search for: fpcmp

..."Hal Finkel" <hfinkel at anl.gov> > > Cc: "Alexey Samsonov" <samsonov at google.com>, "LLVM Developers Mailing > List" <llvmdev at cs.uiuc.edu> > > Sent: Thursday, May 1, 2014 1:56:04 AM > > Subject: Re: [LLVMdev] Is it ok to nuke fpcmp and llvm-PerfectShuffle > utils? > > > > >> Ouch, indeed... Is it ok it has no build rules? > > > > > > I think that it should have build rules. If nothing else, can you > > > file a bug report on this? > > It is building using configure + make....

...undefined reference to `_imp__SymGetSymFromAddr at 16' e:/prg/vc/llvm-cvs-mingw/llvm/mingw-obj/Release/lib/libLLVMSystem.a(Signals.o)(.text+0x79b):Signals.cpp: undefined reference to `_imp__SymGetLineFromAddr at 16' collect2: ld returned 1 exit status llvm[2]: Linking Release executable fpcmp (without symbols) e:/prg/vc/llvm-cvs-mingw/llvm/mingw-obj/Release/lib/libLLVMSystem.a(Process.o)(.text+0x8d):Process.cpp: undefined reference to `GetProcessMemoryInfo at 12' collect2: ld returned 1 exit status * make TOOLS_VERBOSE=1 output: g++ -I/e/prg/vc/llvm-cvs-mingw/llvm/mingw-obj/utils/...

The first one failed on a diff: ******************** TEST (simple) 'sse.expandfft' FAILED! ******************** Execution Context Diff: /home/rkotler/llvmpb3/build/projects/test-suite/tools/fpcmp: Compared: 1.139094e-07 and 1.159249e-07 abs. diff = 2.015500e-09 rel.diff = 1.738626e-02 Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 ******************** TEST (simple) 'sse.expandfft' **************************** The second one too. /llvmpb3/build/projects/test-suite/tools/fpc...

...char*)': llvm-3.1/llvm/lib/Support/DynamicLibrary.cpp:105:1: warning: control reaches end of non-void function [-Wreturn-type] I don't remember how I fixed that. dlfcn.h was already included in DynamicLibrary.cpp. Anyway, somehow I managed. Now building llvm stops at the point of linking fpcmp. It compiles fpcmp.cpp alright, but make comes with the following messages: llvm[2]: Compiling fpcmp.cpp for Release build llvm[2]: Linking Release executable fpcmp (without symbols) ld: 0706-027 The -R $ORIGIN/../lib flag is ignored. ld: 0706-027 The -R /ccvunix/cdndev/user/cdndadm/Tools/llvm-3.0...

>> Ouch, indeed... Is it ok it has no build rules? > > I think that it should have build rules. If nothing else, can you file a bug report on this? It is building using configure + make. In good old days the perfect shuffle tables for PPC were generated through the normal build process like .td files. However, this was pretty useless - they do not change and thus perfectshuffle was

...ote: > ----- Original Message ----- > > From: "Alexey Samsonov" <samsonov at google.com> > > To: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> > > Sent: Wednesday, April 30, 2014 3:30:12 PM > > Subject: [LLVMdev] Is it ok to nuke fpcmp and llvm-PerfectShuffle utils? > > > > > > > > They are not used in the regression test suite anymore, and are not > > even built in CMake build tree. > > You cannot nuke llvm-PerfectShuffle... that is the utility used to > generate the lib/Target/ARM64/ARM64P...

They are not used in the regression test suite anymore, and are not even built in CMake build tree. -- Alexey Samsonov, Mountain View, CA -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20140430/c015da10/attachment.html>

...ng filtered for known failures. On 12/12/2012 05:03 PM, reed kotler wrote: > The first one failed on a diff: > ******************** TEST (simple) 'sse.expandfft' FAILED! > ******************** > Execution Context Diff: > /home/rkotler/llvmpb3/build/projects/test-suite/tools/fpcmp: Compared: > 1.139094e-07 and 1.159249e-07 > abs. diff = 2.015500e-09 rel.diff = 1.738626e-02 > Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 > ******************** TEST (simple) 'sse.expandfft' > **************************** > > The second one too. > > /...

...n failures. > > On 12/12/2012 05:03 PM, reed kotler wrote: >> The first one failed on a diff: >> ******************** TEST (simple) 'sse.expandfft' FAILED! ******************** >> Execution Context Diff: >> /home/rkotler/llvmpb3/build/projects/test-suite/tools/fpcmp: Compared: 1.139094e-07 and 1.159249e-07 >> abs. diff = 2.015500e-09 rel.diff = 1.738626e-02 >> Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 >> ******************** TEST (simple) 'sse.expandfft' **************************** >> >> The second one too. &g...

...built the latest llvm/llvm-gcc yesterday and got a ton of test failures that look like the following ([i686-off-opt] means an i686 Release build of official -- unmodified -- sources): [i686-off-opt]: ./llvm.official/projects/llvm-test/DiffOutput.sh "/ptmp/dag/build.llvm.official/Release/bin/fpcmp -r 0.016" llc sse.isamax [i686-off-opt]: ******************** TEST (llc) 'sse.isamax' FAILED! ******************** [i686-off-opt]: Execution Context Diff: [i686-off-opt]: /ptmp/dag/build.llvm.official/Release/bin/fpcmp: FP Comparison failed, not a numeric difference between 'e&...

