Skip to content

fix https://github.com/NVIDIA/TensorRT-Edge-LLM/issues/87: hard-error instead of silent return when CuTe DSL GEMM is not compiled #97

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

Open
suharvest wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix https://github.com/NVIDIA/TensorRT-Edge-LLM/issues/87: hard-error instead of silent return when CuTe DSL GEMM is not compiled #97

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
12 changes: 8 additions & 4 deletions cpp/kernels/talkerMLPKernels/talkerMLPKernels.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -338,8 +338,10 @@ void invokeTalkerMLP(rt::Tensor const& input, rt::Tensor const& fc1Weight, rt::T
return;
}
#else
LOG_ERROR("CuTe DSL GEMM not compiled. Rebuild with -DENABLE_CUTE_DSL=gemm (or ALL).");
return;
ELLM_CHECK(false,
"invokeTalkerMLP requires CuTe DSL GEMM. Rebuild with -DENABLE_CUTE_DSL=gemm (or ALL); "
"without it the launcher returns without writing the output tensor and downstream "
"sampling consumes uninitialised GPU memory, producing wrong-but-not-crashing audio.");
#endif
}

Expand Down Expand Up @@ -382,8 +384,10 @@ void invokeLinearLayer(
return;
}
#else
LOG_ERROR("CuTe DSL GEMM not compiled. Rebuild with -DENABLE_CUTE_DSL=gemm (or ALL).");
return;
ELLM_CHECK(false,
"invokeLinearLayer requires CuTe DSL GEMM. Rebuild with -DENABLE_CUTE_DSL=gemm (or ALL); "
"without it the launcher returns without writing the output tensor and downstream "
"consumers see uninitialised GPU memory.");
#endif
}

Expand Down