[Rd] Matrix issues when building R with znver3 architecture under GCC 11

Tomas Kalibera tom@@@k@||ber@ @end|ng |rom gm@||@com
Tue Apr 26 09:18:54 CEST 2022


Hi Kieran,

On 4/26/22 06:03, Kieran Short wrote:
> Dear Tomas,
>
> Thanks once again for your insight. I'll take all this on board.
> I'll have a poke around to see what's up with Matrix, but I really 
> don't have time to dig deep.
> However, I'm curious. Assuming I have the necessary resources, how do 
> we check against all CRAN contributed packages - as the dev team does? 
> Is there any advice, documentation, or scripts about how one goes 
> about doing that?

certainly this is something that needs time to set up and tune to run at 
least somewhat reliably. With nearly 19,000 packages, you always get 
some failures (intermittent - when remote systems are down, 
non-deterministic - when there are non-deterministic bugs, stable - when 
packages simply need to be updated). Sometimes some packages lock-up 
during checking (or run into an infinite loop), etc. Some packages may 
cause corruption of the package library, which needs extra precautions. 
Of course you need to install external software for the packages, you 
need to build the right version of R, set up a local mirror of CRAN, etc.

So again it depends on what you want to do. CRAN systems primarily look 
for bugs in packages and some also build binaries (of packages or/and 
R). My set of such scripts (though for experimental builds and checks, 
not regular operations) is under the dev svn 
(https://svn.r-project.org/R-dev-web/trunk/WindowsBuilds/winutf8/ucrt3/) 
with some comments  at 
https://developer.r-project.org/WindowsBuilds/winutf8/ucrt3/howto.html. 
But this is for Windows. You will find several other sets in the dev 
svn, for other platforms. These sets differ as they do slightly 
different things (mine are for testing Rtools). A perhaps surprising 
challenge is to implement timeouts so that they are reliable. This is a 
moving target - new versions of packages sometimes come up with code 
that locks up systems in new ways. And it is platform-specific. 
Inevitably, setting such a system also requires debugging packages, 
because you want to have confidence that the failures are not caused by 
limitations in the scripts - and this is again a moving target as 
packages evolve (different versions of external software may wake up 
bugs, even different depth of the directory tree you run in, etc) and as 
external software evolves.

Then you may want to use CRAN packages as a regression test for R. For 
that, you would want to ensure testing exactly the same set of packages 
(protect library against corruption). And then you need to handle 
intermittent issues due to external sites going up and down (I do that 
by iterating the old and the new version). On the other hand, you would 
probably give up on debugging packages failing on both systems ("before" 
and "after") if they are not too many. This is what one could do to try 
to see if a C compiler optimization causes new crashes/errors/warnings.

Then you may want to do timing. It makes sense to select a subset of 
packages for this, and their long-running examples, such that they do 
not download anything, they are stable (no non-deterministic issues), 
have no warnings or errors (so that you are not measuring error paths), 
they do not lock-up the systems (no timeouts needed). You need to run 
them sequentially, with some repetitions. But you can avoid some of the 
problems above by simply not including packages causing different kinds 
of trouble. One challenge is that if after more repetitions you find 
some non-deterministic failure, you should be able to remove packages 
from the measurements.

And then there are variants of the above. Such as using bug-finding 
tools to find bugs in packages.

So, you may find some scripts in the dev svn, with some comments in 
them, to get some inspiration, but I know it will be a lot of work to 
set it up. An obvious obstacle is that these things run long - a 
debugging turnaround for scripts/set up is long.

If you want to do just some one-off, partially manual experiment, on a 
subset of packages that seem to work for you, that would be easier to do 
something very simple from scratch.

>
> For now, I'm running some lengthy scripts at the moment that require a 
> large number of packages with many dependencies. With this, I hope to 
> both check the speed differences between the different builds and any 
> differences in their outputs.

This may be better anyway as they I assume represent more the workload 
you have. But then it depends on your goal (if you need more specific 
advice, please define it).

