Skip to content

Update PyTorch pin to the version which support elapsed_time; remove our patch for elapsed_time #2952

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
merged 15 commits into from
Dec 10, 2024
Merged

Update PyTorch pin to the version which support elapsed_time; remove our patch for elapsed_time #2952

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
88f56c5
Update PyTorch pin to the version which support elapsed_time; remove …
anmyachev Dec 6, 2024
985a625
Merge branch 'main' of https://github.com/intel/intel-xpu-backend-for…
anmyachev Dec 6, 2024
b3dc591
another commit
anmyachev Dec 6, 2024
797b46a
another commit
anmyachev Dec 6, 2024
bd5b0b9
Update scripts/patch-pytorch.sh
anmyachev Dec 6, 2024
e9c620d
try changes from PR 2962
anmyachev Dec 8, 2024
f446dfd
Merge branch 'main' into amyachev/issue2945
anmyachev Dec 8, 2024
35bf37a
Update .github/pins/pytorch-upstream.txt
anmyachev Dec 9, 2024
6e3c80b
use PyLong_AsUnsignedLongLong
anmyachev Dec 9, 2024
fc8cd2c
Revert "use PyLong_AsUnsignedLongLong"
anmyachev Dec 9, 2024
c13270d
REVERTME just test
anmyachev Dec 9, 2024
0d22d82
test commit which change tensor.data_ptr() behavior
anmyachev Dec 9, 2024
4d27e94
Revert "REVERTME just test"
anmyachev Dec 9, 2024
8472895
specify 'dtype=torch.uint64' as a workaround
anmyachev Dec 9, 2024
49b5a2e
Merge branch 'main' into amyachev/issue2945
anmyachev Dec 10, 2024
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
2 changes: 1 addition & 1 deletion .github/pins/pytorch-upstream.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1 @@
d0fd42eb3ac6939e63879bc055b6901103f713d3
61dc5e9c0a36d590adc47b4110efd94d9eb59306
12 changes: 6 additions & 6 deletions python/tutorials/08-grouped-gemm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,9 +171,9 @@ def group_gemm_fn(group_A, group_B):
g_lds += [A.stride(0), B.stride(0), C.stride(0)]

# note these are device tensors
d_a_ptrs = torch.tensor(A_addrs, device=DEVICE)
d_b_ptrs = torch.tensor(B_addrs, device=DEVICE)
d_c_ptrs = torch.tensor(C_addrs, device=DEVICE)
d_a_ptrs = torch.tensor(A_addrs, device=DEVICE, dtype=torch.uint64)
d_b_ptrs = torch.tensor(B_addrs, device=DEVICE, dtype=torch.uint64)
d_c_ptrs = torch.tensor(C_addrs, device=DEVICE, dtype=torch.uint64)
d_g_sizes = torch.tensor(g_sizes, dtype=torch.int32, device=DEVICE)
d_g_lds = torch.tensor(g_lds, dtype=torch.int32, device=DEVICE)
# we use a fixed number of CTA, and it's auto-tunable
Expand Down Expand Up @@ -277,9 +277,9 @@ def benchmark(N, provider):
g_sizes += [N, N, N]
g_lds += [N, N, N]

d_a_ptrs = torch.tensor(A_addrs, device=DEVICE)
d_b_ptrs = torch.tensor(B_addrs, device=DEVICE)
d_c_ptrs = torch.tensor(C_addrs, device=DEVICE)
d_a_ptrs = torch.tensor(A_addrs, device=DEVICE, dtype=torch.uint64)
d_b_ptrs = torch.tensor(B_addrs, device=DEVICE, dtype=torch.uint64)
d_c_ptrs = torch.tensor(C_addrs, device=DEVICE, dtype=torch.uint64)
d_g_sizes = torch.tensor(g_sizes, dtype=torch.int32, device=DEVICE)
d_g_lds = torch.tensor(g_lds, dtype=torch.int32, device=DEVICE)

Expand Down
1 change: 0 additions & 1 deletion scripts/patch-pytorch.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,4 +16,3 @@ echo "Applying PyTorch patches in $REPO_ROOT"
cd "$REPO_ROOT"

curl -sSL https://github.com/pytorch/pytorch/pull/126516.diff | git apply -
curl -sSL https://github.com/pytorch/pytorch/pull/126456.diff | git apply -