gridMoth Posted June 14, 2019 Share Posted June 14, 2019 (edited) Quote Crash report from lp; Houdini FX Version 17.5.285 [linux-x86_64-gcc6.3] Uptime 15 seconds Fri Jun 14 13:58:31 2019 Caught signal 11 Traceback from 6294 ThreadId=0x7f859d516c00 AP_Interface::coreDumpChaser(UTsignalHandlerArg) <libHoudiniUI.so> AP_Interface::si_CrashHandler::chaser(UTsignalHandlerArg) <libHoudiniUI.so> signalCallback(UTsignalHandlerArg) <libHoudiniUT.so> UT_Signal::UT_ComboSignalHandler::operator()(int, siginfo_t*, void*) const <libHoudiniUT.so> UT_Signal::processSignal(int, siginfo_t*, void*) <libHoudiniUT.so> __funlockfile <libpthread.so.0> So, I'm hoping someone can help me figure out why Houdini is crashing so much with this. This is a high end workstation with dual Xeon scalables and RTX cards. Linux is picking up all three of my cards it seems just fine, but Houdini is not. Definitely not able to connect the OCL variable, and even changing the misc.preference, as well as the .env to use CPU, it fails to change the crashing behaviour. Firstly, the biggest notice is trying to open the About Houdini pane. Fast crash. But then, certain pyro sims will seem okay, but then a wave container might crash. Haven't really dialed in all the variances because it's all likely due to the same thing. Any sort of erode in a heightfield is a crash, but if I load the crash file back up, it looks to have been at least partially sim'd, but middle clicking on the node to see the output is a crash as well. I've read a couple other threads on odforce and elsewhere about Nvidia drivers and possilby even Realtek audio drivers etc., but not sure how to disable or update the intel driver from my motherboard yet. Using Nvidia 430.26 hgpuinfo: Finds my 2080ti as GPU, but fallback on the OCL_DEVICE to CPU (which doesn't seem to really work either). I've got two RTX Titans as well in there, so I've tried setting the OCL_VENDOR to Nvidia Corp, and DEVICENUMBER to 0,1,2 respectively. Anyone else have this type of behaviour, or can give me a bit more insight based on that error output? Guessing this is likely the Nvidia driver, but really not sure. Also tried this on the 17.5.173 build. Would rolling back the Nvidia driver be beneficial perhaps?... that's my next attempt, but wanted to get this post up sooner than later. Much appreciated Edited June 15, 2019 by gridMoth Quote Link to comment Share on other sites More sharing options...
Atom Posted June 15, 2019 Share Posted June 15, 2019 (edited) Signal 11 is generally a bad memory chip, or excessive overclocking. Try dropping your CPU multiplier down a half step, in the bios configuration section of the motherboard. With three GPUs in your system, make sure your power supply has enough watts to support that. Or take out all but one GPU and trouble shoot as usual. Can your system pass the Prime95 test? Edited June 15, 2019 by Atom Quote Link to comment Share on other sites More sharing options...
gridMoth Posted June 15, 2019 Author Share Posted June 15, 2019 (edited) 9 hours ago, Atom said: Signal 11 is generally a bad memory chip, or excessive overclocking. Try dropping your CPU multiplier down a half step, in the bios configuration section of the motherboard. With three GPUs in your system, make sure your power supply has enough watts to support that. Or take out all but one GPU and trouble shoot as usual. Can your system pass the Prime95 test? First off, thanks for responding. I am definitely not overclocking, and will look into the stress test and the bios adjustment. Update: Let the Prime95 test run for about 2 hours and all was clean. Feel like that is enough to be confident in the machine's stabiity. This just seems very centered around Houdini not picking up on my other cards, or being able to determine an OCL device. Doesn't seem like I'd be able to crush—literally a 5 million particle sim for over 2 minutes with it running great, and then have it crash due to lack of power for an "About Houdini" window pane, or doing a simple erosion and a basic terrain, ya know? Edited June 15, 2019 by gridMoth Quote Link to comment Share on other sites More sharing options...
gridMoth Posted June 16, 2019 Author Share Posted June 16, 2019 (edited) Figured it out. Needed to install the opencl nvidia driver seperately. Edited June 16, 2019 by gridMoth Quote Link to comment Share on other sites More sharing options...
Cybernetix Posted September 9, 2020 Share Posted September 9, 2020 (edited) On 6/16/2019 at 3:04 AM, gridMoth said: Figured it out. Needed to install the opencl nvidia driver seperately. Hey Mike, can u point me as to where to find that driver (i need open cl 1.2)? the nvidia website sends me on a wild goose chase ive run clinfo and its looks like everything is fine, but im also getting the same crash as you and would like to try your fix Thanks so much Edited September 9, 2020 by Cybernetix 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.