But, as I wrote before, what would be really helpful for others would be 
if you could narrow down the issue to a tiny reproducible example, so 
that it could be narrowed down and fixed (in Matrix, in GCC, etc). 
Without that, I'd not recommend using the optimized builds, anyway, 
regardless of performance impacts.

Also, if you don't have time to do the measuring etc, if you don't have 
a specific workload with performance issues at hand, I would simply give 
up on it and use the standard builds. The chances this will help a lot 
are small. Measuring this carefully requires a lot of effort. It is a 
long time ago, but I remember a solid evaluation study showing on 
standard application benchmarks, via very careful measurements, that -O3 
optimizations didn't produce faster code than -O2.

Best
Tomas

>
> best regards,
> Kieran
>
> On Wed, Apr 13, 2022 at 8:26 PM Tomas Kalibera 
> <tomas.kalibera using gmail.com> wrote:
>
>
>     On 4/13/22 11:20, Kieran Short wrote:
>>     Hi Tomas,
>>
>>     Many thanks for your thorough response, it is very much
>>     appreciated and what you say makes perfect sense to me.
>>
>>     I was relying on the in-built R compilation checks, I have been
>>     working on the assumption that everything on the R side is
>>     correct (including the matrix package).
>>
>>     Indeed, R 4.1.3 builds and "make check-all" passes with the more
>>     general -march=x86-64 architecture compiled with -O3
>>     optimizations (in my hands, on the Zen3 system). So I had no
>>     underlying reason not to believe R or its packages were the
>>     problem when -march=znver3 was trialed. I found it interesting
>>     that it was only the one factorizing.R script in the Matrix suite
>>     that failed (out of the seemingly hundreds of remaining checks
>>     overall which passed). So I was more wondering if there might
>>     have been prior knowledge within the brain's trust on this list
>>     that "oh the factorizing.R matrix test does ABC error when R or
>>     the package is compiled with GCC using XYZ flags". As you'll read
>>     ahead, you can say that now. :)
>     Right, but something must be broken. You might get specific
>     comments from the Matrix package maintainer, but it would help at
>     least minimizing that failing example to some commands you can run
>     in R console, and showing the differences in outputs.
>>
>>     I don't think I have the capability to determine the root trigger
>>     in R itself, the package, or the compiler (whichever one, or
>>     combination,  it actually is). However, assuming R isn't the
>>     issue, I have done is go through the GCC optimizations and I have
>>     now isolated the culprit optimization which crashes factorizing.R.
>>
>>     It is "-fexpensive-optimizations".
>>
>>     If I use "-fno-expensive-optimizations" paired with -O2 or -O3
>>     optimizations, all "make check-all" checks pass. So I can build a
>>     fully checked and passed R 4.1.3 under my environment now with:
>>
>>     ~/R/R-4.1.3/configure CC=gcc-11.2 CXX=g++-11.2 FC=gfortran-11.2
>>     CXXFLAGS="-O3 -march=znver3 -fno-expensive-optimizations -flto"
>>     CFLAGS="-O3 -march=znver3 -fno-expensive-optimizations -flto"
>>     FFLAGS="-O3 -march=znver3 -fno-expensive-optimizations -flto"
>>     --enable-memory-profiling --enable-R-shlib
>     Ok. The default optimization options used by R on selected current
>     and future versions of GCC and clang also get tested via checking
>     all of CRAN contributed packages. This testing sometimes finds
>     errors not detected by "make check-all", including bugs in GCC.
>     You would need a lot of resources to run these checks, though. In
>     my experience it is not so rare that a bug (in R or GCC) only
>     affects a very small number of packages, often even only one.
>>     I'm yet to benchmark whether the loss of that particular
>>     optimization flag negates the advantages of using znver3 as a
>>     core architecture target over a -x86-64 target in the first place.
>>     So I think I've solved my own problem (at least, it appears that
>>     way based on the checks).
>>     So the remaining question is, what method or package does the
>>     development team use (if any) for testing the speed of various
>>     base R calculations?
>
>     That depends on the developer and the calculations, and on your
