Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Add AVX optimized versions for some x86-64 math functions | Ulrich Drepper | 2011-10-25 | 28 | -34/+294 |
| | |||||
* | Optimize x86-64 SSE4.2+ strcmp a bit more | Ulrich Drepper | 2011-10-25 | 1 | -175/+130 |
| | |||||
* | Use VEX encoding in inline math functions on x86-64 when possible | Ulrich Drepper | 2011-10-25 | 1 | -48/+138 |
| | |||||
* | Cleanup FMA4 patch | Ulrich Drepper | 2011-10-25 | 22 | -1/+26 |
| | | | | | Move the FMA4 code into its own section. Avoid some of the duplication of data resulting from the double use of source files. | ||||
* | Better DLA_FMS | Ulrich Drepper | 2011-10-24 | 1 | -1/+8 |
| | | | | | It's better to use __builtin_fma if it works. Use it for gcc 4.6 and higher. Move the x86-64 dla.h to the correct place. | ||||
* | Use inline asm for DLA_FMS because of broken old compilers | Ulrich Drepper | 2011-10-24 | 1 | -2/+6 |
| | |||||
* | Optimize accurate 64-bit routines for FMA4 on x86-64 | Ulrich Drepper | 2011-10-24 | 31 | -0/+328 |
| | |||||
* | Fix WS | Ulrich Drepper | 2011-10-23 | 1 | -1/+1 |
| | |||||
* | Optimized strnlen and wcscmp for x86-64 | Liubov Dmitrieva | 2011-10-23 | 8 | -317/+987 |
| | |||||
* | Fix WS | Ulrich Drepper | 2011-10-23 | 1 | -5/+5 |
| | |||||
* | Fix signedness in wcscmp comparison | Liubov Dmitrieva | 2011-10-23 | 1 | -45/+64 |
| | |||||
* | Clean up FMA use | Ulrich Drepper | 2011-10-23 | 1 | -6/+3 |
| | | | | | The macro's name should reflect that subtraction is being done. And use __builtin_fma, it seems to work after all. | ||||
* | Clean up last dla.h change | Ulrich Drepper | 2011-10-23 | 1 | -0/+9 |
| | |||||
* | Fix typo in last change | Ulrich Drepper | 2011-10-22 | 1 | -1/+1 |
| | |||||
* | Update ULPs for x86-64 | Ulrich Drepper | 2011-10-22 | 1 | -4/+12 |
| | |||||
* | Fix compilation problems in x86-64 init-arch | Ulrich Drepper | 2011-10-21 | 1 | -1/+2 |
| | |||||
* | Don't use NULL in last s_fma{,f} change | Ulrich Drepper | 2011-10-21 | 2 | -2/+2 |
| | |||||
* | Check for FMA4 support and generate appropriate fma functions | Ulrich Drepper | 2011-10-20 | 4 | -7/+49 |
| | |||||
* | Move fma routines to right place | Ulrich Drepper | 2011-10-20 | 2 | -0/+0 |
| | |||||
* | Optimize x86-64 rawmemchr and add test | Ulrich Drepper | 2011-10-19 | 1 | -4/+3 |
| | |||||
* | Add optimized str{,n}casecmp for AVX on x86-64 | Ulrich Drepper | 2011-10-19 | 3 | -1798/+1870 |
| | |||||
* | Fix PLT use for feraiseexcept on x86_64 | Andreas Schwab | 2011-10-19 | 1 | -1/+1 |
| | |||||
* | Use new internal libc_fe* interfaces in more functions | Ulrich Drepper | 2011-10-18 | 1 | -5/+30 |
| | |||||
* | Provide combined internal feholdexcept/fesetround interface | Ulrich Drepper | 2011-10-18 | 1 | -1/+13 |
| | |||||
* | Pretty print last change to x86-64 mathinline.h | Ulrich Drepper | 2011-10-18 | 1 | -2/+2 |
