Streamk v0.2 #646
Open
Streamk v0.2 #646
+886
−489
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Additionally changes locks to use uint8 instead of int32 for smaller space footprint.
neoblizz
reviewed
Sep 26, 2024
There was a problem hiding this comment.
Note, for gfx90a the load/stores with cache_modifiers do not work. Documented here: https://github.com/ROCm/triton-internal/issues/311
| rm1 = tl.max_contiguous(tl.multiple_of(rm1, BLOCK_SIZE_M), BLOCK_SIZE_M) | ||
| rn1 = tl.max_contiguous(tl.multiple_of(rn1, BLOCK_SIZE_N), BLOCK_SIZE_N) | ||
| P_ = P + pid * BLOCK_SIZE_M * BLOCK_SIZE_N + rm1[:, None] * BLOCK_SIZE_N + rn1[None, :] | ||
| tl.store(P_, acc, cache_modifier=".wt") |
There was a problem hiding this comment.
Note, for gfx90a the load/stores with cache_modifiers do not work. Documented here: https://github.com/ROCm/triton-internal/issues/311
| # todo: try use tl.load once cache modifier landed upstream | ||
| while tl.atomic_cas(locks + next_pid, 1, 1) != 1: | ||
| while (end < tile_iter_end and next_pid < NUM_SMS): | ||
| while tl.load(locks + next_pid, cache_modifier=".cv", volatile=True) != 1: |
There was a problem hiding this comment.
This also does not work in gfx90a: https://github.com/ROCm/triton-internal/issues/311
| EVEN_K: tl.constexpr, | ||
| ): | ||
| pid = tl.program_id(0) | ||
| pid = get_new_pid(pid, num_cus) | ||
| pid = (pid % 8) * (NUM_SMS // 8) + (pid // 8) |
There was a problem hiding this comment.
This is not needed for anything but gfx942, so we will actually remove this if the arch was gfx90a.
| P = torch.zeros((num_cus, block_m * block_n), device="cuda", dtype=torch.float32) | ||
| triton_output = matmul(a, b, c, P, locks, num_cus, block_m, block_n, block_k, group_m, num_warps, num_stages, | ||
| waves_per_eu, mfmaInstrSize, kpack, EVEN_K) | ||
| locks = torch.zeros((num_sms, ), device="cuda", dtype=torch.int32) |
There was a problem hiding this comment.
locks can be less than int32 type, we only need 1 byte: uint8 should work.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
streamk v0.2:
new streamk tuning script to reduce compiling and profiling time
use load/store cache modifier to reimplement spinning lock
add CI test for streamk-kernel
able to use streampipelineV2