Skip to content

Pull requests: google/XNNPACK

New pull request New
257 Open 10,244 Closed
257 Open 10,244 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[XLA:CPU] Add vectorized sin and cos intrinsic kernels for codegen.
#10821 opened Jul 23, 2026 by copybara-service Bot Loading…
Implement tan(x)
#10820 opened Jul 23, 2026 by copybara-service Bot Loading…
Reduce sin polynomial degree
#10819 opened Jul 23, 2026 by copybara-service Bot Loading…
Add Qd8 bf16 qb4w subgraph for fully connected
#10818 opened Jul 23, 2026 by JakeStevens Contributor Loading…
Rename ARM FP8 inline helpers cpuinfo_has_arm_f8dot and cpuinfo_has_arm_f8mm to NEON naming conventions.
#10817 opened Jul 23, 2026 by copybara-service Bot Loading…
Fix catastrophic cancellation in f32 and f16 vcos microkernels near x = pi + 2*pi*n by replacing floating-point theta offsets with integer quadrant range reduction and symmetry/sign folding, ensuring exact numerical equivalence with YNNPACK.
#10810 opened Jul 21, 2026 by copybara-service Bot Loading…
1
Reuse leading gaps in memory planner
#10801 opened Jul 21, 2026 by teerthsharma Loading…
1
Add a schedule dumping function for debugging YNN scheduler decisions.
#10799 opened Jul 21, 2026 by copybara-service Bot Loading…
Copybara import of the project:
#10793 opened Jul 20, 2026 by copybara-service Bot Loading…
reallocate deconv igemm indirection buffer on output-size change
#10791 opened Jul 20, 2026 by aizu-m Contributor Loading…
Fix qb4 packw bias tail in neon
#10787 opened Jul 18, 2026 by JakeStevens Contributor Loading…
7
Use cpuinfo support for ARM fp8
#10785 opened Jul 17, 2026 by copybara-service Bot Loading…
Implement custom op parser registry in TFLite flatbuffer parser.
#10784 opened Jul 17, 2026 by copybara-service Bot Loading…
Make TFLite flatbuffer parser backend-neutral.
#10783 opened Jul 17, 2026 by copybara-service Bot Loading…
perf: optimize decode executor to reduce host-side overhead
#10778 opened Jul 17, 2026 by copybara-service Bot Loading…
[RVV] Add RVV x32-unpool microkernel and benchmark
#10773 opened Jul 17, 2026 by leduclean • Draft
1
litert/tensor/arithmetic.h: bounds-check multiples length in Tile()
#10772 opened Jul 17, 2026 by destro4evr-rgb Contributor Loading…
[RVV] add rvv f32 kernels for CHW ibilinear
#10771 opened Jul 17, 2026 by leduclean • Draft
1
Add f32-ibilinear-chw microkernel benchmark
#10770 opened Jul 17, 2026 by leduclean • Draft
1
Add the golden_parity global configuration for architecture-neutral golden tests.
#10747 opened Jul 14, 2026 by copybara-service Bot Loading…
Inject explicit architecture-neutral arithmetic flags and golden parity defines directly into the ARCore unit test BuildableUnit.
#10745 opened Jul 14, 2026 by copybara-service Bot Loading…
Support enforcing architecture-neutral baseline data cache sizes for Eigen GEMM operations.
#10744 opened Jul 14, 2026 by copybara-service Bot Loading…
Fix support for explicitly disabling x86 SIMD and matrix multiplication architecture extensions.
#10742 opened Jul 14, 2026 by copybara-service Bot Loading…
Implement an XNNPack runner for Gemma 4.
#10738 opened Jul 13, 2026 by copybara-service Bot Loading…
Implement the Gemma 4 graph build.
#10737 opened Jul 13, 2026 by copybara-service Bot Loading…
ProTip! Mix and match filters to narrow down what you’re looking for.