR
+ */
+        vmovups   _sPC2+__svml_sexp10_data_internal(%rip), %ymm4
+        vfmadd213ps _sPC1+__svml_sexp10_data_internal(%rip), %ymm0, %ymm4
+        vpextrd   $2, %xmm6, %r11d
+        vpextrd   $3, %xmm6, %ecx
+        movslq    %r11d, %r11
+        movslq    %ecx, %rcx
+        vfmadd213ps _sPC0+__svml_sexp10_data_internal(%rip), %ymm0, %ymm4
+        vmovd     (%rax,%r11), %xmm7
+        vmovd     (%rax,%rcx), %xmm8
+        vpunpckldq %xmm8, %xmm7, %xmm11
+
+/* remove index bits */
+        vpandn    %ymm1, %ymm2, %ymm0
+        vpunpcklqdq %xmm11, %xmm9, %xmm12
+
+/*  2^N  */
+        vpslld    $18, %ymm0, %ymm1
+        vinsertf128 $1, %xmm12, %ymm15, %ymm14
+
+/* Tj_l+Tj_h*poly */
+        vfmadd213ps %ymm14, %ymm14, %ymm4
+
+/* quick mul 2^N */
+        vpaddd    %ymm1, %ymm4, %ymm0
+
+/*  Finish   */
+        testl     %edx, %edx
+
+/* Go to special inputs processing branch */
+        jne       L(SPECIAL_VALUES_BRANCH)
+                                # LOE rbx r12 r13 r14 r15 edx ymm0 ymm3
+
+/* Restore registers
+ * and exit the function
+ */
+
+L(EXIT):
+        movq      %rbp, %rsp
+        popq      %rbp
+        cfi_def_cfa(7, 8)
+        cfi_restore(6)
+        ret
+        cfi_def_cfa(6, 16)
+        cfi_offset(6, -16)
+
+/* Branch to process
+ * special inputs
+ */
+
+L(SPECIAL_VALUES_BRANCH):
+        vmovups   %ymm3, 32(%rsp)
+        vmovups   %ymm0, 64(%rsp)
+                                # LOE rbx r12 r13 r14 r15 edx ymm0
+
+        xorl      %eax, %eax
+                                # LOE rbx r12 r13 r14 r15 eax edx
+
+        vzeroupper
+        movq      %r12, 16(%rsp)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 0xff, 0xff, 0xff, 0x22
+        movl      %eax, %r12d
+        movq      %r13, 8(%rsp)
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22
+        movl      %edx, %r13d
+        movq      %r14, (%rsp)
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r15 r12d r13d
+
+/* Range mask
+ * bits check
+ */
+
+L(RANGEMASK_CHECK):
+        btl       %r12d, %r13d
+
+/* Call scalar math function */
+        jc        L(SCALAR_MATH_CALL)
+                                # LOE rbx r15 r12d r13d
+
+/* Special inputs
+ * processing loop
+ */
+
+L(SPECIAL_VALUES_LOOP):
+        incl      %r12d
+        cmpl      $8, %r12d
+
+/* Check bits in range mask */
+        jl        L(RANGEMASK_CHECK)
+                                # LOE rbx r15 r12d r13d
+
+        movq      16(%rsp), %r12
+        cfi_restore(12)
+        movq      8(%rsp), %r13
+        cfi_restore(13)
+        movq      (%rsp), %r14
+        cfi_restore(14)
+        vmovups   64(%rsp), %ymm0
+
+/* Go to exit */
+        jmp       L(EXIT)
+        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22
+        /*  DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus)  */
+        .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22
+                                # LOE rbx r12 r13 r14 r15 ymm0
+
+/* Scalar math fucntion call
+ * to process special input
+ */
+
+L(SCALAR_MATH_CALL):
+        movl      %r12d, %r14d
+        movss     32(%rsp,%r14,4), %xmm0
+        call      exp10f@PLT
+                                # LOE rbx r14 r15 r12d r13d xmm0
+
+        movss     %xmm0, 64(%rsp,%r14,4)
+
+/* Process special inputs in loop */
+        jmp       L(SPECIAL_VALUES_LOOP)
+                                # LOE rbx r15 r12d r13d
+END(_ZGVdN8v_exp10f_avx2)
+
+        .section .rodata, "a"
+        .align 32
+
+#ifdef __svml_sexp10_data_internal_typedef
+typedef unsigned int VUINT32;
+typedef struct
+{
+        __declspec(align(32)) VUINT32 _sT[(1<<5)][1];
+        __declspec(align(32)) VUINT32 _sLg2_10[8][1];
+        __declspec(align(32)) VUINT32 _sShifter[8][1];
+        __declspec(align(32)) VUINT32 _sInvLg2_10hi[8][1];
+        __declspec(align(32)) VUINT32 _sInvLg2_10lo[8][1];
+        __declspec(align(32)) VUINT32 _sPC0[8][1];
+        __declspec(align(32)) VUINT32 _sPC1[8][1];
+        __declspec(align(32)) VUINT32 _sPC2[8][1];
+        __declspec(align(32)) VUINT32 _iIndexMask[8][1];
+        __declspec(align(32)) VUINT32 _iAbsMask[8][1];
+        __declspec(align(32)) VUINT32 _iDomainRange[8][1];
+} __svml_sexp10_data_internal;
+#endif
