Skip to content

Matmul + All reduce #96

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
linamy85 wants to merge 3 commits into from
Closed

Matmul + All reduce #96

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
7df6677
Matmul + All Reduce benchmark
linamy85 Feb 2, 2026
1907768
[WIP] separate matmul and all reduce
linamy85 Feb 2, 2026
89eecaf
Optimizing
linamy85 Feb 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 14 additions & 0 deletions Ironwood/configs/training/gemm_all_reduce.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,14 @@
benchmarks:
- benchmark_name: "gemm_all_reduce"
trace_dir: "../microbenchmarks/gemm_all_reduce"
csv_path: "../microbenchmarks/gemm_all_reduce"
xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
num_runs: 10
benchmark_sweep_params:
- {m: 1024, k: 1024, n: 1024, dtype: "bfloat16"}
- {m: 2048, k: 2048, n: 2048, dtype: "bfloat16"}
- {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
- {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
- {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
- {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
73 changes: 73 additions & 0 deletions Ironwood/configs/training/local_separation_test.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,73 @@
benchmarks:
# - benchmark_name: gemm_only
# trace_dir: "../microbenchmarks/gemm_all_reduce"
# csv_path: "../microbenchmarks/gemm_all_reduce"
# xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
# xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
# num_runs: 10
# benchmark_sweep_params:
# - {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
# - {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
# - {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
# - {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
# - benchmark_name: all_reduce_only
# trace_dir: "../microbenchmarks/gemm_all_reduce"
# csv_path: "../microbenchmarks/gemm_all_reduce"
# xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
# xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
# num_runs: 10
# benchmark_sweep_params:
# - {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
# - {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
# - {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
# - {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
- benchmark_name: gemm_all_reduce
trace_dir: "../microbenchmarks/gemm_all_reduce"
csv_path: "../microbenchmarks/gemm_all_reduce"
xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
num_runs: 10
benchmark_sweep_params:
- {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
- {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
- {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
- {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
- benchmark_name: gemm_reducescatter_allgather
trace_dir: "../microbenchmarks/gemm_all_reduce"
csv_path: "../microbenchmarks/gemm_all_reduce"
xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
num_runs: 10
benchmark_sweep_params:
- {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
- {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
- {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
- {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
- benchmark_name: gemm_sharded_all_gather
trace_dir: "../microbenchmarks/gemm_all_reduce"
csv_path: "../microbenchmarks/gemm_all_reduce"
xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
num_runs: 10
benchmark_sweep_params:
- {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
- {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
- {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
- {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
- {m: 65536, k: 65536, n: 65536, dtype: "bfloat16"}
- {m: 131072, k: 131072, n: 131072, dtype: "bfloat16"}
- {m: 262144, k: 262144, n: 262144, dtype: "bfloat16"}
- {m: 524288, k: 524288, n: 524288, dtype: "bfloat16"}
- benchmark_name: gemm_k_sharded_all_reduce
trace_dir: "../microbenchmarks/gemm_all_reduce"
csv_path: "../microbenchmarks/gemm_all_reduce"
xlml_metrics_dir: "../microbenchmarks/gemm_all_reduce"
xla_dump_dir: "../microbenchmarks/gemm_all_reduce/hlo_graphs"
num_runs: 10
benchmark_sweep_params:
- {m: 4096, k: 4096, n: 4096, dtype: "bfloat16"}
- {m: 8192, k: 8192, n: 8192, dtype: "bfloat16"}
- {m: 16384, k: 16384, n: 16384, dtype: "bfloat16"}
- {m: 32768, k: 32768, n: 32768, dtype: "bfloat16"}
- {m: 65536, k: 65536, n: 65536, dtype: "bfloat16"}
- {m: 131072, k: 131072, n: 131072, dtype: "bfloat16"}
Loading