From d0b6a0b1ab49b5958015f31364c5ad73d9cd03eb Mon Sep 17 00:00:00 2001 From: Darren Wihandi <65404740+fairywreath@users.noreply.github.com> Date: Tue, 15 Apr 2025 15:57:45 -0600 Subject: Add cooperative matrix 1 support (#6565) * initial wip for spirv * working tiled example * clean up store and load * minor fixes * fix loadAny name * add initial tests, including broken/unimplemented intrinsics * fix subscript * run tests at 16x16, remove not supported arithmetic tests * minor fixups on implementation * rename CoopMatMatrixUse * Update tests to pass validation layers locally * Add mat-mul-add test and minor fixes * Add more tests * Remove dead code * Add coopMatLoad function and tests, enforce constexpr for matrix layout * Use getVectorOrCoopMatrixElementType in place of getVectorElementType --- docs/user-guide/a3-02-reference-capability-atoms.md | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'docs/user-guide') diff --git a/docs/user-guide/a3-02-reference-capability-atoms.md b/docs/user-guide/a3-02-reference-capability-atoms.md index 1cb7f5bd5..d72a2768f 100644 --- a/docs/user-guide/a3-02-reference-capability-atoms.md +++ b/docs/user-guide/a3-02-reference-capability-atoms.md @@ -424,6 +424,9 @@ Extensions `SPV_NV_cooperative_vector` > Represents the SPIR-V extension for SPV_NV_cooperative_vector. +`SPV_KHR_cooperative_matrix` +> Represents the SPIR-V extension for SPV_KHR_cooperative_matrix. + `spvAtomicFloat32AddEXT` > Represents the SPIR-V capability for atomic float 32 add operations. @@ -535,6 +538,9 @@ Extensions `spvCooperativeVectorTrainingNV` > Represents the SPIR-V capability for cooperative vector training +`spvCooperativeMatrixKHR` +> Represents the SPIR-V capability for cooperative matrices + `spvMaximalReconvergenceKHR` > Represents the SPIR-V capability for maximal reconvergence. @@ -1206,6 +1212,9 @@ Other ---------------------- *Capabilities that may be deprecated* +`cooperative_matrix` +> Capabilities needed to use cooperative matrices + `SPIRV_1_0` > Use `spirv_1_0` instead -- cgit v1.2.3