- User Since
- Jul 4 2014, 11:26 PM (146 w, 3 d)
Sat, Apr 22
Fri, Apr 21
@Tom (vejn) Please open new reports for new issues, putting different issues into the same report is quite confusing.
This could be a driver issue with GCN 1 cards as all cards mentioned in this report are that generation. I cant reproduce this with the cards I have available, and if it is a driver bug there's nothing that we can do anyways. There is a new version of the drivers released recently you guys could try, but we may have to drop support for these cards.
Wed, Apr 19
Tue, Apr 18
Only minor stuff, don't want to be too picky. I think it will be fine to commit, but maybe would be nice if @Sergey Sharybin (sergey) or someone else could test CUDA split first (I would test it but I've been having issues with CUDA lately).
Tue, Apr 11
Sat, Apr 8
Fri, Apr 7
Fri, Mar 31
Only did a quick look right now, will look over the rest later. I like the changes to how kernels are handled.
Wed, Mar 29
Im fine with this too.
Mar 23 2017
Mar 21 2017
Mar 17 2017
Mar 16 2017
Not entirely happy with how locals are done but not sure theres really a good way to go. Maybe have DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS_$N and have kernel functions like kernel_func(KernelGlobals *kg, ccl_local unsigned int* local_1, unsigned int* local_2, ...)? That way we get errors if the number of locals falls out of sync.
Mar 14 2017
Mar 11 2017
Mar 10 2017
@James W E Bird (3dLuver), thanks for the logs, very interesting results. Will try to get this fixed quickly.
Please run blender with the --debug-cycles option and attach the full output after the tile gets stuck. Also run clinfo and attach the output of that aswell.
Mar 9 2017
Looks fine I guess, would be nice to deduplicate things more but not so important.
Mar 8 2017
Mar 7 2017
Mar 4 2017
Mar 3 2017
Feb 24 2017
@Tom (vejn) that is interesting and might point to something we can fix. Its a different issue than the one in this report tho.
Hit the wrong button
Please check you have the correct drivers installed. You say you have an AMD card but the traceback has Nvidia symbols listed? Something must not be right there. If you can please attach the output of clinfo after checking your drivers.
Host memory is system memory, not GPU memory. Either your system ran out of memory or there's a driver bug.
Feb 23 2017
Unfortunately there's not really anything we can do about your system running out of memory, which is whats happening here.
Out of host memory means your computer ran out of memory, theres no bug here.
Updated to replace pointers to KernelGlobals with pointers to KernelCamera in functions that are needed to be shared for calculating differentials and removed the duplicated versions.