# Commit 69e26835e would burn my discrete GPU!

**URL:** https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839
**Category:** Development
**Created:** [January 14, 2018, 8:23pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839 "2018-01-14T20:23:19Z")
**Posts on this page:** 20
**Page:** 1

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 14, 2018, 8:23pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/1 "2018-01-14T20:23:19Z")

</div>

Since commit 69e26835e, Slicer freezes as soon as I move a Volume Rendering model when I use the discrete GPU of my laptop(DRI\_PRIME=1). Temperature increases to near 100°C until the system itself hangs with a noisy fan spinning. kill, killall, pkill and xkill don’t stop Slicer. I have no other choice than powering off by long pressing the power button. I have to wait 10-15 minutes before getting back a usable system, I even thought some internal device was burnt.

This does not happen with commit 69e26835e if I just use the iGPU, but rendering is not sharp as it is less powerful than the dGPU. It does not happen on my desktop PC with a single GPU PCI-E card neither.

No such behaviour happens with preceding commit 486c6ee7a. I can offload to the dGPU and voilà.

Commit 69e26835e introduces many changes, there are probably things to fix.

I build Slicer on Arch Linux, configured as follows :

> cmake -DSlicer\_VTK\_VERSION\_MAJOR:INT=9 -DQt5\_DIR:PATH=/usr/lib/cmake/Qt5 -DSlicer\_USE\_PYTHONQT\_WITH\_OPENSSL:BOOL=0 -DBUILD\_TESTING:BOOL=0 -DCMAKE\_BUILD\_TYPE:STRING=Release …/Slicer

with gcc 7.2.1 20171224, cmake version 3.10.1 and Qt 5.10.0.

For the record, my laptop has an AMD A10-7300 APU, with an AMD Kaveri iGPU, and an AMD Topaz dGPU. I am using the amdgpu kernel module and the open source xf86-video-amdgpu driver.

* * *

Secondly, I need your advice for another issue.

(question moved to [How can we avoid removing all VTK directories when updating slicer](https://discourse.slicer.org/t/how-can-we-avoid-removing-all-vtk-directories-when-updating-slicer/1840))

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [January 15, 2018, 3:17pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/2 "2018-01-15T15:17:32Z")

</div>

The is certainly below Slicer’s level, but we can help in troubleshooting. Overheating is a low-level (hardware, driver, configuration) issue, but if you only encounter this with Slicer and other VTK-based applications then it is worth investigating how VTK uses your GPU differently compared to other applications.

Can you reproduce this with a simple VTK GPU volume rendering example? You can try [this example](https://lorensen.github.io/VTKExamples/site/Cxx/VolumeRendering/SimpleRayCast/) as it is, and then replace `vtkFixedPointVolumeRayCastMapper` by `vtkGPUVolumeRayCastMapper`.

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 15, 2018, 7:55pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/3 "2018-01-15T19:55:34Z")

</div>

I built VTK9 on my laptop, I dont know if it’s the official one from Kitware or the fork at github/Slicer.

SimpleRayCast did not hang when the model is rotated, neither with vtkFixedPointVolumeRayCastMapper nor with vtkGPUVolumeRayCastMapper. That’s for both DRI\_PRIME={0,1}. Example application GPURenderDemo in the VTK tree runs without issue also.

I am downloading the VTK fork at gitub/Slicer to test, I suppose it’s the one Slicer uses.

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [January 15, 2018, 8:03pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/4 "2018-01-15T20:03:52Z")

</div>

You can find the details of VTK repository and git branch/tag in External\_VTKv9.cmake:

> <https://github.com/Slicer/Slicer/blob/b748b88cefa57bc126f3a5c64a4fb59c28e4bb9b/SuperBuild/External_VTKv9.cmake#L103-L121>

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 15, 2018, 9:29pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/5 "2018-01-15T21:29:24Z")

</div>

> [@lassoan](#):
>
> [GitHub - Slicer/VTK: WARNING: This is NOT the official upstream VTK git repository](http://github.com/Slicer/VTK.git)

Ah, VTK is downloaded from github while building Slicer, so I just tested two similar trees, with same results.

