Restructure MoE and Add MoE prepare input kernels #29
Conversation
This file contains hidden or 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
- restructure moe kernels folder - add prepare moe inputs kerel Signed-off-by: kareem <[email protected]>
kareemshaik80
changed the title
kareemshaik80
commented
Oct 30, 2025
Signed-off-by: kareem <[email protected]>
kareemshaik80
changed the title
Signed-off-by: kareem <[email protected]>
Signed-off-by: kareem <[email protected]>
Signed-off-by: kareem <[email protected]>
Signed-off-by: kareem <[email protected]>
Signed-off-by: Shaik, Kareem M <[email protected]>
kareemshaik80
changed the title
kareemshaik80
changed the title
Signed-off-by: kareem <[email protected]>
Signed-off-by: kareem <[email protected]>
airMeng
reviewed
Nov 11, 2025
Collaborator
There was a problem hiding this comment.
activations not only serve for MoE, better leave it unchanged.
Beside I'd prefer only put customized cutlass code under src/sycl/kernels/, and leave pure SYCL code outside
Comment on lines
+6
to
+10
|
|
||
| @pytest.mark.parametrize("num_tokens", [5, 16, 128]) | ||
| @pytest.mark.parametrize("num_experts", [4, 8, 32]) | ||
| @pytest.mark.parametrize("top_k", [2]) | ||
| @pytest.mark.parametrize("hidden_dims", [16, 32, 64]) |
Collaborator
There was a problem hiding this comment.
airMeng
requested changes
Nov 11, 2025
Comment on lines
+173
to
+187
| void operator()(sycl::nd_item<1> item) const { | ||
| int32_t tot_offset = 0; | ||
| int32_t tot_rounded_offset = 0; | ||
| expert_offsets_[0] = 0; | ||
| blockscale_offsets_[0] = 0; | ||
| for (int i = 0; i < num_experts_; ++i) { | ||
| atomic_buffer_[i] = tot_offset; | ||
| int num_tokens = problem_sizes1_[i * 3]; | ||
| int rounded_num_tokens = (num_tokens + (block_size - 1)) / block_size * block_size; | ||
| tot_offset += num_tokens; | ||
| tot_rounded_offset += rounded_num_tokens; | ||
| expert_offsets_[i + 1] = tot_offset; | ||
| blockscale_offsets_[i + 1] = tot_rounded_offset; | ||
| } | ||
| } |
Collaborator
There was a problem hiding this comment.
The function is purely sequential? Try to parallelize it using sycl::exclusive_scan, see https://github.com/intel/llvm/blob/4474e85c51c1c3153af9938164391d1e836cfff4/sycl/doc/extensions/removed/sycl_ext_oneapi_group_algorithms.asciidoc?plain=1#L75
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Uh oh!
There was an error while loading. Please reload this page.