Add P7/8/9/10 divide/modulo doubleword operations to libraries. #188
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
POWER10 added vector divide/divide-extended/modulo instructions. This patch adds static and IFUNC enabled dynamic callable functions to the PVECLIB executable library. This is restricted to the subset of operations that have significant code size or runtime. This is basically the divide/modulo operations; vec_divdud, vec_diveud, vec_divqud, vec_divud, vec_moddud, vec_modud. * configure: Autoreconf regenerated. * Makefile.in: Automake regenerated. * aclocal.m4: Automake regenerated. * src/Makefile.am: Add vec_int64_runtime.c to dependency for PWR library builds. Add Automake regenerated. * src/Makefile.in: Automake regenerated. * src/pveclib/vec_int64_ppc.h: Update copyright dates. Various doxygen text updates for POWER10. (vec_mrgahd, vec_mrgald): New Forward function prototypes. (vec_divdud, vec_diveud, vec_divqud): New extern for dynamic library functions. (vec_divqud_inline): Doxygen text note about usage in quadword long division implementations. (vec_divqud_inline [(_ARCH_PWR10) && (__GNUC__ >= 12)]): POWER10 specific implementation using intrinsics. (vec_divud, vec_moddud, vec_modud): New extern for dynamic library functions. * src/testsuite/arith128_test_i64.c [__VEC_PWR_IMP()]: Define extern and map unit tests to library implementations. * src/testsuite/vec_int64_dummy.c: Update copyright dates. (test_vec_divqud [(_ARCH_PWR10) && (__GNUC__ >= 12)]): POWER10 specific implementation using intrinsics. * src/testsuite/vec_perf_i128.c [__VEC_PWR_IMP()]: Define extern and map timed tests to library implementations. * src/testsuite/vec_pwr10_dummy.c (test_vec_rlqi_64_PWR10, test_vec_slqi_64_PWR10, test_vec_sraqi_64_PWR10, test_vec_srqi_64_PWR10): New compile tests for doubleword shift/rotates special case. (test_vec_divqud_PWR10): New compile test. POWER10 specific implementation using intrinsics. * src/vec_int64_runtime.c: New file. * src/vec_runtime_DYN.c: Include <pveclib/vec_int64_ppc.h>. [VEC_INT64_LIB_LIST(_TARGET)]: Define list of int64 target specific functions. [PVECLIB_DISABLE_POWER7]: Add VEC_INT64_LIB_LIST (_PWR7). Add VEC_INT64_LIB_LIST (_PWR8). [PVECLIB_DISABLE_POWER9]: Add VEC_INT64_LIB_LIST (_PWR9). [PVECLIB_DISABLE_POWER10]: Add VEC_INT64_LIB_LIST (_PWR10). [VEC_RESOLVER_2]: vec_diveud, vec_divqud, vec_divud, vec_modud. [VEC_RESOLVER_3]: vec_divdud, vec_moddud. * src/vec_runtime_PWR10.c: Include vec_int64_runtime.c. * src/vec_runtime_PWR7.c: Include vec_int64_runtime.c. * src/vec_runtime_PWR8.c: Include vec_int64_runtime.c. * src/vec_runtime_PWR9.c: Include vec_int64_runtime.c. Signed-off-by: Steven Munroe <[email protected]>
|
Tulio please a look at this and issue #189. This patch is not bad (size wise) but the int128 equivalent changes will be more complicated. |
|
Seems Tulio is busy. I'll merge this so I can push quadword divide and work on Float128 divide round to odd. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
POWER10 added vector divide/divide-extended/modulo instructions. This patch adds static and IFUNC enabled dynamic callable functions to the PVECLIB executable library. This is restricted to the subset of operations that have significant code size or runtime. This is basically the divide/modulo operations; vec_divdud, vec_diveud, vec_divqud, vec_divud, vec_moddud, vec_modud.