Skip to content

Fix CI by moving B200 to cuda13 and downgrade a100/h100 to cuda12.8 #674

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 1 commit into from
Sep 24, 2025
Merged

Fix CI by moving B200 to cuda13 and downgrade a100/h100 to cuda12.8 #674

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
24 changes: 12 additions & 12 deletions .github/matrix.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,17 +4,17 @@
"runner": "linux.g5.4xlarge.nvidia.gpu",
"python-version": "3.10",
"ref-eager": false,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:12.8.1-devel-ubuntu24.04",
"runtime-version": "cu128",
"container-options": "--gpus all",
"alias": "a10g"
},
{
"runner": "linux.g5.4xlarge.nvidia.gpu",
"python-version": "3.12",
"ref-eager": false,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:12.8.1-devel-ubuntu24.04",
"runtime-version": "cu128",
"container-options": "--gpus all",
"alias": "a10g-dtype-asserts",
"dtype-asserts": true,
Expand All @@ -24,35 +24,35 @@
"runner": "linux.g5.4xlarge.nvidia.gpu",
"python-version": "3.12",
"ref-eager": false,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:12.8.1-devel-ubuntu24.04",
"runtime-version": "cu128",
"container-options": "--gpus all",
"alias": "a10g"
},
{
"runner": "linux.g5.4xlarge.nvidia.gpu",
"python-version": "3.12",
"ref-eager": true,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:12.8.1-devel-ubuntu24.04",
"runtime-version": "cu128",
"container-options": "--gpus all",
"alias": "a10g-ref-eager"
},
{
"runner": "linux.aws.h100",
"python-version": "3.12",
"ref-eager": false,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:12.8.1-devel-ubuntu24.04",
"runtime-version": "cu128",
"container-options": "--gpus all",
"alias": "h100"
},
{
"runner": "linux.dgx.b200",
"python-version": "3.12",
"ref-eager": false,
"image": "nvidia/cuda:12.9.1-devel-ubuntu24.04",
"runtime-version": "cu129",
"image": "nvidia/cuda:13.0.1-devel-ubuntu24.04",
"runtime-version": "cu130",
"container-options": "--gpus all",
"alias": "b200"
},
Expand Down
3 changes: 1 addition & 2 deletions test/test_loops.expected
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1079,7 +1079,6 @@ from __future__ import annotations
import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_compat import libdevice
from helion.runtime import default_launcher as _default_launcher

@triton.jit
Expand Down Expand Up @@ -1115,7 +1114,7 @@ def _helion_three_pass_kernel(x, out, out_stride_0, out_stride_1, x_stride_0, x_
v_8 = v_6 - v_7
v_9 = 1e-06
v_10 = v_8 + v_9
v_11 = libdevice.sqrt(v_10)
v_11 = tl.sqrt_rn(v_10)
for offset_3 in tl.range(0, M.to(tl.int32), _BLOCK_SIZE_3):
indices_3 = offset_3 + tl.arange(0, _BLOCK_SIZE_3).to(tl.int32)
mask_3 = indices_3 < M
Expand Down
2 changes: 1 addition & 1 deletion test/test_reductions.expected
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -164,7 +164,7 @@ def _helion_multi_math_ops_fp16_kernel(x, result, x_size_0, result_stride_0, res
tl.store(result + (indices_0 * result_stride_0 + 0 * result_stride_1), v_2, mask_0)
load_1 = tl.load(x + indices_0 * x_stride_0, mask_0, other=0)
v_3 = tl.cast(load_1, tl.float32)
v_4 = libdevice.sqrt(v_3)
v_4 = tl.sqrt_rn(v_3)
v_5 = tl.cast(v_4, tl.float16)
tl.store(result + (indices_0 * result_stride_0 + 1 * result_stride_1), v_5, mask_0)
load_2 = tl.load(x + indices_0 * x_stride_0, mask_0, other=0)
Expand Down
Loading