Skip to content

Add neonfp16 f16-f32acc dwconv microkernels and scalar fallbacks. - #11102

Open
copybara-service[bot] wants to merge 1 commit into
masterfrom
test_973949321
Open

Add neonfp16 f16-f32acc dwconv microkernels and scalar fallbacks.#11102
copybara-service[bot] wants to merge 1 commit into
masterfrom
test_973949321

Conversation

@copybara-service

Copy link
Copy Markdown
Contributor

Add neonfp16 f16-f32acc dwconv microkernels and scalar fallbacks.

  • Add neonfp16 support to unipass-f16-f32acc.c.in and generate-f16-dwconv.sh.
  • Register neonfp16 microkernels in f16-dwconv-minmax.inc.
  • Update f16_dwconv_config on ARM to test neon_fp16_arith, falling back to neon_fp16 and scalar.
  • Add scalar microkernel fallbacks to all architecture blocks in init_f16_dwconv_config.
  • Update xnn_init_f16_dwconv_config to return config when minmax microkernels are configured.

- Add neonfp16 support to unipass-f16-f32acc.c.in and generate-f16-dwconv.sh.
- Register neonfp16 microkernels in f16-dwconv-minmax.inc.
- Update f16_dwconv_config on ARM to test neon_fp16_arith, falling back to neon_fp16 and scalar.
- Add scalar microkernel fallbacks to all architecture blocks in init_f16_dwconv_config.
- Update xnn_init_f16_dwconv_config to return config when minmax microkernels are configured.

PiperOrigin-RevId: 973949321
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant