Jack Pappas jack-pappas

  • Philadelphia, United States
  • Joined on

Organizations

@tidepowerd
@jack-pappas
  • @jack-pappas 9e6f49f
    Make sure the environment section is a mapping.
@jack-pappas
  • @jack-pappas 218a47e
    CMake build fixes for MinGW.
jack-pappas opened pull request Blosc/c-blosc#136
@jack-pappas
AVX2 shuffle support for types larger than 16 bytes
2 commits with 207 additions and 18 deletions
@jack-pappas
  • @jack-pappas bc2acf0
    Use 'set' intrinsics to create shuffle masks.
jack-pappas opened pull request Blosc/c-blosc#135
@jack-pappas
Implemented peeled AVX2 shuffle
1 commit with 282 additions and 228 deletions
jack-pappas deleted branch large-type-sse2-shuffle at jack-pappas/c-blosc
jack-pappas deleted branch large-type-shuffle-tests at jack-pappas/c-blosc
jack-pappas deleted branch tiled-sse2-shuffle at jack-pappas/c-blosc
jack-pappas deleted branch tiled-sse2-shuffle-tests at jack-pappas/c-blosc
jack-pappas deleted branch tiled-shuffle-preliminary at jack-pappas/c-blosc
jack-pappas deleted branch tiled-peeled-sse2-shuffle at jack-pappas/c-blosc
@jack-pappas
jack-pappas deleted branch peeled-sse2-shuffle at jack-pappas/c-blosc
jack-pappas commented on pull request Blosc/c-blosc#129
@jack-pappas

Yes, I'm planning to port the changes made to the SSE2 implementation in this PR and #133 to the AVX2 implementation as well.

jack-pappas opened pull request Blosc/c-blosc#133
@jack-pappas
Added SSE2 shuffle/unshuffle routines for types larger than 16 bytes
1 commit with 150 additions and 11 deletions
@jack-pappas
@jack-pappas
Add optional support for zlib-ng
@jack-pappas
Enable unit tests based on host CPU feature detection
jack-pappas commented on pull request Blosc/c-blosc#129
@jack-pappas

@FrancescAlted I've rebased the changes onto master (after your merge of #130) and squashed them into a single commit, so this can be merged whenev…

@jack-pappas
  • @jack-pappas 4081fda
    Implement SSE2 shuffle support for buffers containing a number of ele…
@jack-pappas
  • @jack-pappas 4d82a62
    Implement SSE2 shuffle support for buffers containing a number of ele…
@jack-pappas
jack-pappas commented on pull request Blosc/c-blosc#129
@jack-pappas

I tried replacing the _mm_loadu_si128 calls with _mm_load_si128 and the _mm_storeu_si128 calls with _mm_store_si128. After doing so, I compiled blo…

jack-pappas commented on pull request Blosc/c-blosc#129
@jack-pappas

@FrancescAlted I'm happy to make the change either way you want. I can change the loads/stores back to aligned to preserve the current behavior for…