Re: ScOpenclTest::testFinacialIRRFormula test error
On 09/20/2014 01:02 AM, Kohei Yoshida wrote: On Fri, 2014-09-19 at 10:51 +0200, Jakub Kościelak wrote: Hi all, first sorry for my english. When I tried to build LO 4.3.2.1 with --enable-dbgutil, error occurs in test: sc/qa/unit/opencl-test.cxx(1269) : error : Assertion Test name: ScOpenclTest::testFinacialIRRFormula double equality assertion failed - Expected: 29.3418442391175 - Actual : 523 - Delta : 0.00293418442391175 More output in attachment. Software: Windows 8.1, MSVS 2012 PRO, cygwin 32bit. I checked 4.2.6.3 and 4.3.2.2 and there is the same problem. Is this a bug in test case or in Calc ? Likely an issue with your GPU and/or driver for your GPU since that test computes formula results using your GPU. Markus was at one point trying to gather all problematic GPU's to add to our black list, and it's possible that your hardware needs to be on that list. He wrote a detector program you can run to get the detailed information of your GPU... Note that at least the MacOSX-10.8@21-10.7SDK tinderbox runs into exactly the same failure quite regularly, cf. e.g. recent http://tinderbox.libreoffice.org/cgi-bin/gunzip.cgi?tree=MASTERbrief-log=1411378801.4800#950. Stephan ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: ScOpenclTest::testFinacialIRRFormula test error
On Mon, 2014-09-22 at 13:21 +0200, Stephan Bergmann wrote: Note that at least the MacOSX-10.8@21-10.7SDK tinderbox runs into exactly the same failure quite regularly, cf. e.g. recent http://tinderbox.libreoffice.org/cgi-bin/gunzip.cgi?tree=MASTERbrief-log=1411378801.4800#950. Ok. Then maybe the test itself needs to be looked at, rather than blaming the GPU. I did once do a fix on similar issue. Maybe it's a good idea to do something similar for this assertion too. Kohei ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: ScOpenclTest::testFinacialIRRFormula test error
Thanks for Your answer.
My clinfo.log in attachment.
I think, We can add AMD/ATI Radeon HD 7970 with latest drivers to the
Black List:
Wersja pakietu sterowników14.10.1006-140417a-171099C
Wersja programu Catalyst14.4
DostawcaAdvanced Micro Devices, Inc.
Wersja sterownika 2D8.01.01.1390
Ścieżka pliku sterownika 2D
/REGISTRY/MACHINE/SYSTEM/ControlSet001/Control/Class/{4d36e968-e325-11ce-bfc1-08002be10318}/
Wersja biblioteki Direct3D9.14.10.01029
OpenGL - wersja6.14.10.12874
AMD Catalyst Control Center Wersja2014.0417.2226.38446
Wersja sterownika dźwięku AMD9.0.0.9905
Jakub
W dniu 2014-09-20 o 01:02, Kohei Yoshida pisze:
On Fri, 2014-09-19 at 10:51 +0200, Jakub Kościelak wrote:
Hi all,
first sorry for my english.
When I tried to build LO 4.3.2.1 with --enable-dbgutil, error
occurs in test:
sc/qa/unit/opencl-test.cxx(1269) : error : Assertion
Test name: ScOpenclTest::testFinacialIRRFormula
double equality assertion failed
- Expected: 29.3418442391175
- Actual : 523
- Delta : 0.00293418442391175
More output in attachment.
Software: Windows 8.1, MSVS 2012 PRO, cygwin 32bit.
I checked 4.2.6.3 and 4.3.2.2 and there is the same problem.
Is this a bug in test case or in Calc ?
Likely an issue with your GPU and/or driver for your GPU since that test
computes formula results using your GPU.
Markus was at one point trying to gather all problematic GPU's to add to
our black list, and it's possible that your hardware needs to be on that
list. He wrote a detector program you can run to get the detailed
information of your GPU...
I don't remember where to get that program though I need to dig up
the archive a bit to find out.
When I comment this test,
all other tests passed and the build is done.
Use this as a workaround for now. Markus is on vacation until
mid-October.
I build with dbgutil, because with this feature more tests are running,
am I right ?
It's unlikely that that would affect the result of that OpenCL test, but
maybe it does. I don't know.
Maybe I can run extended tests without this switch,
because I don't need all this debug stuff ?
That's a possibility. The vast majority of us recommend using dbgutil
turned on, but that also adds extra baggage that you may or may not need
(plus extra build time and disk space). So, if you are interested in
doing just casual coding, then using a non-dbgutil build may be an
option.