| | |||||
* | Add optimized nearbyint{,f} for x86-64 | Ulrich Drepper | 2011-10-18 | 6 | -1/+105 |
| | |||||
* | Start optimizing the use of the fenv interfaces in libm itself | Ulrich Drepper | 2011-10-18 | 1 | -0/+42 |
| | |||||
* | Fix last change | Andreas Schwab | 2011-10-18 | 1 | -2/+3 |
| | |||||
* | Fix linkage conflict with feraiseexcept | Andreas Schwab | 2011-10-18 | 1 | -6/+10 |
| | |||||
* | Relax asm requirements for recently added x86-64 math interfaces | Ulrich Drepper | 2011-10-17 | 2 | -10/+10 |
| | |||||
* | Makr x86-64 math_private.h more robust | Ulrich Drepper | 2011-10-17 | 1 | -10/+10 |
| | |||||
* | Provide internal optimizations on x86-64 with SSE4.1 | Ulrich Drepper | 2011-10-17 | 6 | -12/+51 |
| | | | | | | | Provide macros so that the internal users can, if possible, directly use the new instructions. Also fix up the mathinline.h header when compiling with SSE4.1 enabled. | ||||
* | Fix last x86-64 mathinline change | Ulrich Drepper | 2011-10-17 | 1 | -2/+2 |
| | | | | Use correct function names. | ||||
* | Use rounds{s,d} for x86 rint, ceil, floor | Ulrich Drepper | 2011-10-16 | 15 | -10/+339 |
| | |||||
* | Fix WS | Ulrich Drepper | 2011-10-15 | 1 | -4/+4 |
| | |||||
* | Optimized memcmp and wmemcmp for x86-64 and x86-32 | Liubov Dmitrieva | 2011-10-15 | 8 | -29/+2242 |
| | |||||
* | Fix lost feraiseexcept symbol | Andreas Schwab | 2011-10-14 | 1 | -1/+2 |
| | |||||
* | Fix PLT use in feupdateenv on x86_64 | Andreas Schwab | 2011-10-13 | 2 | -5/+5 |
| | |||||
* | Check for zero size in memrchr for x86_64 | Andreas Schwab | 2011-10-13 | 1 | -0/+3 |
| | |||||
* | Optimize libm | Ulrich Drepper | 2011-10-12 | 13 | -47/+152 |
| | | | | | libm is now somewhat integrated with gcc's -ffinite-math-only option and lots of the wrapper functions have been optimized. | ||||
* | Optimize use of isnan, isinf, finite | Ulrich Drepper | 2011-10-08 | 1 | -0/+21 |
| | |||||
* | Fix whitespace | Ulrich Drepper | 2011-10-07 | 1 | -1/+1 |
| | |||||
* | Improve 64 bit memchr, memrchr, rawmemchr with SSE2 | Liubov Dmitrieva | 2011-10-07 | 4 | -59/+842 |
| | |||||
* | Don't call ifunc functions in trace mode | Andreas Schwab | 2011-10-05 | 1 | -4/+7 |
| | |||||
* | Fix parse error in bits/mathinline.h with --std=c99 | Andreas Schwab | 2011-09-26 | 1 | -8/+8 |
| | |||||
* | Add fmax and fmin inlines for x86-64 | Ulrich Drepper | 2011-09-15 | 1 | -5/+41 |
| | |||||
* | Use correct section to allow merging | Ulrich Drepper | 2011-09-14 | 1 | -2/+2 |
| | |||||
* | Optimized lrint and llrint for x86-64 | Ulrich Drepper | 2011-09-14 | 1 | -10/+46 |
| | |||||
* | Avoid macro clash between <sys/select.h> and <linux/posix_types.h> | Andreas Schwab | 2011-09-13 | 1 | -3/+3 |
| | |||||
* | Remove --wth-tls option, TLS support is required | Ulrich Drepper | 2011-09-11 | 2 | -9/+4 |
| |