[Kernel] (1/N) Machete - Hopper Optimized Mixed Precision Linear Kernel (#7174)

This commit is contained in:
Lucas Wilkinson
2024-08-20 07:09:33 -06:00
committed by GitHub
parent b6f99a6ffe
commit 5288c06aa0
28 changed files with 4828 additions and 2 deletions
+10
View File
@@ -1,5 +1,15 @@
#pragma once
#if defined(__CUDACC__) || defined(_NVHPC_CUDA)
#define HOST_DEVICE_INLINE __forceinline__ __host__ __device__
#define DEVICE_INLINE __forceinline__ __device__
#define HOST_INLINE __forceinline__ __host__
#else
#define HOST_DEVICE_INLINE inline
#define DEVICE_INLINE inline
#define HOST_INLINE inline
#endif
int64_t get_device_attribute(int64_t attribute, int64_t device_id);
int64_t get_max_shared_memory_per_block_device_attribute(int64_t device_id);