Skip to content

b12561

Pre-release
Pre-release

Choose a tag to compare

@github-actions github-actions released this 20 Nov 10:46
16f80bd

vulkan: chunked GATED_DELTA_NET implementation (#30207)

  • vulkan: add cm2 chunked GATED_DELTA_NET shader

  • fix shmem bank conflict bottleneck

  • optimize further

  • add coopmat maintenance1 variant

  • optimize for RDNA4

  • add gmem version RDNA4

  • optimize for Ampere

  • fix nvidia shmem overuse

  • fix cm2 validation errors

  • improve cm2 perf

  • improve cm1 perf

  • clean up

  • enable by default

  • show cm maintenance1 marker in device info

  • properly check for cm_m1 features

  • add tests

  • add v32 mode, improve chunked condition

  • allow other subgroup sizes

  • vulkan: specialize the GDN CM2 workgroup size

Assisted-by: Codex

  • vulkan: coopmat2 GDN implementation

  • work around Nvidia shader compiler crash

  • remove shared shader cm2 code


Co-authored-by: Jeff Bolz jbolz@nvidia.com

Website:

Attestations:

macOS/iOS:

Linux:

Android:

Windows:

openEuler:

  • DISABLED
  • openEuler x86 (310p)
  • openEuler x86 (910b, ACL Graph)
  • openEuler aarch64 (310p)
  • openEuler aarch64 (910b, ACL Graph)

UI: