NVCOMP failure when use gdeflate::decompressCPU()

Accelerated Computing GPU-Accelerated Libraries
1

1. I tried nvcomp GDeflate comp/decomp in GPU and CPU, it failed when I use gdeflate::decompressCPU() to decompress data. the project I tested is from CUDALibrarySamples/nvCOMP/examples/gdeflate_cpu_compression.cu at master · NVIDIA/CUDALibrarySamples · GitHub. But errors occurs when I add decompressCPU(), how to make it work?

test environment: win10+vs2019+cuda12.1+nvcomp_4.1.0.6, RTX A6000

…

  // Compress on the CPU using gdeflate CPU batched API
  gdeflate::compressCPU(
      input_data_cpu.ptrs(),
      input_data_cpu.sizes(),
      chunk_size,
      input_data_cpu.size(),
      compress_data_cpu.ptrs(),
      compress_data_cpu.sizes());

  printf("batch_size %d \n", input_data_cpu.size());

  // deCompress on the CPU using gdeflate CPU batched API
  gdeflate::decompressCPU(
      compress_data_cpu.ptrs(),
      compress_data_cpu.sizes(),
      output_data_cpu.size(),
      output_data_cpu.ptrs(),
      output_data_cpu.sizes());

…

2. Furthermore, I also tested nvcompBatchedGzipDecompressAsync() in CUDALibrarySamples/nvCOMP/examples/gzip_gpu_decompression.cu at master · NVIDIA/CUDALibrarySamples · GitHub, but there is no nvcompBatchedGzipCompressAsync(), will you add this feature in future?

2

the decompressCPU() erros shows below:

There was an unhandled exception: Microsoft C++ exception: std::runtime_error at memory location 0x000000A96452F208.