-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all,
thanks a lot for your support sp far, but so far it seems nothing really worked. I always get stuck at the same point, which is: other GL apps run fine through optirun, but darktable's openCL support fails when it comes to compiling blendop.cl. However, I could make some progress (still testing) by removing individual kernels from blendop.cl. The problem seems to be with one kernel only: __kernel void blendop_mask_RAW( ... ). In particular, the problem is with the last statement: write_imagef(mask, (int2)(x, y), gopacity); If that last line is commented out, dt compiles all openCL code flawlessly and seems to work just fine. The only significant difference I see between those blendop_mask_* kernels is that the RAW version doesn't call read_imagef() first.. well, I actually have no clue when it comes to OpenCL.. it's just what I noticed for now. Can someone please suggest a way forward from here? Thanks a lot, Jens PS: Johannes and Kevin: you both mentioned running dt on optimus cards as well. Does one of you happen to also have the same card (i.e. GT640M)? On 10/29/2012 05:11 PM, Jens Fendler wrote: > Hi all, > > I just got a new laptop, which I would like to use for DT > processing. One problem, however, seems to be the NVidia Optimus > graphics card, which gives me problems with the OpenCL acceleration > in darktable. Everything else I've tried so far sems to work fine > with bumblebee's "optirun". (Though I haven't yet tried any other > OpenCL apps yet). > > When starting DT, the compilation of blendop.cl fails with error > code -30 (kernels before that have been compiled fine). I was > assuming optirun being necessary for running OpenCL, but nothing > changes even when trying to run it on its own - OpenCL still looks > OK (i.e. detects the nvidia card), but blendop.cl compilation > fails. Subsequently, dt continues without OpenCL support. > > I then tried to disable only blendop.cl by commenting it out in > programs.conf. When now running without optirun, dt compiles all > remaining kernels, and I get the message "...FINALLY: opencl is > AVAILABLE...". But then dt tells me it can't create most of the > kernels (error codes either -44 or -46), while some kernels seem to > work just fine. This combination (i.e. not using optirun, > blendop.cl disabled) seems to be the only on in which I get at > least some OpenCL working. > > Running though optirun with blendop.cl disabled (being the > combination I expected to work as well) is actually the wost: while > I still get the message OpenCL would be available, dt then > segfaults.. > > I have attached log files (created with -d opencl) for each of the > 4 combinations, as well as the stacktrace in the hope that someone > here might have an idea what the problem could be. > > If it helps, I'd be willing to run almost every test you guys would > have for me to track this issue down.. please let me know if you > need any further information, files, etc.. and I'll provide what I > can. > > Thanks, Jens > > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ iEYEARECAAYFAlCP9FIACgkQbyTZFOIr/d1sTQCguLKdrpJJ07RyTpQwKn1TCR7b cf4AoJK8IzFgDAYww9sNCgF3+N/gTHfo =C9rj -----END PGP SIGNATURE----- ------------------------------------------------------------------------------ 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_sfd2d_oct _______________________________________________ darktable-devel mailing list darktable-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-devel
