Use CUDA runtime API to retrieve function pointer to driver API (#1700)

* Query pfn to driver api

* use default for older toolkits

---------

Co-authored-by: shunfans <shunfans@nvidia.com>

CMakeLists.txt

@@ -234,6 +234,7 @@ set(CUTLASS_NVCC_EMBED_CUBIN ON CACHE BOOL "Embed compiled CUDA kernel binaries
 set(CUTLASS_NVCC_EMBED_PTX ON CACHE BOOL "Embed compiled PTX into executables.")
 set(CUTLASS_NVCC_KEEP OFF CACHE BOOL "Keep intermediate files generated by NVCC.")
 set(CUTLASS_ENABLE_F16C OFF CACHE BOOL "Enable F16C x86 extensions in host code.")
+set(CUTLASS_ENABLE_DIRECT_CUDA_DRIVER_CALL OFF CACHE BOOL "Enable CUTLASS to directly call driver API)
 
 ################################################################################
 #

include/cute/atom/copy_traits_sm90_im2col.hpp

@@ -40,6 +40,8 @@
 
 #include "cute/algorithm/prefetch.hpp"
 #include "cutlass/fast_math.h"
+#include "cutlass/cuda_host_adapter.hpp"
+
 namespace cute
 {
 
@@ -450,7 +452,7 @@ make_im2col_tma_copy_desc(
   CUtensorMapFloatOOBfill tma_oob_fill    = to_CUtensorMapFloatOOBfill(aux_params.oobfill_);
   CUtensorMapSwizzle      tma_swizzle     = TMA::to_CUtensorMapSwizzle(detail::get_tma_swizzle_bits(smem_swizzle));
 
-  CUresult encode_result = cuTensorMapEncodeIm2col(
+  CUresult encode_result = CUTLASS_CUDA_DRIVER_WRAPPER_CALL(cuTensorMapEncodeIm2col)(
       &tma_desc,
       tma_format,
       num_total_modes,

include/cute/atom/copy_traits_sm90_tma.hpp

@@ -41,6 +41,7 @@
 #include <cute/algorithm/prefetch.hpp>
 
 #include <cute/numeric/integral_ratio.hpp>
+#include <cutlass/cuda_host_adapter.hpp>
 
 namespace cute
 {
@@ -983,7 +984,7 @@ make_tma_copy_desc(Tensor<GEngine,GLayout> const& gtensor,         // The origin
 
     // TMA smem swizzle type
     CUtensorMapSwizzle smem_swizzle = TMA::to_CUtensorMapSwizzle(get_tma_swizzle_bits(swizzle));
-    CUresult result = cuTensorMapEncodeTiled(
+    CUresult result = CUTLASS_CUDA_DRIVER_WRAPPER_CALL(cuTensorMapEncodeTiled)(
         &tma_desc,
         tma_format,
         tma_dim,

include/cutlass/cuda_host_adapter.hpp

@@ -82,6 +82,78 @@ namespace cutlass {
 
 /////////////////////////////////////////////////////////////////////////////////////////////////
 
+#if !defined(__CUDACC_RTC__)
+
+#include <cudaTypedefs.h>
+#include <driver_types.h>
+
+#define CUTLASS_CUDA_DRIVER_STRINGIFY(tok) #tok
+
+#if defined(CUTLASS_ENABLE_DIRECT_CUDA_DRIVER_CALL)
+
+#define CUTLASS_CUDA_DRIVER_WRAPPER_DECL(func, ver) \
+  template <typename... Args>                       \
+  CUresult call_##func(Args... args) {              \
+    return func(args...);                           \
+  }
+
+#else // defined(CUTLASS_ENABLE_DIRECT_CUDA_DRIVER_CALL)
+
+#if (__CUDACC_VER_MAJOR__ >= 12 && __CUDACC_VER_MINOR__ >= 5)
+
+#define CUTLASS_CUDA_DRIVER_WRAPPER_DECL(func, ver)             \
+  template <typename... Args>                                   \
+  CUresult call_##func(Args... args) {                          \
+    cudaDriverEntryPointQueryResult cuda_status;                \
+    void* pfn = nullptr;                                        \
+    cudaError_t cuda_err = cudaGetDriverEntryPointByVersion(    \
+        CUTLASS_CUDA_DRIVER_STRINGIFY(func),                    \
+        &pfn, ver,                                              \
+        cudaEnableDefault,                                      \
+        &cuda_status);                                          \
+    if (cuda_status != cudaDriverEntryPointSuccess ||           \
+        cuda_err != cudaSuccess) {                              \
+      return CUDA_ERROR_UNKNOWN;                                \
+    }                                                           \
+    return reinterpret_cast<PFN_##func##_v##ver>(pfn)(args...); \
+  }
+
+#else
+
+#define CUTLASS_CUDA_DRIVER_WRAPPER_DECL(func, ver)             \
+  template <typename... Args>                                   \
+  CUresult call_##func(Args... args) {                          \
+    cudaDriverEntryPointQueryResult cuda_status;                \
+    void* pfn = nullptr;                                        \
+    cudaError_t cuda_err = cudaGetDriverEntryPoint(             \
+        CUTLASS_CUDA_DRIVER_STRINGIFY(func),                    \
+        &pfn,                                                   \
+        cudaEnableDefault,                                      \
+        &cuda_status);                                          \
+    if (cuda_status != cudaDriverEntryPointSuccess ||           \
+        cuda_err != cudaSuccess) {                              \
+      return CUDA_ERROR_UNKNOWN;                                \
+    }                                                           \
+    return reinterpret_cast<PFN_##func>(pfn)(args...);          \
+  }
+
+#endif // (__CUDACC_VER_MAJOR__ >= 12 && __CUDACC_VER_MINOR__ >= 5)
+
+#endif // defined(CUTLASS_ENABLE_DIRECT_CUDA_DRIVER_CALL)
+
+#if (__CUDACC_VER_MAJOR__ >= 12)
+CUTLASS_CUDA_DRIVER_WRAPPER_DECL(cuTensorMapEncodeTiled, 12000);
+CUTLASS_CUDA_DRIVER_WRAPPER_DECL(cuTensorMapEncodeIm2col, 12000);
+#endif
+
+#undef CUTLASS_CUDA_DRIVER_STRINGIFY
+
+#define CUTLASS_CUDA_DRIVER_WRAPPER_CALL(func) cutlass::call_##func
+
+#endif // !defined(__CUDACC_RTC__)
+
+/////////////////////////////////////////////////////////////////////////////////////////////////
+
 /// This class manages runtime CUlaunchAttribute that can be supplied to CudaHostAdapter
 /// CudaHostLaunchAttributes will be an empty struct in earlier CTK where CUlaunchAttribute
 /// is not introduced.