diff options
| author | Joseph Myers <joseph@codesourcery.com> | 2013-12-19 13:36:10 +0000 |
|---|---|---|
| committer | Joseph Myers <joseph@codesourcery.com> | 2013-12-19 13:36:10 +0000 |
| commit | f88acd39da2a509081e541b84ecbf204ef20f9e8 (patch) | |
| tree | 383c0bcce4b3e74c9150d01550e7c412e47981bd /sysdeps/x86_64/fpu/e_expl.S | |
| parent | c688b4196014e0162a1ff11120f6c9516be0c6cb (diff) | |
| download | glibc-f88acd39da2a509081e541b84ecbf204ef20f9e8.tar.gz glibc-f88acd39da2a509081e541b84ecbf204ef20f9e8.tar.xz glibc-f88acd39da2a509081e541b84ecbf204ef20f9e8.zip | |
Fix x86/x86_64 expm1 inaccuracy near 0 in directed rounding modes (bug 16293).
Bug 16293 is inaccuracy of x86/x86_64 versions of expm1, near 0 in directed rounding modes, that arises from frndint rounding the exponent to 1 or -1 instead of 0, resulting in large cancellation error. This inaccuracy in turn affects other functions such as sinh that use expm1. This patch fixes the problem by setting round-to-nearest mode temporarily around the affected calls to frndint. I don't think this is needed for other uses of frndint, such as in exp itself, as only for expm1 is the cancellation error significant. Tested x86_64 and x86 and ulps updated accordingly. * sysdeps/i386/fpu/e_expl.S (IEEE754_EXPL) [USE_AS_EXPM1L]: Set round-to-nearest mode when using frndint. * sysdeps/i386/fpu/s_expm1.S (__expm1): Likewise. * sysdeps/i386/fpu/s_expm1f.S (__expm1f): Likewise. * sysdeps/x86_64/fpu/e_expl.S (IEEE754_EXPL) [USE_AS_EXPM1L]: Likewise. * math/auto-libm-test-in: Add more tests of expm1. Do not expect sinh test to fail. * math/auto-libm-test-out: Regenerated. * math/libm-test.inc (TEST_COND_x86_64): Remove macro. (TEST_COND_x86): Likewise. (expm1_tonearest_test_data): New array. (expm1_test_tonearest): New function. (expm1_towardzero_test_data): New array. (expm1_test_towardzero): New function. (expm1_downward_test_data): New array. (expm1_test_downward): New function. (expm1_upward_test_data): New array. (expm1_test_upward): New function. (main): Run the new test functions. * sysdeps/i386/fpu/libm-test-ulps: Update. * sysdeps/x86_64/fpu/libm-test-ulps: Likewise.
Diffstat (limited to 'sysdeps/x86_64/fpu/e_expl.S')
| -rw-r--r-- | sysdeps/x86_64/fpu/e_expl.S | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/sysdeps/x86_64/fpu/e_expl.S b/sysdeps/x86_64/fpu/e_expl.S index a919780390..1c21f03ddc 100644 --- a/sysdeps/x86_64/fpu/e_expl.S +++ b/sysdeps/x86_64/fpu/e_expl.S @@ -127,9 +127,20 @@ ENTRY(IEEE754_EXPL) #endif 3: FLDLOG /* 1 log2(base) */ fmul %st(1), %st /* 1 x log2(base) */ +#ifdef USE_AS_EXPM1L + /* Set round-to-nearest temporarily. */ + fstcw -4(%rsp) + movl $0xf3ff, %edx + andl -4(%rsp), %edx + movl %edx, -8(%rsp) + fldcw -8(%rsp) +#endif frndint /* 1 i */ fld %st(1) /* 2 x */ frndint /* 2 xi */ +#ifdef USE_AS_EXPM1L + fldcw -4(%rsp) +#endif fld %st(1) /* 3 i */ fldt MO(c0) /* 4 c0 */ fld %st(2) /* 5 xi */ |