Skip to content

Pull requests: google/XNNPACK

New pull request New
127 Open 7,219 Closed
127 Open 7,219 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Don't use a fixed-size array for Slinky input/output values, as degenerate cases may need larger values; instead, allocate the space dynamically.
#7599 opened Dec 12, 2024 by copybara-service bot Loading…
Add f32 fgemv SME kleidi kernel
#7596 opened Dec 11, 2024 by copybara-service bot Loading…
1
Sparsity detection should distinguish between fp32 and fp16.
#7595 opened Dec 11, 2024 by copybara-service bot Loading…
Replace vld1_dup_X with vdup_n_X
#7592 opened Dec 11, 2024 by copybara-service bot Loading…
Hook up batch-matrix-multiply to use qp8-based microkernels.
#7587 opened Dec 10, 2024 by copybara-service bot Loading…
Add void statement to empty else clause.
#7583 opened Dec 10, 2024 by copybara-service bot Loading…
Integreate KleidiAI int4 SME into XNNPack.
#7582 opened Dec 10, 2024 by copybara-service bot Loading…
QC4W GEMM test/benchmark use packw microkernels
#7581 opened Dec 10, 2024 by copybara-service bot Loading…
Prototype XNNPack gemm compiler.
#7569 opened Dec 6, 2024 by copybara-service bot Loading…
[RFC] Multithread packing
#7545 opened Dec 2, 2024 by mcr229 Loading…
10
[Packing Refactor] Move all Blockwise Packing to pack_weights_and_bias
#7544 opened Dec 2, 2024 by mcr229 Loading…
6
Adding Tiled Scalar Packing Kernels for QB4W
#7521 opened Nov 27, 2024 by mcr229 Loading…
3
Integration of Kleidi F16 SME kernels
#7520 opened Nov 27, 2024 by copybara-service bot Loading…
1
Flip the order in which the GEMM tiles are computed, i.e. instead of traversing the tiles in n direction first, traverse them in m direction, and then in n direction.
#7513 opened Nov 26, 2024 by copybara-service bot Loading…
F32-IGEMM AVX512 generate up to 16x64
#7493 opened Nov 21, 2024 by copybara-service bot Loading…
implement qs8 x8c4 pack using avxvnni
#7462 opened Nov 14, 2024 by kylo5aby Loading…
17
Replace lut yaml with table header
#7446 opened Nov 8, 2024 by pratham-mcw Loading…
4
Replace conv-hwc yaml with table header
#7445 opened Nov 8, 2024 by RahulSundarMCW Loading…
5
Replace lut32norm yaml with table header
#7444 opened Nov 8, 2024 by RahulSundarMCW Loading…
3
VMulCAddC-Replaced yaml files with header table
#7411 opened Nov 4, 2024 by nitheshsrikanth-mcw Loading…
16
Support kblock for x32 GIO packing
#7405 opened Nov 4, 2024 by xujuntwt95329 Loading…
Support RVV f32-dwconv-unipass
#7402 opened Nov 4, 2024 by yhng3010 Loading…
1
[WIP] HVX x32 transposec
#7395 opened Oct 31, 2024 by ejparkqc Loading…
Packing use inlined for loops to copy and set bias and weights
#7379 opened Oct 29, 2024 by copybara-service bot Loading…
Fix build error
#7376 opened Oct 29, 2024 by RahulSundarMCW Loading…
1
ProTip! What’s not been updated in a month: updated:<2024-11-12.