search for: livermoreloop

Hi Daniel, I'm trying to add LivermoreLoops test to the benchmark suite (tar ball attached), but I'm getting the error below: --- Tested: 2 tests -- FAIL: SingleSource/Benchmarks/LivermoreLoops/lloops.compile_time (1 of 2) FAIL: SingleSource/Benchmarks/LivermoreLoops/lloops.execution_time (2 of 2) When I use the option to only run thi...

...) functions so that seems expected to me. The compile failures end up getting buried in the logs, but they will either be in the test.log file in the top-level sandbox directory, or inside the <filename>.llvm.o.compile files in the Output directory for that test (i.e. SingleSource/Benchmarks/LivermoreLoops/Output/...). - Daniel On Mon, Nov 5, 2012 at 3:51 PM, Renato Golin <rengolin at systemcall.org>wrote: > Hi Daniel, > > I'm trying to add LivermoreLoops test to the benchmark suite (tar ball > attached), but I'm getting the error below: > > --- Tested: 2 tests -...

...seman commits the fast math patch we will be able to implement floating point reductions. That's great news! Attached is the whole benchmark, divided into 24 kernels and running on LNT with FP comparison and timings. Unpack the file onto SingleSource/Benchmarks and change the Makefile to add LivermoreLoops to the tests. Run the LNT tests with --only-test SingleSource/Benchmarks/LivermoreLoops to see it pass. The heuristics are dumb accumulations of the array values, just to get a value that will change considerably with small FP errors, so we can easily test and adjust how much FP error is being in...

Hi Renato! Thanks for working on this! It's really important to have more array-ish benchmarks. On Nov 16, 2012, at 12:28 PM, Renato Golin <rengolin at systemcall.org> wrote: > Daniel, Nadav, Hal, > > So, after some painstakingly boring re-formatting, I've split the 24 > kernels into 24 files (and left a horrible header file with code in > it, which I'll

Daniel, Nadav, Hal, So, after some painstakingly boring re-formatting, I've split the 24 kernels into 24 files (and left a horrible header file with code in it, which I'll clean up later). Since we're taking times in the benchmark tool, and we're trying to assert the quality of the FP approximation by the vectorization, I'll try to come up with a reasonable watermark for each

...SULT-exec-time: user 164.9400 TEST-RESULT-exec-real-time: real 165.3131 make[3]: Leaving directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/Linpack' make[3]: Entering directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/LivermoreLoops

...k the results and does report if they were as expected (not sure yet how that's calculated in detail). Nevertheless, should be good to have this test, at least to make sure we're not breaking floating point loops with vectorization in the future. Attached is a tar ball with the contents of LivermoreLoops to be included inside test-suite/SingleSource/Benchmarks. Daniel, can I just add this to the SVN repository, or are there other things that need to be done as well? It might need some care to fully use the testing infrastructure, though. cheers, --renato On 5 November 2012 22:12, Nadav Rotem &l...

...it should belong. :) > The compile failures end up getting buried in the logs, but they will either > be in the test.log file in the top-level sandbox directory, or inside the > <filename>.llvm.o.compile files in the Output directory for that test (i.e. > SingleSource/Benchmarks/LivermoreLoops/Output/...). Yes, I saw those logs and they were fine (3 warnings, 0 errors), so I got confused. Is there a place where you log the verbose Makefile output? That would have given me a better clue as to what was happening. cheers, --renato

That would be great! On Nov 5, 2012, at 2:11 PM, Renato Golin <rengolin at systemcall.org> wrote: > On 5 November 2012 17:41, Nadav Rotem <nrotem at apple.com> wrote: >> 1. We do not allow reductions on floating point types. We should allow them when unsafe-math is used. >> 2. All of the arrays are located in a struct. At the moment we don't detect that these

...box/bin/lnt runtest \ --submit ~/devel/llvm/test/perfdb \ nt \ -j8 \ --sandbox sandbox \ --test-suite ~/devel/llvm/test/test-suite \ --cc ~/devel/llvm/build/bin/clang \ --cxx ~/devel/llvm/build/bin/clang++ \ --only-test SingleSource/Benchmarks/LivermoreLoops Is that right? cheers, --renato -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20130108/dffedbd7/attachment.html>

On 5 November 2012 17:41, Nadav Rotem <nrotem at apple.com> wrote: > 1. We do not allow reductions on floating point types. We should allow them when unsafe-math is used. > 2. All of the arrays are located in a struct. At the moment we don't detect that these arrays are disjoin, and this prevents vectorization. Indeed, they look like simple changes. If no one is dying to get them

...164.9400 >> TEST-RESULT-exec-real-time: real 165.3131 >> make[3]: Leaving directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/Linpack' >> make[3]: Entering directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/LivermoreLoops >> _______________________________________________ >> LLVM Developers mailing list >> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

...e: user 164.9400 > TEST-RESULT-exec-real-time: real 165.3131 > make[3]: Leaving directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/Linpack' > make[3]: Entering directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/LivermoreLoops > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

...it ~/devel/llvm/test/perfdb \ > nt \ > -j8 \ > --sandbox sandbox \ > --test-suite ~/devel/llvm/test/test-suite \ > --cc ~/devel/llvm/build/bin/clang \ > --cxx ~/devel/llvm/build/bin/clang++ \ > --only-test SingleSource/Benchmarks/LivermoreLoops > > Is that right? > > cheers, > --renato

...-real-time: real 165.3131 >>> make[3]: Leaving directory >>> `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/Linpack' >>> make[3]: Entering directory >>> `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/LivermoreLoops >>> _______________________________________________ >>> LLVM Developers mailing list >>> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu >>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >

...gt; TEST-RESULT-exec-real-time: real 165.3131 >>>> make[3]: Leaving directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/Linpack' >>>> make[3]: Entering directory `/home/rkotler/llvmpb3/build/projects/test-suite/SingleSource/Benchmarks/LivermoreLoops >>>> _______________________________________________ >>>> LLVM Developers mailing list >>>> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu >>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >> >

Another issue is that the "fast nightly test" builders aren't working properly: they report success even though they failed (this started happening about one month ago, I don't know what the reason is). While this should be investigated, I'm tempted to just move everything over to LNT instead... Ciao, Duncan.

...if they > were as expected (not sure yet how that's calculated in detail). > Nevertheless, should be good to have this test, at least to make sure > we're not breaking floating point loops with vectorization in the > future. > > Attached is a tar ball with the contents of LivermoreLoops to be > included inside test-suite/SingleSource/Benchmarks. Daniel, can I just > add this to the SVN repository, or are there other things that need to > be done as well? It might need some care to fully use the testing > infrastructure, though. > > cheers, > --renato > &gt...

To weigh in here... On Thu, Jan 3, 2013 at 8:15 AM, David Blaikie <dblaikie at gmail.com> wrote: > +Daniel & Michael who work on the LNT infrastructure & might have some > thoughts on the differences & their merits & motivations. > > On Thu, Jan 3, 2013 at 4:05 AM, Renato Golin <renato.golin at linaro.org> > wrote: > > David, > > > >