Anyway, my main use case is not my laptop. So I’ll still be able to keep pace with Slicer development, until the nightly builds on your website come with VTK9. I need that for Volume Rendering as it just works without the GPU proprietary driver, which is not the case for VTK7 currently in the nightly builds.

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 11:24am UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/6 "2018-01-16T11:24:34Z")

</div>

To my dismay, I just discovered that Slicer freezes when rotating a Volume Rendering model on my work desktop PC also. It has a single AMD RX480 GPU. Using therefore a build from a previous commit.

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [January 16, 2018, 1:48pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/7 "2018-01-16T13:48:15Z")

</div>

What is the size of your volume?  
How much GPU memory usage do you allow in volume rendering settings?

Could you try if you have similar issues with volume rendering in ParaView?

---

<div class="post-metadata">

### Author: ![Sankhesh\_Jhaveri](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/sankhesh_jhaveri/32/1081_2.png) [@Sankhesh\_Jhaveri](https://discourse.slicer.org/u/Sankhesh_Jhaveri)
#### Post date: [January 16, 2018, 2:03pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/8 "2018-01-16T14:03:44Z")

</div>

Going over past conversations in this thread, were you able to run SimpleRayCast on the desktop with AMD GPU. Did you that have issues too?

If you have a VTK build, you could also run the volume rendering tests

```
$ ctest -V -R vtkRenderingVolume*
```

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 5:04pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/9 "2018-01-16T17:04:50Z")

</div>

> [@lassoan](#):
>
> What is the size of your volume?

I’m using CTA-cardio.nrrd from your repo.

> [@lassoan](#):
>
> How much GPU memory usage do you allow in volume rendering settings?

On my work machine, 8 GB, corresponding to the VRAM of my GPU.  
On my laptop, 3GB : 1 GB for the iGPU and 2 GB for the dGPU.

> [@lassoan](#):
>
> ParaView

Downloaded PataView, but can’t find time now to learn how to create a Volume Rendering model there.

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 5:08pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/10 "2018-01-16T17:08:05Z")

</div>

> [@Sankhesh\_Jhaveri](#):
>
> run SimpleRayCast on the desktop with AMD GPU.

SimpleRayCast results on my work desktop PC with AMD RX480 GPU :

vtkFixedPointVolumeRayCastMapper : normal behaviour  
vtkGPUVolumeRayCastMapper : immediate UI freeze leading to hard reboot on first run, immediate UI freeze with only mouse active but useless on second run.

I’ll try again at home on my laptop when possible.

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [January 16, 2018, 5:09pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/11 "2018-01-16T17:09:50Z")

</div>

How much GPU memory do you specify in Slicer (Volume rendering module, Advanced section, Techniques tab)?

To test in ParaView: You can only load uncompressed volumes, so re-save CTA-cardio with compression disabled. Then, load the volume and in Properties/Display/Representation, choose “Volume”.

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 5:11pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/12 "2018-01-16T17:11:50Z")

</div>

> [@Sankhesh\_Jhaveri](#):
>
> you could also run the volume rendering tests

ctest -V -R vtkRenderingVolume\*  
UpdateCTestConfiguration from :/home/user/src/SimpleRayCast/DartConfiguration.tcl  
UpdateCTestConfiguration from :/home/user/src/SimpleRayCast/DartConfiguration.tcl  
Test project /home/user/src/SimpleRayCast  
Constructing a list of tests  
Updating test list for fixtures  
Added 0 tests to meet fixture requirements  
Checking test dependency graph…  
Checking test dependency graph end  
No tests were found!!!

Can’t understand what it means.

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [January 16, 2018, 5:14pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/13 "2018-01-16T17:14:47Z")

</div>

> [@chir.set](#):
>
> No tests were found!!!

