You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Updated the samples using the cudaDeviceProp fields which are deprecated and removed in CUDA 13.0, replacing the fields with the equivalents in "cudaDeviceGetAttribute":
Deprecated "cudaDeviceProp" fields int clockRate; // - Replaced with "cudaDevAttrClockRate" int deviceOverlap; // - Replaced with "cudaDevAttrGpuOverlap */ int kernelExecTimeoutEnabled; // - Replaced with "cudaDevAttrKernelExecTimeout int computeMode; // - Replaced with "cudaDevAttrComputeMode" */ int memoryClockRate; // - Replaced with "cudaDevAttrMemoryClockRate" int cooperativeMultiDeviceLaunch; // - Deprecated, cudaLaunchCooperativeKernelMultiDevice is deprecated.
0_Introduction
UnifiedMemoryStreams
simpleHyperQ
simpleIPC
simpleMultiCopy
systemWideAtomics
1_Utilitie
deviceQuery
2_Concepts_and_Techniques
streamOrderedAllocationIPC
4_CUDA_Libraries
simpleCUBLASXT
5_Domain_Specific
simpleVulkan
vulkanImageCUDA
Updated the samples using the CUDA driver API "cuCtxCreate" with adding the parameter "CUctxCreateParams" as "cuCtxCreate" is updated to "cuCtxCreate_v4" by default in CUDA 13.0:
Common
nvrtc_helper.h
0_Introduction
UnifiedMemoryStreams
matrixMulDrv
simpleTextureDrv
vectorAddDrv
vectorAddMMAP
2_Concepts_and_Techniques
EGLStream_CUDA_CrossGPU
EGLStream_CUDA_Interop
threadMigration
3_CUDA_Features
graphMemoryFootprint
memMapIPCDrv
4_CUDA_Libraries
jitLto
7_libNVVM
cuda-c-linking
device-side-launch
simple
uvmlite
8_Platform_Specific/Tegra
EGLSync_CUDAEvent_Interop
Updated the sample using CUDA API "cudaGraphAddNode"/"cudaStreamGetCaptureInfo" with adding "cudaGraphEdgeData" pointer parameter as they are updated to "cudaGraphAddNode_v2"/"cudaStreamGetCaptureInfo_v3" by default in CUDA 13.0:
3_CUDA_Features
graphConditionalNodes
Updated the samples using CUDA API "cudaMemAdvise"/"cudaMemPrefetchAsync" with changing the parameter "int device" to "cudaMemLocation location" as they are updated to "cudaMemAdvise_v2"/"cudaMemPrefetchAsyn_v2" by default in CUDA 13.0.
4_CUDA_Libraries
conjugateGradientMultiDeviceCG
6_Performance
UnifiedMemoryPerf
Replaced "thrust::identity()" with "cuda::std::identity()" as it is deprecated in CUDA 13.0.
2_Concepts_and_Techniques
segmentationTreeThrust
Updated the the headers file and samples for CUFFT error codes update.
Deprecated CUFFT errors:
CUFFT_INCOMPLETE_PARAMETER_LIST
CUFFT_PARSE_ERROR
CUFFT_LICENSE_ERROR
New added CUFFT errors:
CUFFT_MISSING_DEPENDENCY
CUFFT_NVRTC_FAILURE
CUFFT_NVJITLINK_FAILURE
CUFFT_NVSHMEM_FAILURE
Header files and samples that are related with this change:
Common/helper_cuda.h
4_CUDA_Libraries
simpleCUFFT
simpleCUFFT_2d_MGPU
simpleCUFFT_MGPU
simpleCUFFT_callback
Updated toolchain for cross-compilation for Tegra QNX platforms.