vulkan: linux builds + small subgroup size fixes #11767
Conversation
github-actions
bot
added
Vulkan
| l_warptile = { 128, 128, 128, 16, device->subgroup_size * 2, 64, 2, tm_l, tn_l, tk_l, device->subgroup_size }; | ||
| m_warptile = { 128, 64, 64, 16, device->subgroup_size, 32, 2, tm_m, tn_m, tk_m, device->subgroup_size }; | ||
| s_warptile = { subgroup_size_16, 32, 32, 16, 32, 32, 2, tm_s, tn_s, tk_s, device->subgroup_size }; | ||
| l_warptile = { 128, 128, 128, 16, subgroup_size_8 * 2, 64, 2, tm_l, tn_l, tk_l, subgroup_size_8 }; |
There was a problem hiding this comment.
I imagine the coopmat path doesn't handle faking the subgroup size, maybe add an assert to that effect? Coopmat implementations probably have at least 8 invocations per subgroup, so this seems fine.
There was a problem hiding this comment.
Does our coopmat shader even work with a subgroup size of 8? We should probably find the actual limit and set up the assert based on that.
Honestly I don't know exactly why the regular mul_mat shaders break down with a subgroup size less than 8, but with the Vulkan backend becoming more and more popular I'd rather have it run slowly than fail mysteriously on those devices.
There was a problem hiding this comment.
The warptile parameters are not independent. There is probably a minimum there, coming from the hardcoded values.
Vulkan requires either the SDK or additional packages to build on Linux, so let's release official binaries so people can easily try it out.
Meanwhile our mat mul shaders don't work with subgroup sizes smaller than 8. With this fix all tests are passing even with
device->subgroup_sizeforced to 1.