Skip to content

Pull requests: google/XNNPACK

New pull request New
156 Open 7,820 Closed
156 Open 7,820 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

EXPECT_EQ -> ASSERT_EQ to avoid test log explosion
#8254 opened Apr 11, 2025 by copybara-service bot Loading…
Re-enable more fp16 kernels after workaround
#8253 opened Apr 11, 2025 by copybara-service bot Loading…
Build everything in //test/... and //bench/...
#8250 opened Apr 11, 2025 by copybara-service bot Loading…
Expand the description of gemm microkernels. NFC
#8247 opened Apr 10, 2025 by copybara-service bot Loading…
Remove gemminc kernels
#8246 opened Apr 10, 2025 by copybara-service bot Loading…
Copybara import of the project:
#8245 opened Apr 10, 2025 by copybara-service bot Loading…
QS8 gemm use V73 HVX for FP32 quantization
#8244 opened Apr 10, 2025 by copybara-service bot Loading…
QS8 gemm use V73 int to float and multiply for quantization
#8242 opened Apr 10, 2025 by copybara-service bot Loading…
QS8 gemm use V73 int to float instruction for quantization
#8241 opened Apr 10, 2025 by copybara-service bot Loading…
Refactor xnn_log_* to avoid materializing args if the required log level is not set.
#8239 opened Apr 9, 2025 by copybara-service bot Loading…
Clean up the subgraph benchmarks a bit:
#8238 opened Apr 9, 2025 by copybara-service bot Loading…
Strided slice can handle 0 sized inputs & outputs
#8236 opened Apr 9, 2025 by copybara-service bot Loading…
Disable xnn_x32_packw_gemm_goi_ukernel_x32__avx512f_u4_prfm which crashes sometimes
#8225 opened Apr 8, 2025 by copybara-service bot Loading…
Internal change only.
#8224 opened Apr 8, 2025 by copybara-service bot Loading…
QS8-IGEMM support NR for Hexagon HVX
#8219 opened Apr 7, 2025 by copybara-service bot Loading…
QS8-GEMM support NR for Hexagon HVX
#8218 opened Apr 7, 2025 by copybara-service bot Loading…
Copybara import of the project:
#8213 opened Apr 4, 2025 by copybara-service bot Loading…
SME2 IGEMM QS8 integration prototype
#8206 opened Apr 4, 2025 by doviethoa-at-work • Draft
Add benchmarks for both the characteristic and actual L1/L2 cache sizes.
#8202 opened Apr 4, 2025 by copybara-service bot Loading…
Enable QS8-GEMM for HVX
#8194 opened Apr 3, 2025 by copybara-service bot Loading…
Remove scaled dot product attention operator
#8188 opened Apr 3, 2025 by copybara-service bot Loading…
GEMM HVX use qfloat accumulators
#8168 opened Mar 31, 2025 by copybara-service bot Loading…
Reformat C too
#8155 opened Mar 31, 2025 by copybara-service bot Loading…
[RVV] Add qu8-gemm/qu8-igemm kernels for rvv
#8105 opened Mar 23, 2025 by ken-unger Loading…
3
Support Sparse Inference for RVV
#8081 opened Mar 20, 2025 by NonerKao Loading…
8
ProTip! Add no:assignee to see everything that’s not assigned.