| View Issue Details [ Jump to Notes ] | [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
| 0003743 | VTK | (No Category) | public | 2006-09-11 17:48 | 2006-10-10 10:52 | ||||
| Reporter | Sean McBride | ||||||||
| Assigned To | Mathieu Malaterre | ||||||||
| Priority | low | Severity | minor | Reproducibility | always | ||||
| Status | closed | Resolution | no change required | ||||||
| Platform | OS | OS Version | |||||||
| Product Version | |||||||||
| Target Version | Fixed in Version | ||||||||
| Summary | 0003743: dashboard: TestHyperOctreeDual fails on Mac on Intel | ||||||||
| Description | Every night, my Mac OS X build on i386 is failing the 'TestHyperOctreeDual' test. So is atamai's build, so I don't think its my setup. See: <http://www.vtk.org/Testing/Sites/stout.atamai/Darwin-i386-c++-Carbon/20060911-0300-Nightly/Results/__Graphics_Testing_Cxx_TestHyperOctreeDual.html> [^] <http://www.vtk.org/Testing/Sites/RogueResearchi386/MacOSX10.4.7-cocoa-gcc4.0.1-univ/20060911-1645-Experimental/Results/__Graphics_Testing_Cxx_TestHyperOctreeDual.html> [^] The test passes on ppc. There is no cross compilation involved. The failures occur on an intel machine running intel code. | ||||||||
| Tags | No tags attached. | ||||||||
| Project | |||||||||
| Type | |||||||||
| Attached Files | |||||||||
| Relationships | |
| Relationships |
| Notes | |
|
(0004935) Mathieu Malaterre (developer) 2006-09-16 11:08 |
I would say the OpenGL driver on x86 is not as robust as the one on ppc. Could you run the test using Mesa ? |
|
(0004959) Sean McBride (developer) 2006-09-18 11:41 |
Was this bug marked 'close' by accident? I assume so, since the bug is still there. I am not familiar with Mesa, how would I retest using Mesa? Do I just have to add some kind of VTK_MESA to the cmake file? I will do a few more tests to see if your hypothesis about it being a graphics driver bug is likely. |
|
(0004961) Sean McBride (developer) 2006-09-18 11:55 |
I have found that the "GraphicsCxxTests" executable contains the "TestHyperOctreeDual" test, and I've found the starting image at /VTKData/Baseline/Graphics/TestHyperOctreeDual.png. But how do I get the "TestImage" that is shown in the nightly builds? I'd like to run the "GraphicsCxxTests" on several different machine to compare output. Thanks. |
|
(0004963) Sean McBride (developer) 2006-09-18 14:33 |
OK, so the machine where it is failing has this GPU: Chipset Model: ATY,RadeonX1600 Bus: spdisplays_pcie_device VRAM (Total): 128 MB Vendor: ATI (0x1002) Device ID: 0x71c5 Revision ID: 0x0000 spdisplays_efi-version: 01.00.068 I tried TestHyperOctreeDual built on a ppc machine and it fails on the intel machine too (running in Rosetta). So this rules out a build problem and an endian problem. Mathieu, how do I test with Mesa? |
|
(0004970) 2006-09-19 12:14 |
OK, this is almost certainly a bug in 10.4.7. I tried in 10.5 and it works. Also, atamai tells me the test started failing for them when they went from 10.4.6 to 10.4.7! I'll file a bug with Apple. |
|
(0004971) Sean McBride (developer) 2006-09-19 12:16 |
OK, this is almost certainly a bug in 10.4.7. I tried in 10.5 and it works. Also, atamai tells me the test started failing for them when they went from 10.4.6 to 10.4.7! I'll file a bug with Apple. |
|
(0004993) Sean McBride (developer) 2006-09-21 10:14 |
Apple bug: <rdar://4743040> |
|
(0005089) Mathieu Malaterre (developer) 2006-10-02 19:03 |
So clearly people have to hope this is fixed in 10.4.8. |
|
(0005326) Sean McBride (developer) 2006-10-10 10:52 |
It is in fact fixed in 10.4.8! Goodie! |
| Notes |
| Issue History | |||
| Date Modified | Username | Field | Change |
| 2011-06-16 13:11 | Zack Galbreath | Category | => (No Category) |
| Issue History |
| Copyright © 2000 - 2018 MantisBT Team |