Re: [Bf-committers] How much ram is needed for compiling cuda kernels ?
Thanks, that is the elegant solution for my problem... I did as you said and the kernel was compiled on first start without a problem... -- > > Message: 6 > Date: Fri, 19 Apr 2013 08:49:09 -0400 > From: trouble daemon > Subject: Re: [Bf-committers] How much ram is needed for compiling cuda > kernels ? > To: bf-blender developers > Message-ID: > nd...@mail.gmail.com> > Content-Type: text/plain; charset=ISO-8859-1 > > Hi Marko, > > I use ccmake (the ncurses version of cmake-menu) here in Linux. For my > cuda compiling, I generally leave WITH_CYCLES_CUDA_BINARIES set to > OFF, which has the effect of delaying building of the cuda kernel > until I actually run blender and enable cycles for the first time. > After the first run, this should have the effect of only comping the > cuda that is actually needed for your system. > > If you would rather build your kernel during compile instead, set that > above line to ON, and then in advanced settings, change > CYCLES_CUDA_BINARIES_ARCH to sm_21 only (the newer default is: > "sm_20;sm_21;sm_30" for me atm, but I think it had sm_13 in there > until recently). > > Hope this helps, gl o/ > > > Dan > > On Fri, Apr 19, 2013 at 8:31 AM, Marko Radojcic > wrote: > > Thanks. This is the answer I was looking for. > > > > My GT 430 has 2.1 compute capability so 1.3 is not needed for me, and > > apparently I didn't force the sm_21 value properly. > > > > What shoud I edit to force compiling only 2.1 cuda kernel ? > > I obviously didn't properly understand the wiki page explaining it ... > > > > I am using cmake, but I can work with scons if the setting is easier to > > achieve... > > > > > > > > > >> We dropped official support for sm 1.3, partly because it is so > >> difficult to compile for. Which build system are you using? Perhaps > >> this is not properly configured. > >> > >> > How much RAM is required to build the kernels? - that is my main > >> question. > >> > > >> > Is 2.5GB limit within the OS or the build environment ? > >> > > >> > I'm using gcc 4.6.2 and CUDA 5.0 > >> > >> I don't know the number, it changes each time we add more features. > >> For sm 2.x or newer RAM usage is a lot lower than sm 1.3 so I suggest > >> just trying to find out how to configure things to avoid the issue, as > >> building sm 1.3 on 32 bit is something we don't support anymore. > >> > >> Brecht. > >> > >> > >> > > ___ > > Bf-committers mailing list > > Bf-committers@blender.org > > http://lists.blender.org/mailman/listinfo/bf-committers > ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] How much ram is needed for compiling cuda kernels ?
Thanks. This is the answer I was looking for. My GT 430 has 2.1 compute capability so 1.3 is not needed for me, and apparently I didn't force the sm_21 value properly. What shoud I edit to force compiling only 2.1 cuda kernel ? I obviously didn't properly understand the wiki page explaining it ... I am using cmake, but I can work with scons if the setting is easier to achieve... > We dropped official support for sm 1.3, partly because it is so > difficult to compile for. Which build system are you using? Perhaps > this is not properly configured. > > > How much RAM is required to build the kernels? - that is my main > question. > > > > Is 2.5GB limit within the OS or the build environment ? > > > > I'm using gcc 4.6.2 and CUDA 5.0 > > I don't know the number, it changes each time we add more features. > For sm 2.x or newer RAM usage is a lot lower than sm 1.3 so I suggest > just trying to find out how to configure things to avoid the issue, as > building sm 1.3 on 32 bit is something we don't support anymore. > > Brecht. > > > ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] How much ram is needed for compiling cuda kernels ?
Hi! First of all I would like to appologize for a message from badoo website that was sent by my e-mail address a week ago. I didn't know that website would spam all my contact. I certanly didn't tell it to. Now to the business... I am using Ubuntu 12.10 32 bit. My computer has 4 GB of RAM and PAE kernel, so it actually sees the whole 4 GB. I regularly compile "my own" version of blender from svn and use prebuild versions with CUDA and OpenCL support. Now I tried to compile the CUDA build myself. Although I have set the variable to use only "sm_21" kernel type, the build step said it was building for compute capability 1.3, which is fine in general, but it produces "memory allocation error" - at that moment the process monitor said that "be" compiler is using 2.5 GB of RAM. How much RAM is required to build the kernels? - that is my main question. Is 2.5GB limit within the OS or the build environment ? I'm using gcc 4.6.2 and CUDA 5.0 Cycles build without CUDA support compiles and runs fine. Thanks for any help you can provide. Best wishes, Marko. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] Is Cycles going to support OpenCL 1.0 ?
Hi! I've been playing with Cycles these days, but I only have HD 4850 Radeon that supports OpenCL 1.0, not 1.1, so I cannot test the full potential of GPU rendering in cycles for ati... Is OpenCL 1.1 necessary ? If so, it eliminates all pre- HD 5xxx ATI GPUs which is a shame because many of those are still quite decent and powefull GPUs... Greetings,... Marko ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Official Automated Builds
Hi, I read the comments about official automated builds and I think it's definitely a good idea, but I'm just a occasional developer so still I think it is great that main developers are considering pros and cons and how to do of such an idea. It is possible that such solution might create more download traffic and that it might create a nightmare for build systems, but I remember taking part in BGE development and I might say that such automations might attract more people to develop and test Blender which is another great opportunity for Blender to grow and improve. Best of wishes to all of you. Marko ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] [#25436] Crash when using high resolution smoke sim
Hi, I tried to reproduce the bug and a similar one occured. I'm using Ubuntu Linux 10.10 and I have ATI graphics card. I ran the file attached with the bug report in today's svn blender version by using CLI: ~/blender-build/cmake/bin/blender --redner-anim filename.blend UI opened and I used alt+a to play the animation without any modifications to the file. It didn't crash, and baking was not crashing while in blender, but when I exited blender it complained about corrupted double linked list It might be a problem with graphics settings of the Ubuntu desktop because X server is complaining about: AtiCallFGLComposite in the backtrace. In situations like this, as for what I can remember usually removing compiz package and installing original ATI graphics drivers would solve most of the problems. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] GPU computing
Hi! As far as developing for different platforms, I can provide Linux OpenCL coverage, at least participate in it. I am working with ATI Stream in Linux and Windows. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] GPU computing
Hi! I was working on such an idea as well, some time ago, I think that Blender would most benefit from GPU raytracing and fluid simulations... I recommend OpenCL, it is a framework that integrates CPU, ATI GPU and NVIDIA GPU functionality. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