Kohei
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 1.2 AMD-APP (1445.5)
Platform Name: AMD Accelerated Parallel
Processing
Platform Vendor: Advanced Micro Devices, Inc.
Platform Extensions: cl_khr_icd
cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing
cl_amd_event_callback cl_amd_offline_devices cl_amd_hsa
Platform Name: AMD Accelerated Parallel
Processing
Number of devices: 2
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name:AMD Radeon HD 7900 Series
Device Topology: PCI[ B#1, D#0, F#0 ]
Max compute units: 32
Max work items dimensions: 3
Max work items[0]: 256
Max work items[1]: 256
Max work items[2]: 256
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int:1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double:1
Max clock frequency: 1100Mhz
Address bits: 32
Max memory allocation: 1073741824
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 8
Max image 2D width:16384
Max image 2D height: 16384
Max image 3D width:2048
Max image 3D
ScOpenclTest::testFinacialIRRFormula test error
Hi all, first sorry for my english. When I tried to build LO 4.3.2.1 with --enable-dbgutil, error occurs in test: sc/qa/unit/opencl-test.cxx(1269) : error : Assertion Test name: ScOpenclTest::testFinacialIRRFormula double equality assertion failed - Expected: 29.3418442391175 - Actual : 523 - Delta : 0.00293418442391175 More output in attachment. Software: Windows 8.1, MSVS 2012 PRO, cygwin 32bit. I checked 4.2.6.3 and 4.3.2.2 and there is the same problem. Is this a bug in test case or in Calc ? When I comment this test, all other tests passed and the build is done. I build with dbgutil, because with this feature more tests are running, am I right ? Maybe I can run extended tests without this switch, because I don't need all this debug stuff ? Regards, Jakub Kościelak F:/4321/sc/qa/unit/opencl-test.cxx(1269) : error : Assertion Test name: ScOpenclTest::testFinacialIRRFormula double equality assertion failed - Expected: 29.3418442391175 - Actual : 523 - Delta : 0.00293418442391175 Failures !!! Run: 223 Failure total: 1 Failures: 1 Errors: 0 warn:legacy.osl:10148:5868:unotools/source/config/configmgr.cxx:167: OSL_ASSERT: items_.empty() Error: a unit test failed, please do one of: export DEBUGCPPUNIT=TRUE# for exception catching export CPPUNITTRACE=gdb --args# for interactive debugging on Linux export CPPUNITTRACE=\[full path to devenv.exe]\ /debugexe # for interactive debugging in Visual Studio export VALGRIND=memcheck# for memory checking and retry using: make CppunitTest_sc_opencl_test make[1]: *** [F:/4321/workdir/CppunitTest/sc_opencl_test.test] Error 1 make: *** [build] Error 2 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: ScOpenclTest::testFinacialIRRFormula test error
On Fri, 2014-09-19 at 10:51 +0200, Jakub Kościelak wrote: Hi all, first sorry for my english. When I tried to build LO 4.3.2.1 with --enable-dbgutil, error occurs in test: sc/qa/unit/opencl-test.cxx(1269) : error : Assertion Test name: ScOpenclTest::testFinacialIRRFormula double equality assertion failed - Expected: 29.3418442391175 - Actual : 523 - Delta : 0.00293418442391175 More output in attachment. Software: Windows 8.1, MSVS 2012 PRO, cygwin 32bit. I checked 4.2.6.3 and 4.3.2.2 and there is the same problem. Is this a bug in test case or in Calc ? Likely an issue with your GPU and/or driver for your GPU since that test computes formula results using your GPU. Markus was at one point trying to gather all problematic GPU's to add to our black list, and it's possible that your hardware needs to be on that list. He wrote a detector program you can run to get the detailed information of your GPU... I don't remember where to get that program though I need to dig up the archive a bit to find out. When I comment this test, all other tests passed and the build is done. Use this as a workaround for now. Markus is on vacation until mid-October. I build with dbgutil, because with this feature more tests are running, am I right ? It's unlikely that that would affect the result of that OpenCL test, but maybe it does. I don't know. Maybe I can run extended tests without this switch, because I don't need all this debug stuff ? That's a possibility. The vast majority of us recommend using dbgutil turned on, but that also adds extra baggage that you may or may not need (plus extra build time and disk space). So, if you are interested in doing just casual coding, then using a non-dbgutil build may be an option. Kohei ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
