If it possible we give this PTX code to nVidia, and ask them when it can be supported? Not sure if it is a bug on the nVidia's side, or it is leftout intentionally.
Chuck On 3/22/2013 5:22 PM, Peter Colberg wrote: > On Sat, Mar 23, 2013 at 01:55:30AM +0200, Pekka Jääskeläinen wrote: >> Also, the kernel uses "automatic locals" in contrast to the >> "host allocated" kernel local args, if that makes any difference. > That is the point. Automatic local memory arrays (with static size) > are supported by both CUDA and OpenCL, so in this case cuModuleLoad* > will have no problem loading PTX generated from OpenCL using LLVM > NVPTX. > > The issue is local and constant memory passed by parameter. The CUDA > driver rejects these cases, despite supporting them in OpenCL. > > This gives the impression that the real issue with this pocl backend > will not be the LLVM NVPTX backend, but the CUDA driver, which is > outside the realm of LLVM. In the above case, the CUDA driver rejects > valid PTX code without giving any reason. > > Peter > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_d2d_mar > _______________________________________________ > pocl-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/pocl-devel > ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_mar _______________________________________________ pocl-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/pocl-devel
