similar to: Benchmark LNT weird thread behaviour

Folks, I'm running the libc++ tests in tree [1] and I'm seeing two class of errors: 1. -lrt is used, but there is no such "librt.a" in the path. Results are that some symbols are not found, or that the DSO object wasn't. Adding a symlink in build/lib/librt.a to build/clang/3.8.0/lib/linux/libclang_rt.builtins-aarch64.a worked. 2. "collate_byname failed to construct

Hi, I'm new to llvm and am trying to run benchmarks from the test-suite using lnt to check loop-vectorization for various benchmarks. Test are compiling and executing fine, but I am not getting optimization remarks while using flags like -Rpass-missed=loop-vectorize and -Rpass-analysis=loop-vectorize I've tried running it like this: lnt runtest test-suite --sandbox SANDBOX --cc

Hi, So, I owe you all an apology. I was totally unaware that the llvm.org server, running at UIUC, was still in operation and hosting an active LNT instance. I was under the assumption the LNT server ran elsewhere. As a result, the ability for the LNT bots to submit run information to the "old" LNT server is now gone. I am very sorry for this oversight. In speaking with Chris

As for memory/storage, I think you want these presented as early as possible so people don't go setting up a box only to find out, half a week later, that they can never do what they intended to do (build Clang in debug mode). The reason I use 32-bit Python is that it is what you need, if you want to make a Windows buildbot slave because the Win32 bindings are needed and I don't think

Mikael, Thanks for working on. My gross comments... - Canonical URL of LLVM project is; http://llvm.org/ , not www.llvm.org. - "test-suite and cygwin" could be split out. It would be special thing. - Memory/storage requirements would be moved to "A.B Debugging". Less memory would be enough without debug build. - Python (x64 binary) is available, too. I am using. Python

Hi Logan, So, I just realised clang 3.8 comes with an unwind.h which doesn't have _URC_OK defined (introduced in r262178, just after 3.8 split). But on that commit, the personality routine depends on it for EHABI calls, which is defined by default on ARM environments. The end result is that I can only use Clang 3.9+ to compile libunwind 3.9+.

I want to get lnt to use qemu for the execution. In that case, RHOST= is not set. But I change the Arch because I am going to run in cross mode. Then I'm setting RUNUNDER to be a script which runs qemu. In this case it builds timeit-target as a Mips which fails because this is running on x86. ~/mysandbox/bin/lnt runtest nt --sandbox ~/mysandbox --cc /local/llvmpb_a/install/bin/clang

Hi, I'm trying to run the benchmark suite: http://llvm.org/docs/TestingGuide.html#test-suite-quickstart I'm doing it the lnt way, as described at: http://llvm.org/docs/lnt/quickstart.html I don't know what to expect but the results seems to be quite noisy and unstable. E.g I've done two runs on two different commits that only differ by a space in CODE_OWNERS.txt on my 12

I updated the docs to show a pip/svn install like this: pip install svn+http://llvm.org/svn/llvm-project/lnt/trunk <svn+http://llvm.org/svn/llvm-project/lnt/trunk> > On Jul 18, 2017, at 11:24 AM, Kristof Beyls <kristof.beyls at arm.com> wrote: > > >> On 18 Jul 2017, at 20:21, Daniel Dunbar <daniel at zuster.org <mailto:daniel at zuster.org>> wrote:

Dear, I am currently working on one specific architecture. When I make changes and try to "make" it takes a very long time as It compiles all the files in the source. My question is since I am only editing a few files in the architecture directory and some files in include directory, Can I speed up my "make" by running "make" only for a few files? Can such changes

Hi Chris, llvm.org/perf gives again error 500s. I remember you had a look the last time. Would you mind looking again? Also, is there something which we could do to make LNT more robust. Do you have an idea what the last issue was about? Thanks, Tobias -------- Original Message -------- Subject: buildbot failure in LLVM on polly-perf-O3-polly-scev-codegen-isl Date: Fri, 21 Feb 2014 17:08:44

The issues seems to be this line in the tools Makefile timeit-target: timeit.c $(LD_ENV_OVERRIDES) $(LCC) -o $@ $< $(LDFLAGS) $(CFLAGS) $(TARGET_FLAGS) -O3 It should not add target flags if we are simulating the target on the host. On 06/06/2013 06:59 PM, reed kotler wrote: > I want to get lnt to use qemu for the execution. > > In that case, RHOST= is not set. > > But I

On 1 July 2013 02:02, Chris Matthews <chris.matthews at apple.com> wrote: > One thing that LNT is doing to help “smooth” the results for you is by > presenting the min of the data at a particular revision, which (hopefully) > is approximating the actual runtime without noise. > That's an interesting idea, as you said, if you run multiple times on every revision. On ARM,

Two other things: 1) I get massively more stable execution times on 16.04 than on 14.04 on both x86 and ARM because 16.04 does far fewer gratuitous moves from one core to another, even without explicit pinning. 2) turn off ASLR: "echo 0 > /proc/sys/kernel/randomize_va_space". As well as getting stable addresses for debugging repeatability, it also stabilizes execution time

> On Feb 27, 2017, at 1:36 AM, Kristof Beyls via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > Hi Mikael, > > Some noisiness in benchmark results is expected, but the numbers you see seem to be higher than I'd expect. > A number of tricks people use to get lower noise results are (with the lnt runtest nt command line options to enable it between brackets): > *

I'd be fine pulling the pypi (sic) entry, shall I make it so? - Daniel On Tue, Jul 18, 2017 at 11:06 AM Chris Matthews <chris.matthews at apple.com> wrote: > Yep, we should pull that package down, or start to update it. > > The docs builds work for me. Is it possible that is not in your venv or > something like that? > > > On Jul 18, 2017, at 10:48 AM, Matthias

LNT executes: cc_version = capture([cc, '-v', '-E'] + cc_flags + ['-x', 'c', '/dev/null', '-###'], include_stderr=True).strip() Then parses the result Perhaps that output has changed in an unexpected way? > On Aug 6, 2014, at 3:06 PM, Chris Matthews <chris.matthews at apple.com> wrote:

Hi folks, It seems all LNT buildbots are broken due to the error: error: The 'six==1.9.0' distribution was not found and is required by LNT The module seems to be in version 1.10 and probably the old binary is no longer there. Do we really need it to be 1.9? Can't that be >1.9? cheers, --renato

So you're referring to llvm_test_prepare? There aren't many examples of that being used. I tried adding a call to it in my test directory's CMakeLists.txt like so: llvm_test_prepare(WORKDIR ${CMAKE_CURRENT_BINARY_DIR} export MY_VAR=42 ) In the hopes that something would happen, even an error, but it had no effect. Nothing at http://llvm.org/docs/lnt/tests.html explains how to

I will take a peek at the server log. On Jan 16, 2014, at 4:09 PM, Tobias Grosser <tobias at grosser.es> wrote: > On 01/17/2014 01:07 AM, Tobias Grosser wrote: >> On 01/17/2014 12:59 AM, Chris Matthews wrote: >>> Do you guys have examples of a LNT results file that causes the failure? >> >> Hi Chris, >> >> thanks for the quick reply. Here a link