I'm really sorry for the wide e-mail, but I've exhausted the normal channels. I get: llvm[2]: Linking Release executable tblgen (without symbols) llvm[2]: ======= Finished Linking Release Executable tblgen (without symbols) llvm[2]: Compiling fpcmp.cpp for Release build llvm[2]: Linking Release executable fpcmp (without symbols) llvm[2]: ======= Finished Linking Release Executable fpcmp (without symbols) llvm[2]: Compiling PerfectShuffle.cpp for Release build llvm[2]: Linking Release executable llvm-PerfectShuffle (without symbols) llvm[2...

.../===----------------------------------------------------------------------===// //=== WARNING: Implementation here must contain only generic UNIX code that Then I start fixing up link errors like the following that now require -lbsd: make[2]: Entering directory `/mnt/ssd/work/svn/lgs/build/utils/fpcmp' llvm[2]: Linking Debug+Asserts executable fpcmp /mnt/ssd/work/svn/lgs/build/Debug+Asserts/lib/libLLVMSupport.a(Process.o): In function `llvm::sys::Process::GetRandomNumber()': /mnt/ssd/work/svn/lgs/llvm/lib/Support/Unix/Process.inc:371: undefined reference to `arc4random' collect2: err...

...et-independent configuration. Maybe we could do this: - Have a reference output with -ffp-contract=off - Run program compiled with -ffp-contract=off and the default - Compare the former against the reference output for validation - Compare the latter (the default build) against the former with fpcmp with some tolerance This also has the advantage that we still only need the hashed reference outputs in the repository. -Hal > -- Hal Finkel Lead, Compiler Technology and Programming Languages Leadership Computing Facility Argonne National Laboratory

...5 > > 74c74 > > < TR=0.79, TI=0.16, P0=7894.33, Q0=1592.81 > > --- > > > TR=0.79, TI=0.16, P0=7894.32, Q0=1592.81 > > This works on Linux (it did last night) so not sure why FreeBSD gives > these rounding errors. Can you investigate? Does the test use fpcmp? It does: /usr/home/llvm/obj/../test/Programs/RunSafely.sh 500 /dev/null Output/power.out-llc Output/power.llc /usr/home/llvm/obj/../test/Programs/DiffOutput.sh "/usr/home/llvm/obj/tools/Debug/fpcmp -r 0.00001" llc power /usr/home/llvm/obj/../test/Programs/RunSafely.sh 500 /dev/null...

Hi, I'm able to build the llvm tools on the MinGW platform: burg, fpcmp, tblgen, llvm-as, llvm-dis, opt, gccas, llc, llvm-link, lli, gccld, llvm-stub, analyze and extract. I wonder if these tools are sufficient to start build the cfrontend? Henrik. _________________________________________________________________ Undg� pop-ups med MSN Toolbar - http://toolbar.msn...

...x00002b062061fd4a in __cxa_throw () from /usr/lib64/libstdc++.so.6 [x86_64-off-opt]: #6 0x0000000000400603 in main () [x86_64-off-opt]: /ptmp/dag/llvm-project.official/llvm/trunk/projects/test- suite/DiffOutput.sh "/ptmp/dag/build.llvm.trunk.official.opt/x86_64-unknown- linux-gnu/Release/bin/fpcmp " cbe simple_throw [x86_64-off-opt]: ******************** TEST (cbe) 'simple_throw' FAILED! ******************** [x86_64-off-opt]: Execution Context Diff: [x86_64-off-opt]: /ptmp/dag/build.llvm.trunk.official.opt/x86_64-unknown- linux-gnu/Release/bin/fpcmp: FP Comparison failed, not...

...y "./configure CC=clang CXX=clang" but get link > > errors. > > Try using CXX=clang++ instead. Always use clang++ for C++ code. :) > clang++ does not exist. I tried "CXX=clang -x c++" but fail with errors, clang is trying to compile .o files. "llvm/utils/fpcmp/Release/fpcmp.o:4:2235: warning: ull character ignored" with ./configure --enable-optimized --with-clang=/usr/bin/clang CC=/usr/bin/clang CXX="/usr/bin/clang -x c++" Tried clean svn, but fails too. Did I something wrong?

I forgot to mention that you can also run "make TEST=simple report" which will generate a nice report. Do you know why these tests fail ? You can step into the test directory and run 'make TEST=simple' from there. It will save you some time. On Dec 12, 2012, at 4:04 PM, reed kotler <rkotler at mips.com> wrote: > I'm getting three failures. > > TEST-FAIL:

...port.txt 2>&1 > > That does indeed seem to get me further. Thanks! Failures below. -Dave ******************** TEST (simple) 'sse.expandfft' FAILED! ******************** Execution Context Diff: /home/greened/src/llvm-30-rc3/build/Release/bin/fpcmp: Compared: 1.139094e-07 and 1.159249e-07 abs. diff = 2.015500e-09 rel.diff = 1.738626e-02 Out of tolerance: rel/abs: 1.600000e-02/0.000000e+00 TEST-FAIL: exec /home/greened/src/llvm-30-rc3/build/projects/llvm-test/SingleSource/UnitTests/Vector/SSE/sse.expandfft ******************** TEST (simple...

Nikola Smiljanic <popizdeh at gmail.com> writes: > I'm not sure if this makes a difference but I ran this > > make -j 2 -k LLVMCC_OPTION=clang ENABLE_BUILT_CLANG=1 ENABLE_PARALLEL_REPORT=1 TEST=simple report > ../../report.txt 2>&1 That does indeed seem to get me further. Thanks! -Dave