>     goals - what you want to measure or show. I don't have a simple
>     advice. If you are considering this for your own work, I'd
>     recommend measuring some of your workloads. Also you can
>     extrapolate from your workloads (from where time is spent in them)
>     what would be a relevant benchmark. For example, if most time is
>     spent in BLAS, then it is about finding a good optimized
>     implementation (and for that checking the impact of the
>     optimizations). Similarly, if it is some R package (base,
>     recommended, or contributed), it may be using a computational
>     kernel written in C or Fortran, something you could test
>     separately or with a specific benchmark. I think it would be
>     unlikely that CPU-specific C compiler optimizations would
>     substantially speed up the R interpreter itself.
>
>     For just deciding whether -fno-expensive-optimization negates the
>     gains, you might look at some general computational/other
>     benchmarks (not R). If it negated it even on benchmarks used by
>     others to present the gains, then it probably is not worth it.
>
>     One of the things I did in the past was looking at timings of
>     selected CRAN packages (longer running examples, packages with
>     most reverse dependencies) and then looking into the reasons for
>     the individual bigger differences. That was when looking at the
>     impacts of the byte-code compiler. Unlikely worth the effort in
>     this case. Also, primarily, I think the bug should be traced down
>     and fixed, wherever it is. Only then the measuring would make sense.
>
>     Best
>     Tomas
>
>
>
>>
>>     best regards,
>>     Kieran
>>
>>     On Wed, Apr 13, 2022 at 4:00 PM Tomas Kalibera
>>     <tomas.kalibera using gmail.com> wrote:
>>
>>         Hi Kieran,
>>
>>         On 4/12/22 02:36, Kieran Short wrote:
>>         > Hello,
>>         >
>>         > I'm new to this list, and have subscribed particularly
>>         because I've come
>>         > across an issue with building R from source with an
>>         AMD-based Zen
>>         > architecture under GCC11. Please don't attack me for my
>>         linux operating
>>         > system choice, but it is Ubuntu 20.04 with Linux Kernel
>>         5.10.102.1 -
>>         > microsoft-standard-WSL2. I've built GCC11 using GCC8 (the
>>         standard GCC
>>         > under Ubuntu20.04 WSL release), under Windows11 with wslg.
>>         WSL2/g runs as a
>>         > hypervisor with ports to all system resources including
>>         display, GPU (cuda,
>>         > etc).
>>         >
>>         > The reason why I am posting this email is that I am trying
>>         to compile R
>>         > using the AMD Zen3 platform architecture rather than
>>         x86/64, because it has
>>         > processor-specific optimizations that improve performance
>>         over the standard
>>         > x86/64 in benchmarks. The Zen3 architecture optimizations
>>         are not available
>>         > in earlier versions of GCC (actually, they have possibly
>>         been backported to
>>         > GCC10 now). Since Ubuntu 20.04 doesn't have GCC11, I
>>         compiled the GCC11
>>         > compiler using the native GCC8.
>>         >
>>         > The GCC11 I have built can build R 4.1.3 with a standard x86-64
>>         > architecture and pass all tests with "make check-all".
>>         > I configured that with:
>>         >> ~/R/R-4.1.3/configure CC=gcc-11.2 CXX=g++-11.2
>>         FC=gfortran-11.2
>>         > CXXFLAGS="-O3 -march=x86-64" CFLAGS="-O3 -march=x86-64"
>>         FFLAGS="-O3
>>         > -march=x86-64" --enable-memory-profiling --enable-R-shlib
>>         > and built with
>>         >> make -j 32 -O
>>         >> make check-all
>>         > ## PASS.
>>         >
>>         > So I can build R in my environment with GCC11.
>>         > In configure, I am using references to "gcc-11.2"
>>         "gfortran-11.2" and
>>         > "g++-11.2" because I compiled GCC11 compilers with these
>>         suffixes.
>>         >
>>         > Now, I'm using a 32 thread (16 core) AMD Zen3 CPU (a
>>         5950x), and want to
>>         > use it to its full potential. Zen3 optimizations are
>>         available as a
>>         > -march=znver3 option n GCC11. The znver3 optimizations
