Skip to content
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

fix[dace][next]: Updated gt_auto_optimize() #1889

Merged
merged 3 commits into from
Feb 26, 2025
Merged

fix[dace][next]: Updated gt_auto_optimize() #1889

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2d5ad5d
Fixed an issue in the auto optimizer.
philip-paul-mueller Feb 25, 2025
abdcd60
Merge branch 'main' into fixed-auto-optimize
philip-paul-mueller Feb 26, 2025
07a7d5d
Update src/gt4py/next/program_processors/runners/dace/transformations…
philip-paul-mueller Feb 26, 2025
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
18 changes: 11 additions & 7 deletions src/gt4py/next/program_processors/runners/dace/transformations/auto_optimize.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,10 +66,11 @@ def gt_auto_optimize(
one with stride one.
5. If requested the function will now apply loop blocking, on the dimension
indicated by `leading_dim`.
6. If requested the SDFG will be transformed to GPU. For this the
6. The strides of temporaries are set to match the compute order.
7. If requested the SDFG will be transformed to GPU. For this the
`gt_gpu_transformation()` function is used, that might apply several other
optimizations.
7. Afterwards some general transformations to the SDFG are applied.
8. Afterwards some general transformations to the SDFG are applied.
This includes:
- Use fast implementation for library nodes.
- Move small transients to stack.
Expand Down Expand Up @@ -235,7 +236,13 @@ def gt_auto_optimize(
validate_all=validate_all,
)

# Phase 6: Going to GPU
# Phase 6: Setting the strides of transients
# It is important that we set the strides before the GPU transformation.
# Because this transformation will also apply `CopyToMap` for the Memlets
# that the DaCe runtime can not handle.
gtx_transformations.gt_change_transient_strides(sdfg, gpu=gpu)

# Phase 7: Going to GPU
if gpu:
# TODO(phimuell): The GPU function might modify the map iteration order.
# This is because how it is implemented (promotion and
Expand All @@ -251,7 +258,7 @@ def gt_auto_optimize(
try_removing_trivial_maps=True,
)

# Phase 7: General Optimizations
# Phase 8: General Optimizations
# The following operations apply regardless if we have a GPU or CPU.
# The DaCe auto optimizer also uses them. Note that the reuse transient
# is not done by DaCe.
Expand All @@ -267,9 +274,6 @@ def gt_auto_optimize(
# TODO(phimuell): Fix the bug, it uses the tile value and not the stack array value.
dace_aoptimize.move_small_arrays_to_stack(sdfg)

# Now we modify the strides.
gtx_transformations.gt_change_transient_strides(sdfg, gpu=gpu)

if make_persistent:
gtx_transformations.gt_make_transients_persistent(sdfg=sdfg, device=device)

Expand Down