Added documentation for nvrtc cache mechanism Moved the following functions in CUDA backt to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function.
9prady9 added a commit to 9prady9/arrayfire that referenced this pull request
Apr 8, 2019
* Add CUDA runtime compilation support using nvrtc Moved the following functions in CUDA backend to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function. (cherry picked from commit 7797d01 )
umar456 pushed a commit to 9prady9/arrayfire that referenced this pull request
Apr 17, 2019
* Add CUDA runtime compilation support using nvrtc Moved the following functions in CUDA backend to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function. (cherry picked from commit 7797d01 )
umar456 pushed a commit to 9prady9/arrayfire that referenced this pull request
Apr 17, 2019
* Add CUDA runtime compilation support using nvrtc Moved the following functions in CUDA backend to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function. (cherry picked from commit 7797d01 )
umar456 pushed a commit to 9prady9/arrayfire that referenced this pull request
Apr 17, 2019
* Add CUDA runtime compilation support using nvrtc Moved the following functions in CUDA backend to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function. (cherry picked from commit 7797d01 )
umar456 pushed a commit that referenced this pull request
Apr 17, 2019
* Add CUDA runtime compilation support using nvrtc Moved the following functions in CUDA backend to use runtime compilation * Transpose (In place transpose hasn't been ported yet) * Convolutions * Scan and Scan by Key The eventual goal is to use math.hpp even inside jit kernels and remove the code-path controlled by isJIT parameter of compileKernel function. (cherry picked from commit 7797d01 )
Closed