>>         improve performance
>>         > in Phoronix Test Suite benchmarks (I'm not aware of anyone
>>         that has
>>         > compiled R with them). See:
>>         >
>>         https://www.phoronix.com/scan.php?page=article&item=amd-5950x-gcc11
>>         <https://www.phoronix.com/scan.php?page=article&item=amd-5950x-gcc11>
>>         >
>>         > However, the R 4.1.3 build (made with "make -j 32 -O"),
>>         configured with
>>         > -march=znver3, produces an R that fails "make check-all".
>>         >
>>         >> ~/R/R-4.1.3/configure CC=gcc-11.2 CXX=g++-11.2
>>         FC=gfortran-11.2
>>         > CXXFLAGS="-O2 -march=znver3" CFLAGS="-O2 -march=znver3"
>>         FFLAGS="-O2
>>         > -march=znver3" --enable-memory-profiling --enable-R-shlib
>>         > or
>>         >> ~/R/R-4.1.3/configure CC=gcc-11.2 CXX=g++-11.2
>>         FC=gfortran-11.2
>>         > CXXFLAGS="-O3 -march=znver3" CFLAGS="-O3 -march=znver3"
>>         FFLAGS="-O3
>>         > -march=znver3" --enable-memory-profiling --enable-R-shlib
>>         >
>>         > The fail is always in the factorizing.R Matrix.R tests, and
>>         in particular,
>>         > there are a number of errors and a fatal error.
>>         > I have attached the output because I cannot really
>>         understand what is going
>>         > wrong. But results returned from matrix calculations are
>>         obviously odd with
>>         > -march=znver3 in GCC 11. There is another
>>         backwards-compatible architecture
>>         > option "znver2" and this has EXACTLY the same result.
>>         >
>>         > While there are other warrnings and errors (many in
>>         assert.EQ() ), the
>>         > factorizing.R script continues. The fatal error (at line
>>         2662 in the
>>         > attached factorizing.Rout.fail text file) is:
>>         >
>>         >> ## problematic rank deficient rankMatrix() case -- only
>>         seen in large
>>         > cases ??
>>         >> Z. <- readRDS(system.file("external", "Z_NA_rnk.rds",
>>         package="Matrix"))
>>         >> tools::assertWarning(rnkZ. <- rankMatrix(Z., method =
>>         "qr")) # gave errors
>>         > Error in assertCondition(expr, classes, .exprString = d.expr) :
>>         >    Failed to get warning in evaluating rnkZ. <-
>>         rankMatrix(Z., method  ...
>>         > Calls: <Anonymous> -> assertCondition
>>         > Execution halted
>>         >
>>         > Can anybody shed light on what might be going on here?
>>         'make check-all'
>>         > passes all the other checks. It is just factorizing.R in
>>         Matrix that fails
>>         > (other matrix tests run ok).
>>         > Sorry this is a bit long-winded, but I thought details
>>         might be important.
>>
>>         R gets used and tested most with the default optimizations,
>>         without use
>>         of model-specific instructions and with -O2 (GCC). It happens
>>         time to
>>         time that some people try other optimization options and run
>>         into
>>         problems. In principle, there are these cases (seen before):
>>
>>         (1) the test in R package (or R) is wrong - it
>>         (unintentionally) expects
>>         behavior which has been observed in builds with default
>>         optimizations,
>>         but is not necessarily the only correct one; in case of
>>         numerical
>>         tolerances set empirically, they could simply be too tight
>>
>>         (2) the algorithm in R package or R has a bug - the result is
>>         really
>>         wrong and it is because the algorithm is (unintentionally)
>>         not portable
>>         enough, it (unintentionally) only works with default
>>         optimizations or
>>         lower; in case of numerical results, this can be because it
>>         expects more
>>         precision from the floating point computations than mandated
>>         by IEEE, or
>>         assumes behavior not mandated
>>
>>         (3) the optimization by design violates some properties the
>>         algorithm
>>         knowingly depends on; with numerical computations, this can
>>         be a sort of
>>         "fast" (and similarly referred to) mode which violates IEEE
>>         floating
>>         point standard by design, in the aim of better performance;
>>         due to the
