Closed pytho12 closed 9 months ago
Hello, this is expected behavior since T4 does not support HW accelerated decode. You may want to fallback to NVJPEG_BACKEND_DEFAULT
. You can refer to this code snippet.
In the nvJPEG case, why is there no encoding case?
In the nvJPEG case, why is there no encoding case? I would be grateful if you could provide encoding cases and test performance cases for nvJPEG.
Will look into adding a new nvjpeg sample which focuses on encode. For nvjpeg encode, you can take a look at the Image-Resize sample. This sample uses the nvjpeg encode APIs.
you can also refer to the nvjpeg encoder sample here - https://github.com/NVIDIA/cuda-samples/tree/master/Samples/4_CUDA_Libraries/nvJPEG_encoder
marking as close. Feel free to reopen if you need additional information. Will add the encoder sample as part of this issue - https://github.com/NVIDIA/CUDALibrarySamples/issues/161
Hello, I am using a T4 card. When using status = nvjpegCreateEx(NVJPEG_BACKEND_HARDWARE, &dev_allocator, NULL, 0, &nv_handle ); error: NVJPEG_STATUS_ARCH_MISMATCH