AuburnSounds / intel-intrinsics

The Dlang SIMD library

https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#techs=MMX,SSE,SSE2,SSE3,SSSE3,SSE4_1

AuburnSounds/intel-intrinsics Issues

Could _mm_pow_s/ps support 0.0f input?
Updated a month ago1
AVX2
Updated a month ago2
Undeprecate the _mm_srl_epi32-style SSE2 function
Closed a month ago2
_mm_pow_ps with small numbers
Updated a month ago
Doing nothing vs going pragma(inline, true) vs function literals
Updated 3 months ago
Enabling AVX vectors in DMD D_SIMD
Updated 6 months ago1
Major speed regression
Closed 6 months ago1
Share assembly blocks between DMD and LDC when needed
Closed 9 months ago2
Evaluate faster _mm_pow_ss and _mm_exp_ss
Updated 9 months ago
DMD 2.102 bug on CI
Updated 10 months ago5
Forbid _mm_realloc with zero size, and _mm_realloc_discard
Closed 10 months ago1
LDC build times woes
Closed a year ago9
Latest LDC + -b unittest-release-inst -a x86 => crash in CI
Updated a year ago1
Fix _mm_madd_epi16 on arm64
Closed a year ago
Remaining ARM64 paths in avxintrin.d
Closed a year ago
Builtins not found with the LDC-1.32.0 compiler
Closed a year ago2
ARM64: Use integer widen builtins
Closed a year ago3
Revive arm32 support officially
Updated a year ago
Cannot compile with the latest version of ldc2(1.32.0)
Closed a year ago4
Using sdfmt?
Closed a year ago
Undefined identifier _m256 in avxintrin
Closed a year ago1
AVX emulated with ARM64, some items remaining
Closed a year ago2
Different speed results on macOS(M1) and Linux(X86_64)
Closed a year ago4
Add a bonus intrinsics for int4 x int4
Closed a year ago1
Function taking __m256i
Closed a year ago
is big realloc really better than free + malloc
Closed a year ago
Separate and optimize scanline conversion function
Closed a year ago1
AVX intrinsics (not AVX2)
Closed a year ago1
Speed-up float comparison for GDC and DMD
Updated a year ago
_mm256_movemask_ps failure on arm64
Closed a year ago2
SSE2 (line < 431)
Closed a year ago
How would be a way to have divergent code path for say using AVX?
Closed a year ago3
Fix semantics of masked load/store
Closed a year ago1
Add support for SSE2 _mm_storeu_si64
Closed a year ago2
Fake vectors should be aligned... when possible
Updated 2 years ago
Investigate cost of failed inlined intrinsics that do not turn to a good sequence
Closed 2 years ago1
Remove portable shufflevector
Closed 2 years ago2
Use __builtin_ia32_addsubpd instead of implicit
Closed 2 years ago4
Faster set/setr
Closed 2 years ago1
excessive loadUnaligned usage with LDC
Closed 2 years ago1
Fix the remaining buggy GDC paths
Closed 2 years ago3
Fix ssemathfun port
Closed 2 years ago1
Latest release fails on Buildkite
Closed 2 years ago6
Unittest failure blocks DMD buildkite
Closed 2 years ago3
Fix CI with -mcpu=native
Closed 2 years ago3
Idea: UDAs to automatically do inlining with LDC
Updated 2 years ago6
_mm_cmpistrX has no need to use _mm_cmpestrX
Closed 2 years ago
Bug in greaterOrEqualMask
Closed 2 years ago1
Weird code for _mm_cvtepi32_ps on haswell/skylake
Closed 2 years ago4
Bench pragma(inline) on real project (debug performance)
Closed 2 years ago6