group-h: Implemented optimized matrix multiplication#3
Open
z-haq wants to merge 1 commit into
Open
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Implemented blocked_matmul with cache blocking (block size 32) and parallel_matmul using OpenMP. Both pass validation against the reference output for all 10 test cases.
Results are in the README. The optimizations work best on larger matrices — on case 6 (256×256×256) blocked gets 3.79x and parallel gets 4.73x over naive. On smaller cases like 4 (32×128×32) the gains are minimal or parallel is actually slower than blocked, which makes sense since thread overhead isn't worth it for tiny workloads.
Blocked uses ikj loop order inside the block to keep the A value in a register and access B/C sequentially. Parallel just parallelizes the outer row loop with schedule(static) since rows have equal work.
Timing uses omp_get_wtime() and repeats until at least 50ms of wall time to average out noise.