本文章主要是記錄,cuda 編程程序中遇到的相關概念,名字解釋和問題;主要是是用來備忘:
- cuda PTX :并行執行緒執行(Parallel Thread eXecution,PTX)代碼是編譯后的GPU代碼的一種中間形式,它可以再次編譯為原生的GPU微碼,
- 下面幾條鏈接,是對nvcc 編譯程序中,編譯引數的解釋:-gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61;本編譯引數適用于 1080ti顯卡;此種模式為cuda fatbinary模式,用于優化cudabin的執行;
- https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#virtual-architecture-feature-list
- https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#gpu-feature-list
- nvcc 編譯程序理解:https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#gpu-compilation

-
上圖是:Two-Staged Compilation with Virtual and Real Architectures,詳細情況,請看上面鏈接;

-
上圖是:Just-in-Time Compilation of Device Code
- https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html#fatbinaries (CUDA fatbinaries 模式)
- https://www.cnblogs.com/zjutzz/p/10272424.html (nvcc 與 C++ 混合編譯Demo, 較為優秀)
保持更新,更多內容,請關注 cnblogs.com/xuyaowen;
轉載請註明出處,本文鏈接:https://www.uj5u.com/houduan/87998.html
標籤:C++
上一篇:POJ1852
下一篇:Run-Time Check Failure #0 - The value of ESP was not properly saved across a function call錯誤