By default, Slicer does not build test applications when VTK is built. In CMake’s VTK project, set BUILD\_TESTING variable to ON, then configure and build VTK. After that, volume rendering tests will be found.

---

<div class="post-metadata">

### Author: ![jcfr](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/jcfr/32/17825_2.png) [@jcfr](https://discourse.slicer.org/u/jcfr)
#### Post date: [January 16, 2018, 5:35pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/14 "2018-01-16T17:35:33Z")

</div>

> [@chir.set](#):
>
> SimpleRayCast results on my work desktop PC with AMD RX480 GPU :
> 
> vtkFixedPointVolumeRayCastMapper : normal behaviour
> 
> vtkGPUVolumeRayCastMapper : immediate UI freeze leading to hard reboot on first run, immediate UI freeze with only mouse active but useless on second run.

@Sankhesh_Jhaveri @lassoan Look like the example allows to reproduce the problem.

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 5:48pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/15 "2018-01-16T17:48:10Z")

</div>

> [@lassoan](#):
>
> Then, load the volume and in Properties/Display/Representation, choose “Volume”.

After accepting the dialog about time consuming process in ParaView, it starts computing and the UI freezes after about 5 secs. No model is displayed yet.

---

<div class="post-metadata">

### Author: ![ihnorton](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/ihnorton/32/9_2.png) [@ihnorton](https://discourse.slicer.org/u/ihnorton)
#### Post date: [January 16, 2018, 5:50pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/16 "2018-01-16T17:50:32Z")

</div>

> [@chir.set](#):
>
> I am using the amdgpu kernel module and the open source xf86-video-amdgpu driver.

Can you try the Radeon driver instead, as a differential? (or maybe you did already?)

[https://wiki.archlinux.org/index.php/AMD\_Catalyst](https://wiki.archlinux.org/index.php/AMD_Catalyst)

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 5:52pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/17 "2018-01-16T17:52:38Z")

</div>

> [@lassoan](#):
>
> After that, volume rendering tests will be found.

ctest -V -R vtkRenderingVolume\* &\> ~/tmp/ctest.log

It shows a few windows for less than 1 sec each, until the UI freezes. Mouse cursor is till available but useless. Here is what could be logged :

