Skip to content

C++11 fix warnings #1904

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
tridao merged 8 commits into from
Sep 25, 2025
Merged

C++11 fix warnings #1904

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
3832d79
errors are with C++11 narrowing warnings (treated as errors in strict…
johnnynunez Sep 23, 2025
16aee90
errors are with C++11 narrowing warnings (treated as errors in strict…
johnnynunez Sep 23, 2025
c900d27
errors are with C++11 narrowing warnings (treated as errors in strict…
johnnynunez Sep 23, 2025
932d926
errors are with C++11 narrowing warnings (treated as errors in strict…
johnnynunez Sep 23, 2025
d510f68
Update flash_api_stable.cpp
johnnynunez Sep 23, 2025
e16064e
Merge branch 'Dao-AILab:main' into main
johnnynunez Sep 23, 2025
170a081
upstream cutlass v4.2.1 csrc
johnnynunez Sep 24, 2025
2eff3f7
Merge remote-tracking branch 'origin/main'
johnnynunez Sep 24, 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
6 changes: 3 additions & 3 deletions hopper/flash_api.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -609,7 +609,7 @@ mha_fwd_get_scheduler_metadata(

// Otherwise the kernel will be launched from cuda:0 device
// Cast to char to avoid compiler warning about narrowing
at::cuda::CUDAGuard device_guard{(char)seqused_k.get_device()};
at::cuda::CUDAGuard device_guard{static_cast<c10::DeviceIndex>(seqused_k.get_device())};

auto opts = seqused_k.options();
// This needs to be set after get_num_splits
Expand Down Expand Up @@ -876,7 +876,7 @@ mha_fwd(at::Tensor q, // (b, s_q, h, d) or (total_q, h, d) if there is cu_seql

// Otherwise the kernel will be launched from cuda:0 device
// Cast to char to avoid compiler warning about narrowing
at::cuda::CUDAGuard device_guard{(char)q.get_device()};
at::cuda::CUDAGuard device_guard{static_cast<c10::DeviceIndex>(q.get_device())};

at::Tensor softmax_lse;
if (!is_varlen_q) {
Expand Down Expand Up @@ -1463,7 +1463,7 @@ std::tuple<at::Tensor, at::Tensor, at::Tensor, at::Tensor, at::Tensor, at::Tenso

// Otherwise the kernel will be launched from cuda:0 device
// Cast to char to avoid compiler warning about narrowing
at::cuda::CUDAGuard device_guard{(char)q.get_device()};
at::cuda::CUDAGuard device_guard{static_cast<c10::DeviceIndex>(q.get_device())};

auto opts = q.options();
// Need softmax_d to have total_q_padded_rounded since we want its address to be aligned by 16/8 bytes for TMA / LDG.64
Expand Down