+__svml_sexp10_data_internal:
+        /*== _sT ==*/
+        .long 0x3f800000  // 2^( 0 /32 )
+        .long 0x3f82cd87  // 2^( 1 /32 )
+        .long 0x3f85aac3  // 2^( 2 /32 )
+        .long 0x3f88980f  // 2^( 3 /32 )
+        .long 0x3f8b95c2  // 2^( 4 /32 )
+        .long 0x3f8ea43a  // 2^( 5 /32 )
+        .long 0x3f91c3d3  // 2^( 6 /32 )
+        .long 0x3f94f4f0  // 2^( 7 /32 )
+        .long 0x3f9837f0  // 2^( 8 /32 )
+        .long 0x3f9b8d3a  // 2^( 9 /32 )
+        .long 0x3f9ef532  // 2^( 10/32 )
+        .long 0x3fa27043  // 2^( 11/32 )
+        .long 0x3fa5fed7  // 2^( 12/32 )
+        .long 0x3fa9a15b  // 2^( 13/32 )
+        .long 0x3fad583f  // 2^( 14/32 )
+        .long 0x3fb123f6  // 2^( 15/32 )
+        .long 0x3fb504f3  // 2^( 16/32 )
+        .long 0x3fb8fbaf  // 2^( 17/32 )
+        .long 0x3fbd08a4  // 2^( 18/32 )
+        .long 0x3fc12c4d  // 2^( 19/32 )
+        .long 0x3fc5672a  // 2^( 20/32 )
+        .long 0x3fc9b9be  // 2^( 21/32 )
+        .long 0x3fce248c  // 2^( 22/32 )
+        .long 0x3fd2a81e  // 2^( 23/32 )
+        .long 0x3fd744fd  // 2^( 24/32 )
+        .long 0x3fdbfbb8  // 2^( 25/32 )
+        .long 0x3fe0ccdf  // 2^( 26/32 )
+        .long 0x3fe5b907  // 2^( 27/32 )
+        .long 0x3feac0c7  // 2^( 28/32 )
+        .long 0x3fefe4ba  // 2^( 29/32 )
+        .long 0x3ff5257d  // 2^( 30/32 )
+        .long 0x3ffa83b3  // 2^( 31/32 )
+        .align 32
+        .long 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78, 0x42d49a78  /* _sLg2_10*2^K   */
+        .align 32
+        .long 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000  /* _sShifter) */
+        .align 32
+        .long 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000, 0x3c1a2000  /* _sInvLg2_10hi/2^K hi (24-K-7) bits*/
+        .align 32
+        .long 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc, 0x341a84fc  /* _sInvLg2_10lo/2^K  lo bits */
+        // otherwise exp10(0) won't produce exact 1.0
+        .align 32
+        .long 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868, 0x2fecc868  /* _sPC0 */
+        .align 32
+        .long 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b, 0x40135e1b  /* _sPC1 */
+        .align 32
+        .long 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2, 0x4029a8d2  /* _sPC2 */
+        .align 32
+        .long 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f, 0x0000001f  /* _iIndexMask =(2^K-1)*/
+        //common
+        .align 32
+        .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff   /* _iAbsMask */
+        .align 32
+        .long 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818, 0x4217b818   /* _iDomainRange=-log10(max_denormal=0x007fffff) RZ */
+        .align 32
+        .type	__svml_sexp10_data_internal,@object
+        .size	__svml_sexp10_data_internal,.-__svml_sexp10_data_internal
diff --git a/sysdeps/x86_64/fpu/svml_d_exp102_core.S b/sysdeps/x86_64/fpu/svml_d_exp102_core.S
new file mode 100644
index 0000000000..157fb3b7c0
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_d_exp102_core.S
@@ -0,0 +1,29 @@
+/* Function exp10 vectorized with SSE2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_d_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVbN2v_exp10)
+WRAPPER_IMPL_SSE2 exp10
+END (_ZGVbN2v_exp10)
+
+#ifndef USE_MULTIARCH
+ libmvec_hidden_def (_ZGVbN2v_exp10)
+#endif
diff --git a/sysdeps/x86_64/fpu/svml_d_exp104_core.S b/sysdeps/x86_64/fpu/svml_d_exp104_core.S
new file mode 100644
index 0000000000..9b9d0a5d4b
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_d_exp104_core.S
@@ -0,0 +1,29 @@
+/* Function exp10 vectorized with AVX2, wrapper version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_d_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVdN4v_exp10)
+WRAPPER_IMPL_AVX _ZGVbN2v_exp10
+END (_ZGVdN4v_exp10)
+
+#ifndef USE_MULTIARCH
+ libmvec_hidden_def (_ZGVdN4v_exp10)
+#endif
diff --git a/sysdeps/x86_64/fpu/svml_d_exp104_core_avx.S b/sysdeps/x86_64/fpu/svml_d_exp104_core_avx.S
new file mode 100644
index 0000000000..1ba1a819ed
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_d_exp104_core_avx.S
@@ -0,0 +1,25 @@
+/* Function exp10 vectorized in AVX ISA as wrapper to SSE4 ISA version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_d_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVcN4v_exp10)