> UpdateCTestConfiguration from :/home/user/src/VTK-build/DartConfiguration.tcl  
> Parse Config file:/home/user/src/VTK-build/DartConfiguration.tcl  
> Add coverage exclude regular expressions.  
> Add coverage exclude: vtk[^.]+(Java|Python).cxx  
> Add coverage exclude: ._vtkOpenGLState._  
> Add coverage exclude: ._Testing.Cxx.cxx  
> Add coverage exclude: .Testing.Cxx.h  
> Add coverage exclude: .moc\_.cxx  
> Add coverage exclude: ./Rendering/OpenGL/vtkgl.  
> Add coverage exclude: ./Utilities/.  
> Add coverage exclude: ./ThirdParty/._  
> Add coverage exclude: ._vtkOpenGLPolyDataMapper._  
> SetCTestConfiguration:CMakeCommand:/usr/bin/cmake  
> UpdateCTestConfiguration from :/home/user/src/VTK-build/DartConfiguration.tcl  
> Parse Config file:/home/user/src/VTK-build/DartConfiguration.tcl  
> Test project /home/user/src/VTK-build  
> Constructing a list of tests  
> Done constructing a list of tests  
> Updating test list for fixtures  
> Added 0 tests to meet fixture requirements  
> Checking test dependency graph…  
> Checking test dependency graph end  
> test 41  
> Start 41: vtkRenderingVolume-HeaderTest
> 
> 41: Test command: /bin/python2 “/home/user/src/VTK/Testing/Core/HeaderTesting.py” “/home/user/src/VTK/Rendering/Volume” “VTKRENDERINGVOLUME\_EXPORT”  
> 41: Test timeout computed to be: 3600  
> 41: Use export macro: VTKRENDERINGVOLUME\_EXPORT  
> 1/89 Test #41: vtkRenderingVolume-HeaderTest … Passed 0.81 sec  
> test 353  
> Start 353: vtkRenderingVolumeOpenGL2-HeaderTest
> 
> 353: Test command: /bin/python2 “/home/user/src/VTK/Testing/Core/HeaderTesting.py” “/home/user/src/VTK/Rendering/VolumeOpenGL2” “VTKRENDERINGVOLUMEOPENGL2\_EXPORT”  
> 353: Test timeout computed to be: 3600  
> 353: Use export macro: VTKRENDERINGVOLUMEOPENGL2\_EXPORT  
> 2/89 Test #353: vtkRenderingVolumeOpenGL2-HeaderTest … Passed 0.19 sec  
> test 1208  
> Start 1208: vtkRenderingVolumeCxx-ProjectedTetrahedraZoomIn
> 
> 1208: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “ProjectedTetrahedraZoomIn” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary”  
> 1208: Test timeout computed to be: 3600  
> 1208: Loading /home/user/src/VTK-build/ExternalData//Testing/Data/ironProt.vtk  
> 1208: Loading /home/user/src/VTK-build/ExternalData//Testing/Data/neghip.slc  
> 1208: Generic Warning: In /home/user/src/VTK/Rendering/Volume/Testing/Cxx/ProjectedTetrahedraZoomIn.cxx, line 197  
> 1208: This test will always pass.  
> 1208:  
> 3/89 Test #1208: vtkRenderingVolumeCxx-ProjectedTetrahedraZoomIn … Passed 2.30 sec  
> test 1209  
> Start 1209: vtkRenderingVolumeCxx-TestFinalColorWindowLevel
> 
> 1209: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestFinalColorWindowLevel” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestFinalColorWindowLevel.png”  
> 1209: Test timeout computed to be: 3600  
> 1209: 0Standard0.182799  
> 1209: 0.342806  
> 4/89 Test #1209: vtkRenderingVolumeCxx-TestFinalColorWindowLevel … Passed 0.59 sec  
> test 1210  
> Start 1210: vtkRenderingVolumeCxx-TestFixedPointRayCastLightComponents
> 
> 1210: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestFixedPointRayCastLightComponents” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestFixedPointRayCastLightComponents.png”  
> 1210: Test timeout computed to be: 3600  
> 1210: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)  
> 1210: 0Standard0.152745  
> 1210: 0.236041  
> 5/89 Test #1210: vtkRenderingVolumeCxx-TestFixedPointRayCastLightComponents … Passed 0.77 sec  
> test 1211  
> Start 1211: vtkRenderingVolumeCxx-TestGPURayCastAdditive
> 
> 1211: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestGPURayCastAdditive” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestGPURayCastAdditive.png”  
> 1211: Test timeout computed to be: 3600  
> 1211: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)  
> 1211: 5.27712Standard0.062804  
> 1211: 0.128531  
> 6/89 Test #1211: vtkRenderingVolumeCxx-TestGPURayCastAdditive … Passed 0.34 sec  
> test 1212  
> Start 1212: vtkRenderingVolumeCxx-TestGPURayCastCompositeBinaryMask
> 
> 1212: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestGPURayCastCompositeBinaryMask” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestGPURayCastCompositeBinaryMask.png”  
> 1212: Test timeout computed to be: 3600  
> 1212: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)  
> 1212: 0.00261438Standard0.0450981  
> 1212: 0.061025  
> 7/89 Test #1212: vtkRenderingVolumeCxx-TestGPURayCastCompositeBinaryMask … Passed 0.33 sec  
> test 1213  
> Start 1213: vtkRenderingVolumeCxx-TestGPURayCastCompositeMaskBlend
> 
> 1213: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestGPURayCastCompositeMaskBlend” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestGPURayCastCompositeMaskBlend.png”  
> 1213: Test timeout computed to be: 3600  
> 1213: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)  
> 1213: 0Standard0.069319  
> 1213: 0.14196  
> 8/89 Test #1213: vtkRenderingVolumeCxx-TestGPURayCastCompositeMaskBlend … Passed 0.38 sec  
> test 1214  
> Start 1214: vtkRenderingVolumeCxx-TestGPURayCastCompositeMask
> 
> 1214: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestGPURayCastCompositeMask” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestGPURayCastCompositeMask.png”  
> 1214: Test timeout computed to be: 3600  
> 1214: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)  
> 1214: 0.392157Standard0.078526  
> 1214: 0.153189  
> 9/89 Test #1214: vtkRenderingVolumeCxx-TestGPURayCastCompositeMask … Passed 0.33 sec  
> test 1215  
> Start 1215: vtkRenderingVolumeCxx-TestGPURayCastCompositeToMIP
> 
> 1215: Test command: /home/user/src/VTK-build/bin/vtkRenderingVolumeCxxTests “TestGPURayCastCompositeToMIP” “-D” “/home/user/src/VTK-build/ExternalData//Testing” “-T” “/home/user/src/VTK-build/Testing/Temporary” “-V” “/home/user/src/VTK-build/ExternalData/Rendering/Volume/Testing/Data/Baseline/TestGPURayCastCompositeToMIP.png”  
> 1215: Test timeout computed to be: 3600  
> 1215: CTEST\_FULL\_OUTPUT (Avoid ctest truncation of output)

