leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 27 Packages Projects Releases Wiki Activity

Fix some sign conversion warnings (#1172)

Browse Source 
* Fix sign conversion warnings

* Fix type conversion warnings

* Fix sign conversion warnings

* Change smem_size_ to constexpr

* clang warnings

* undo cast change

* one miss change

* missing part

---------

Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
This commit is contained in:
cyyever
2023-11-30 00:28:40 -05:00
committed by GitHub
co-authored by Haicheng Wu
parent 99c4eebe3b
commit 10b850f9c7
5 changed files with 22 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/cutlass/gemm/device/gemm_universal_base.h
+3 -12
View File
@@ -112,7 +112,8 @@ protected:
CUTLASS_THREAD_LOCAL static int sm_occupancy_;
/// Kernel dynamic shared memory allocation requirement
CUTLASS_THREAD_LOCAL static int smem_size_;
/// Update the kernel function's shared memory configuration for the current device
static constexpr size_t smem_size_ = sizeof(typename GemmKernel::SharedStorage);
/// Initialize static thread-local members for the thread's current device,
/// if necessary.
@@ -143,11 +144,8 @@ protected:
return Status::kErrorInternal;
}
// Update the kernel function's shared memory configuration for the current device
smem_size_ = int(sizeof(typename GemmKernel::SharedStorage));
// If requires more than 48KB: configure for extended, dynamic shared memory
if (smem_size_ >= (48 << 10))
if constexpr (smem_size_ >= (48 << 10))
{
cudart_result = cudaFuncSetAttribute(
Kernel2<GemmKernel>,
@@ -377,7 +375,6 @@ public:
}
};
/////////////////////////////////////////////////////////////////////////////////////////////////
/// Static initializers
/////////////////////////////////////////////////////////////////////////////////////////////////
@@ -394,12 +391,6 @@ CUTLASS_THREAD_LOCAL int GemmUniversalBase<GemmKernel_>::device_sms_ = -1;
template <typename GemmKernel_>
CUTLASS_THREAD_LOCAL int GemmUniversalBase<GemmKernel_>::sm_occupancy_ = -1;
/// Kernel dynamic shared memory allocation requirement
template <typename GemmKernel_>
CUTLASS_THREAD_LOCAL int GemmUniversalBase<GemmKernel_>::smem_size_ = -1;
/////////////////////////////////////////////////////////////////////////////////////////////////
} // namespace device