+WRAPPER_IMPL_AVX _ZGVbN2v_exp10
+END (_ZGVcN4v_exp10)
diff --git a/sysdeps/x86_64/fpu/svml_d_exp108_core.S b/sysdeps/x86_64/fpu/svml_d_exp108_core.S
new file mode 100644
index 0000000000..a530dc12de
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_d_exp108_core.S
@@ -0,0 +1,25 @@
+/* Function exp10 vectorized with AVX-512, wrapper to AVX2.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_d_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVeN8v_exp10)
+WRAPPER_IMPL_AVX512 _ZGVdN4v_exp10
+END (_ZGVeN8v_exp10)
diff --git a/sysdeps/x86_64/fpu/svml_s_exp10f16_core.S b/sysdeps/x86_64/fpu/svml_s_exp10f16_core.S
new file mode 100644
index 0000000000..e5043bc875
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_s_exp10f16_core.S
@@ -0,0 +1,25 @@
+/* Function exp10f vectorized with AVX-512. Wrapper to AVX2 version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_s_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVeN16v_exp10f)
+WRAPPER_IMPL_AVX512 _ZGVdN8v_exp10f
+END (_ZGVeN16v_exp10f)
diff --git a/sysdeps/x86_64/fpu/svml_s_exp10f4_core.S b/sysdeps/x86_64/fpu/svml_s_exp10f4_core.S
new file mode 100644
index 0000000000..75e6637a82
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_s_exp10f4_core.S
@@ -0,0 +1,29 @@
+/* Function exp10f vectorized with SSE2, wrapper version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_s_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVbN4v_exp10f)
+WRAPPER_IMPL_SSE2 exp10f
+END (_ZGVbN4v_exp10f)
+
+#ifndef USE_MULTIARCH
+ libmvec_hidden_def (_ZGVbN4v_exp10f)
+#endif
diff --git a/sysdeps/x86_64/fpu/svml_s_exp10f8_core.S b/sysdeps/x86_64/fpu/svml_s_exp10f8_core.S
new file mode 100644
index 0000000000..d481d2dee9
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_s_exp10f8_core.S
@@ -0,0 +1,29 @@
+/* Function exp10f vectorized with AVX2, wrapper version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_s_wrapper_impl.h"
+
+	.text
+ENTRY (_ZGVdN8v_exp10f)
+WRAPPER_IMPL_AVX _ZGVbN4v_exp10f
+END (_ZGVdN8v_exp10f)
+
+#ifndef USE_MULTIARCH
+ libmvec_hidden_def (_ZGVdN8v_exp10f)
+#endif
diff --git a/sysdeps/x86_64/fpu/svml_s_exp10f8_core_avx.S b/sysdeps/x86_64/fpu/svml_s_exp10f8_core_avx.S
new file mode 100644
index 0000000000..65944bd4d2
--- /dev/null
+++ b/sysdeps/x86_64/fpu/svml_s_exp10f8_core_avx.S
@@ -0,0 +1,25 @@
+/* Function exp10f vectorized in AVX ISA as wrapper to SSE4 ISA version.
+   Copyright (C) 2021 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_s_wrapper_impl.h"
+
+        .text
+ENTRY (_ZGVcN8v_exp10f)
+WRAPPER_IMPL_AVX _ZGVbN4v_exp10f
+END (_ZGVcN8v_exp10f)
diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx.c
new file mode 100644
index 0000000000..7cdda9895b
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx.c
@@ -0,0 +1 @@
+#include "test-double-libmvec-exp10.c"
diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx2.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx2.c
new file mode 100644
index 0000000000..7cdda9895b
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx2.c
@@ -0,0 +1 @@
+#include "test-double-libmvec-exp10.c"
diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx512f.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx512f.c
new file mode 100644
index 0000000000..7cdda9895b
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp10-avx512f.c
@@ -0,0 +1 @@
+#include "test-double-libmvec-exp10.c"
diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp10.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp10.c
new file mode 100644
index 0000000000..b1461ed85e
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp10.c
@@ -0,0 +1,3 @@
+#define LIBMVEC_TYPE double
+#define LIBMVEC_FUNC exp10
+#include "test-vector-abi-arg1.h"
diff --git a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
index 2f7172bd7b..256e8f07c9 100644
--- a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVbN2v_atan)
 VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVbN2v_asin)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVbN2vv_hypot)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVbN2v_exp2)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10), _ZGVbN2v_exp10)
 