>>         nature of the algorithm depending on IEEE, and poor luck, the
>>         results
>>         end up completely wrong
>>
>>         (4) there is a bug in the C or Fortran compiler (GCC as we
>>         use GCC) that
>>         only exhibits with the unusual optimizations; the compiler
>>         produces
>>         wrong code
>>
>>         So, when you run into a problem like this and want to get
>>         that fixed,
>>         the first thing is to identify which case of the above it is,
>>         in case of
>>         1 and 2 also differentiate between base R and a package (and
>>         which
>>         concrete package). Different people maintain these things and
>>         you would
>>         ideally narrow down the problem to a very small, isolated,
>>         reproducible
>>         example to support your claim where the bug is. If you do
>>         this right,
>>         the problem can often get fixed very fast.
>>
>>         Such an example for (1) could be: few lines of standalone R
>>         code using
>>         Matrix that produces correct results, but the test is not
>>         happy. With
>>         pointers to the real check in the tests that is wrong. And an
>>         explanation why the result is wrong.
>>
>>         For (2)-(4) it would be a minimal standalone C/Fortran
>>         example including
>>         only the critical function/part of algorithm that is not
>>         correct/not
>>         portable/not compiled correctly, with results obtained with
>>         optimizations where it works and where it doesn't. Unless you
>>         find an
>>         obvious bug in R easy to explain (2), when the example would
>>         not have to
>>         be standalone. With such standalone C example, you could
>>         easily test the
>>         results with different optimizations and compilers, it is
>>         easier to
>>         analyze, and easier to produce a bug report for GCC. What
>>         would make it
>>         harder in this case is that it needs special hardware, but
>>         you could
>>         still try with the example, and worry about that later (one
>>         option is
>>         running in an emulator, and again a standalone example really
>>         helps
>>         here). In principle, as it needs special hardware, the
>>         chances someone
>>         else would do this work is smaller. Indeed, if it turns out
>>         to be (3),
>>         it is unlikely to get resolved, but at least would get
>>         isolated (you
>>         would know what not to run).
>>
>>         As a user, if you run into a problem like this and do not
>>         want to get it
>>         fixed, but just work it around somehow. First, it may be
>>         dangerous,
>>         possibly one would get incorrect results from computations,
>>         but say in
>>         applications where they are verified externally. You could
>>         try disabling
>>         individual specific optimization until the tests pass. You
>>         could try
>>         with later versions of gcc-11 (even unreleased) or gcc-12.
>>         Still, a lot
>>         of this is easier with a small example, too. You could ignore
>>         the
>>         failing test. And it may not be worth it - it may be that you
>>         could get
>>         your speedups in a different, but more reliable way.
>>
>>         Using wsl2 on its own should not necessarily be a problem and
>>         the way
>>         you built gcc from the description should be ok, but at some
>>         point it
>>         would be worth checking under Linux and running natively -
>>         because even
>>         if these are numerical differences, they could be in
>>         principle caused by
>>         running on Windows (or in wsl2), at least in the past such
>>         differences
>>         were seen (related to (2) above). I would recommend checking
>>         on Linux
>>         natively once you have at least a standalone R example.
>>
>>         Best
>>         Tomas
>>
>>
>>         >
>>         > best regards,
>>         > Kieran
>>         > ______________________________________________
>>         > R-devel using r-project.org mailing list
>>         > https://stat.ethz.ch/mailman/listinfo/r-devel
>>
	[[alternative HTML version deleted]]



More information about the R-devel mailing list