Add challenge 89: Flash Attention Forward (Medium)#232
Open
claude[bot] wants to merge 1 commit intomainfrom
Open
Add challenge 89: Flash Attention Forward (Medium)#232claude[bot] wants to merge 1 commit intomainfrom
claude[bot] wants to merge 1 commit intomainfrom
Conversation
Introduces a Flash Attention forward-pass challenge teaching the online-softmax tiling algorithm that avoids materializing the full seq_len × seq_len attention matrix in global memory. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
seq_len × seq_lenattention matrix in global memoryQ,K,Veach[num_heads, seq_len, head_dim]; output same shape--action submiton NVIDIA Tesla T4: all functional and performance tests passnum_heads=16,seq_len=4,096,head_dim=64Why this challenge is distinct from existing attention challenges
Test plan
.cu,.pytorch.py,.triton.py,.jax.py,.cute.py,.mojo)challenge.htmlpasses all checklist items (starts with<p>, correct<h2>sections, example matchesgenerate_example_test(), performance bullet)challenge.pyhas all 6 methods;generate_functional_test()returns 10 test cases covering edge cases, powers-of-2, non-powers-of-2, and realistic sizesgenerate_performance_test()fits 5× in 16 GB VRAM (4 tensors × 16 × 4096 × 64 × 4B ≈ 64 MB)pre-commit run --all-filespassesrun_challenge.py --action submitpasses all tests on NVIDIA Tesla T4🤖 Generated with Claude Code