 #define VEC_INT_TYPE __m128i
 
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
index e2d519faac..9de1dab2c2 100644
--- a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
@@ -35,6 +35,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVdN4v_atan)
 VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVdN4v_asin)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVdN4vv_hypot)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVdN4v_exp2)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10), _ZGVdN4v_exp10)
 
 #ifndef __ILP32__
 # define VEC_INT_TYPE __m256i
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
index 1ce4d8b413..43865ab099 100644
--- a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVcN4v_atan)
 VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVcN4v_asin)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVcN4vv_hypot)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVcN4v_exp2)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10), _ZGVcN4v_exp10)
 
 #define VEC_INT_TYPE __m128i
 
diff --git a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
index 6c87cec648..5dbdacf617 100644
--- a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVeN8v_atan)
 VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVeN8v_asin)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVeN8vv_hypot)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVeN8v_exp2)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10), _ZGVeN8v_exp10)
 
 #ifndef __ILP32__
 # define VEC_INT_TYPE __m512i
diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx.c
new file mode 100644
index 0000000000..be3cdaa80d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx.c
@@ -0,0 +1 @@
+#include "test-float-libmvec-exp10f.c"
diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx2.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx2.c
new file mode 100644
index 0000000000..be3cdaa80d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx2.c
@@ -0,0 +1 @@
+#include "test-float-libmvec-exp10f.c"
diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx512f.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx512f.c
new file mode 100644
index 0000000000..be3cdaa80d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f-avx512f.c
@@ -0,0 +1 @@
+#include "test-float-libmvec-exp10f.c"
diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp10f.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f.c
new file mode 100644
index 0000000000..06f447eb8d
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp10f.c
@@ -0,0 +1,3 @@
+#define LIBMVEC_TYPE float
+#define LIBMVEC_FUNC exp10f
+#include "test-vector-abi-arg1.h"
diff --git a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c
index 597d7d7598..c159c8f583 100644
--- a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVeN16v_atanf)
 VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVeN16v_asinf)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVeN16vv_hypotf)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVeN16v_exp2f)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10f), _ZGVeN16v_exp10f)
 
 #define VEC_INT_TYPE __m512i
 
diff --git a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c
index 3500eec810..c745ef744a 100644
--- a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVbN4v_atanf)
 VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVbN4v_asinf)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVbN4vv_hypotf)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVbN4v_exp2f)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10f), _ZGVbN4v_exp10f)
 
 #define VEC_INT_TYPE __m128i
 
diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c
index 921b9c65d6..c9226cf4dc 100644
--- a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c
@@ -35,6 +35,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVdN8v_atanf)
 VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVdN8v_asinf)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVdN8vv_hypotf)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVdN8v_exp2f)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10f), _ZGVdN8v_exp10f)
 
 /* Redefinition of wrapper to be compatible with _ZGVdN8vvv_sincosf.  */
 #undef VECTOR_WRAPPER_fFF
diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c
index 6cbcb57521..92970c5ace 100644
--- a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c
+++ b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c
@@ -32,6 +32,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVcN8v_atanf)
 VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVcN8v_asinf)
 VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVcN8vv_hypotf)
 VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVcN8v_exp2f)
+VECTOR_WRAPPER (WRAPPER_NAME (exp10f), _ZGVcN8v_exp10f)
 
 #define VEC_INT_TYPE __m128i