Robert Crovella questions

Robert Crovella answers

• CUDA How to launch a new kernel call in one kernel function?
• Getting calling a __host__ function from a global function is not allowed when using cudaMallocManaged
• Unified Memory and Streams in C
• CUDA why just reading (zero write) from unified memory cause next kernel to become slower
• Troubles with cudaMemcpyToSymbol
• How do we use cuPrintf()?
• Why does transform_reduce result differ from the result of transform & reduce?
• Are atomic operations in CUDA guaranteed to be scheduled per warp?
• Does CUDA quietly downcasts double to float?
• cuda 11 kernel doesn't run