Now I have to finish my paper work !

---

<div class="post-metadata">

### Author: ![chir.set](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/chir.set/32/66982_2.png) [@chir.set](https://discourse.slicer.org/u/chir.set)
#### Post date: [January 16, 2018, 6:17pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/18 "2018-01-16T18:17:42Z")

</div>

> [@ihnorton](#):
>
> Can you try the Radeon driver instead, as a differential?

Well the RX480 GPU won’t work with the radeon kernel module. I’ll try that on my laptop ASAP.

(BTW, the page you referenced is much outdated. Catalyst is dead, Xorg is versioned to 1.1.19.)

---

<div class="post-metadata">

### Author: ![Sankhesh\_Jhaveri](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/sankhesh_jhaveri/32/1081_2.png) [@Sankhesh\_Jhaveri](https://discourse.slicer.org/u/Sankhesh_Jhaveri)
#### Post date: [January 17, 2018, 8:36pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/19 "2018-01-17T20:36:29Z")

</div>

We were able to test this on the same GPU on a mac with proprietary AMD drivers. No issues.  
It could be a driver issue.

There have been many changes to the volume rendering code in VTK that went into commit 69e26835e. Now that you have a VTK build, would you be able to run `git bisect` to isolate the VTK commit that introduced this issue?

---

<div class="post-metadata">

### Author: ![jcfr](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/jcfr/32/17825_2.png) [@jcfr](https://discourse.slicer.org/u/jcfr)
#### Post date: [January 17, 2018, 8:58pm UTC](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839/20 "2018-01-17T20:58:58Z")

</div>

Thanks

Here are more details about git bisect: [https://blog.kitware.com/regression-hunting-from-slicer-cdash-error-to-itk-master/](https://blog.kitware.com/regression-hunting-from-slicer-cdash-error-to-itk-master/)

And an example of test script that could be adapted: [https://gist.github.com/jcfr/6029698](https://gist.github.com/jcfr/6029698)

@Sankhesh_Jhaveri I guess the bisect test script could rebuild and only execute the test process `TestGPURayCastCompositeToMIP`. Then, if it doesn’t stop after a while, it could be killed if it doesn’t exit. Do you have test script that could be reused for this type of problem ?

but if the machine freeze, I guess he would have to be a manual bisect with a systematic restart of the machine …

[Next page](https://discourse.slicer.org/t/commit-69e26835e-would-burn-my-discrete-gpu/1839.md?page=2)
