feat(cuda): upgrade x86_64-cuda to 11.8

GitOrigin-RevId: 21f3ee70312eaa39454e30e7e698fce53ae7ac53

CMakeLists.txt

@@ -767,6 +767,16 @@ if(MGE_WITH_CUDA)
         set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_61,code=sm_61")
         set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_70,code=sm_70")
         set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_75,code=sm_75")
+      elseif(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER "11.8.0"
+             OR ${CMAKE_CUDA_COMPILER_VERSION} VERSION_EQUAL "11.8.0")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_61,code=sm_61")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_70,code=sm_70")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_75,code=sm_75")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_80,code=sm_80")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_86,code=sm_86")
+        set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_89,code=sm_89")
+        set(MGE_CUDA_GENCODE
+            "${MGE_CUDA_GENCODE} -gencode arch=compute_89,code=compute_89")
       elseif(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER "11.1.0"
              OR ${CMAKE_CUDA_COMPILER_VERSION} VERSION_EQUAL "11.1.0")
         set(MGE_CUDA_GENCODE "${MGE_CUDA_GENCODE} -gencode arch=compute_61,code=sm_61")

dnn/cuda-stub/src/libcuda.cpp

@@ -59,6 +59,8 @@ CUresult on_init_failed(int func_idx) {
 #include "./libcuda-wrap_11.2.h"
 #elif CUDA_VERSION == 11040
 #include "./libcuda-wrap_11.4.h"
+#elif CUDA_VERSION == 11080
+#include "./libcuda-wrap_11.8.h"
 #else
 #error "cuda stub not support this cuda version, you can close cuda stub to passby"
 #endif

dnn/src/cuda/kernel_common/diagnostic_prologue.cuh

@@ -3,6 +3,7 @@
 #else
 #define MEGDNN_DIAGNOSTIC_PROLOGUE_INCLUDED
 #endif
+#include <cuda.h>
 
 //! see
 //! https://stackoverflow.com/questions/49836419/how-to-hide-nvccs-function-was-declared-but-never-referenced-warnings

scripts/whl/manylinux2014/build_wheel_common.sh

 #!/bin/bash -e
-set -x
+set -ex
 
 CWD=$(dirname $0)
 BASEDIR=$(readlink -f ${CWD}/../../..)
@@ -12,7 +12,7 @@ CUDA_LIB_DIR="/usr/local/cuda/lib64/"
 TensorRT_LIB_DIR="/opt/tensorrt/lib/"
 
 SDK_NAME="unknown"
-x86_64_support_version="cu101 cu111 cu112 cpu cu111_cudnn821_tensorRT825 cu114"
+x86_64_support_version="cu101 cu111 cu112 cpu cu111_cudnn821_tensorRT825 cu114 cu118"
 aarch64_support_version="cu102_JetsonNano cu111 cpu"
 if [[ -z ${IN_CI} ]]
 then
@@ -225,6 +225,44 @@ elif [ $SDK_NAME == "cu114" ];then
         -gencode arch=compute_86,code=sm_86 \
         -gencode arch=compute_86,code=compute_86\" "
 
+elif [ $SDK_NAME == "cu118" ];then
+    BUILD_GCC8="ON"
+    REQUIR_CUDA_VERSION="11080"
+    REQUIR_CUDNN_VERSION="8.6.0"
+    REQUIR_TENSORRT_VERSION="8.5.3.1"
+    REQUIR_CUBLAS_VERSION="11.11.3.6"
+
+    # override the default cuda/cudnn/trt lib dir
+    CUDNN_LIB_DIR="/opt/cudnn/lib/"
+    CUDA_LIB_DIR="/usr/local/cuda/targets/x86_64-linux/lib/"
+    TensorRT_LIB_DIR="/opt/tensorrt/lib/"
+
+    CUDA_COPY_LIB_LIST="\
+        ${CUDA_LIB_DIR}/libnvrtc.so.11.2:\
+        ${CUDA_LIB_DIR}/libcublasLt.so.11:\
+        ${CUDA_LIB_DIR}/libcublas.so.11:\
+        ${CUDNN_LIB_DIR}/libcudnn_adv_infer.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn_adv_train.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn_cnn_infer.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn_cnn_train.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn_ops_infer.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn_ops_train.so.8:\
+        ${CUDNN_LIB_DIR}/libcudnn.so.8:\
+        ${TensorRT_LIB_DIR}/libnvinfer_plugin.so.8:\
+        ${TensorRT_LIB_DIR}/libnvonnxparser.so.8\
+        ${TensorRT_LIB_DIR}/libnvinfer_builder_resource.so.8.5.3:\
+        ${TensorRT_LIB_DIR}/libnvparsers.so.8:\
+        ${TensorRT_LIB_DIR}/libnvinfer.so.8"
+
+    EXTRA_CMAKE_FLAG=" -DMGE_WITH_CUDNN_SHARED=ON -DMGE_WITH_CUBLAS_SHARED=ON \
+        -DMGE_CUDA_GENCODE=\"-gencode arch=compute_61,code=sm_61 \
+        -gencode arch=compute_70,code=sm_70 \
+        -gencode arch=compute_75,code=sm_75 \
+        -gencode arch=compute_80,code=sm_80 \
+        -gencode arch=compute_86,code=sm_86 \
+        -gencode arch=compute_89,code=sm_89 \
+        -gencode arch=compute_89,code=compute_89\" "
+
 
 elif [ $SDK_NAME == "cpu" ];then
     echo "use $SDK_NAME without cuda support"
@@ -308,7 +346,7 @@ if [ ${BUILD_WHL_CPU_ONLY} = "OFF" ]; then
     CUBLAS_VERSION_CONTEXT=$(head -150 ${CUBLAS_VERSION_PATH})
     CUDA_VERSION_CONTEXT=$(head -300 ${CUDA_VERSION_PATH})
     CUDNN_VERSION_CONTEXT=$(head -62 ${CUDNN_VERSION_PATH})
-    TENSORRT_VERSION_CONTEXT=$(tail -12 ${TENSORRT_VERSION_PATH})
+    TENSORRT_VERSION_CONTEXT=$(tail -20 ${TENSORRT_VERSION_PATH})
 
     if [ "$REQUIR_CUDA_VERSION" -ge "11000" ];then
         CUDA_API_VERSION=$(echo $CUDA_VERSION_CONTEXT | grep -Eo "define CUDA_VERSION * +([0-9]+)")

src/tensorrt/include/megbrain/tensorrt/tensorrt_opr.h

@@ -36,8 +36,9 @@ enum class Empty : int32_t {};
 #define TENSOR_RT_MANAGE_ALL_WORKSPACE 0
 #endif
 
-#if NV_TENSOR_RT_VERSION >= 8000
-#error "if trt8 fix https://github.com/NVIDIA/TensorRT/issues/2290, try TENSOR_RT_MANAGE_ALL_WORKSPACE=0"
+//! issue detail: https://github.com/NVIDIA/TensorRT/issues/2290
+#if NV_TENSOR_RT_VERSION >= 8400
+#define TENSOR_RT_MANAGE_ALL_WORKSPACE 0
 #endif
 
 namespace mgb {