Skip to content

[Refactor][OPS] split tileops/ops/elementwise.py into per-cluster package #1388

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

Merged
lcy-seso merged 2 commits into from
May 9, 2026
Merged

[Refactor][OPS] split tileops/ops/elementwise.py into per-cluster package #1388

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
97f0c7d
[Refactor][OPS] split elementwise.py into per-cluster package
lcy-seso May 8, 2026
25131e0
[Refactor][OPS] update manifest source.op for elementwise split
lcy-seso May 8, 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
114 changes: 57 additions & 57 deletions tileops/manifest/elementwise_binary.yaml
View file
Open in desktop

Large diffs are not rendered by default.

16 changes: 8 additions & 8 deletions tileops/manifest/elementwise_multi_input.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,11 +20,11 @@ WhereFwdOp:
shape_rules:
# PyTorch's torch.where broadcasts condition/input/other together;
# the output shape is the broadcast of all three.
- "output.shape == broadcast_shapes(condition.shape, input.shape, other.shape)"
- "output.shape == broadcast_shapes(condition.shape, input.shape, other.shape)"

workloads:
- {input_shape: [4096, 4096], dtypes: [float16, bfloat16, float32], label: "elementwise-16M"}
- {input_shape: [16384, 16384], dtypes: [float16, bfloat16], label: "elementwise-256M"}
- {input_shape: [4096, 4096], dtypes: [float16, bfloat16, float32], label: "elementwise-16M"}
- {input_shape: [16384, 16384], dtypes: [float16, bfloat16], label: "elementwise-256M"}

roofline:
# Mixed-dtype op (bool condition + float input/other) — inline mode cannot
Expand All @@ -37,7 +37,7 @@ WhereFwdOp:
kernel: tileops/kernels/elementwise.py
kernel_map:
where: WhereFwdKernel
op: tileops/ops/elementwise.py
op: tileops/ops/elementwise/where.py
test: tests/ops/test_special_elementwise.py
bench: benchmarks/ops/bench_independent_elementwise.py
bench_manifest_driven: false
Expand All @@ -61,11 +61,11 @@ LerpTensorFwdOp:
outputs:
output: {dtype: "same_as(input)"}
shape_rules:
- "output.shape == broadcast_shapes(input.shape, end.shape, weight.shape)"
- "output.shape == broadcast_shapes(input.shape, end.shape, weight.shape)"

workloads:
- {input_shape: [4096, 4096], dtypes: [float16, bfloat16, float32], label: "elementwise-16M"}
- {input_shape: [16384, 16384], dtypes: [float16, bfloat16], label: "elementwise-256M"}
- {input_shape: [4096, 4096], dtypes: [float16, bfloat16, float32], label: "elementwise-16M"}
- {input_shape: [16384, 16384], dtypes: [float16, bfloat16], label: "elementwise-256M"}

roofline:
func: "tileops.perf.formulas.lerp_tensor_fwd_roofline"
Expand All @@ -74,7 +74,7 @@ LerpTensorFwdOp:
kernel: tileops/kernels/elementwise.py
kernel_map:
lerp_tensor: LerpTensorFwdKernel
op: tileops/ops/elementwise.py
op: tileops/ops/elementwise/arithmetic.py
test: tests/ops/test_binary_arith.py
bench: benchmarks/ops/bench_binary_arith.py
bench_manifest_driven: false
Loading
Loading