search for: flexibla

I agree that the "R Installation and Administration" guide statement about FlexiBLAS needs an update. Standard R in CentOS Stream 9, which is now downstream from Fedora, gives the following result for Matrix products in sessionInfo(): > sessionInfo() R version 4.3.1 (2023-06-16) Platform: x86_64-redhat-linux-gnu (64-bit) Running under: CentOS Stream 9 Matrix products: default...

On 10/11/23 23:12, George Ostrouchov wrote: > I agree that the "R Installation and Administration" guide statement about FlexiBLAS needs an update. Standard R in CentOS Stream 9, which is now downstream from Fedora, gives the following result for Matrix products in sessionInfo(): > >> sessionInfo() > R version 4.3.1 (2023-06-16) > Platform: x86_64-redhat-linux-gnu (64-bit) > Running under: CentOS Stream 9 &g...

Hi, R has been built against FlexiBLAS in rawhide [1, 2]. These are the R packages that need to be rebuilt: R-ape-0:5.4-2.fc33.x86_64 R-expm-0:0.999.4-7.fc33.x86_64 R-gee-0:4.13.20-4.fc33.x86_64 R-gss-0:2.2.2-3.fc33.x86_64 R-igraph-0:1.2.5-3.fc33.x86_64 R-msm-0:1.6.8-5.fc33.x86_64 R-preprocessCore-0:1.50.0-4.fc33.x86_64 R-qtl-0:1.46.2...

...s://www.tcl-lang.com/software/tcltk/9.0.html. > Yes, you just need to install tcl8-devel and tk8-devel (and tell dnf to remove version 9 if it's already there) and compile as usual. > Building against external BLAS is also not self-evident, so may need > guidance. Building against FlexiBLAS is recommended. So you need to install flexiblas-devel, then tell R via --with-blas=flexiblas --with-lapack. See the configure part of our spec for a complete reference of what we do: https://src.fedoraproject.org/rpms/R/blob/rawhide/f/R.spec#_326 I?aki > > Any suggestions welcome! >...

...h-arch_64=x86-64-v2 --with-arch_32=x86-64 --build=x86_64-redhat-linux --with-build-config=bootstrap-lto --enable-link-serialization=1 Thread model: posix Supported LTO compression algorithms: zlib zstd gcc version 11.4.1 20230605 (Red Hat 11.4.1-2) (GCC) I have been building R 4.4.0 and 4.4.1 with Flexiblas and with the built in R BLAS/LAPACK. R BLAS: ./configure --prefix=/tools/R/$RVER --enable-R-shlib --enable-memory-profiling --with-pcre2=/tools/pcre2/10.42 Flexiblas: PKG_CONFIG_PATH=/tools/flexiblas/3.4.2/lib64/pkgconfig ./configure --prefix=/tools/R/flexiblas/4.4.1 --enable-R-shlib --enabl...

...R-devel that allows R to be linked to Tcl/Tk 9. I'll give it a few days to bed in and then port it over to R-4-5-branch so that it can go in R 4.5.1. Martyn > > > Building against external BLAS is also not self-evident, so may need > > guidance. > > > Building against FlexiBLAS is recommended. So you need to install > flexiblas-devel, then tell R via --with-blas=flexiblas --with-lapack. > > See the configure part of our spec for a complete reference of what we do: > > https://src.fedoraproject.org/rpms/R/blob/rawhide/f/R.spec#_326 > > I?aki > >...

Dear Miguel Esteva, I couldn't get a Red Hat "ubi9" container to install enough dependencies to build R. Is there a way to reproduce your setup on a virtual machine somewhere? On Fri, 3 May 2024 00:42:43 +0000 Miguel Esteva via R-help <r-help at r-project.org> wrote: > *** caught segfault *** > > address 0x1801fa8f70, cause 'memory not mapped' > >

I've not worked changing underlying computational infrastructure, but developers who do might want to use ideas from FlexiBLAS. Apologies in advance if this is well-known. Best JN > From: Martin Koehler koehlerm at mpi-magdeburg.mpg.de > Date: January 07, 2016 > Subject: FlexiBLAS Version 1.3.0 Release > > It is our pleasure to announce the new release of FlexiBLAS. We do not > provide yet another BLAS...

...ADS=1 is a solution. There isn't much R can do about the behaviour of the BLAS, because there is no standard interface to set the number of threads. Some BLASes (like ATLAS) don't even offer it as a tunable number at all [*]. A system administrator could link the installation of R against FlexiBLAS [**], provide safe defaults in the environment variables and educate the users about its tunables [***], but that's a choice just like it had been a choice to link R against a parallel variant of OpenBLAS on a shared computer. This is described in R Installation and Administration, section A.3...

...this in 83986 [1] needs a fix in Lapack.c: if (dladdr((void *) F77_NAME(ilaver), &dl_info)) { char buf[PATH_MAX+1]; char *res = realpath(dl_info.dli_fname, buf); if (res) { SEXP nfo = R_NilValue; if (strstr(res, "flexiblas")) nfo = R_flexiblas_info(); if (isNull(nfo)) nfo = mkChar(""); ans = ScalarString(nfo); break; } } It will always return an empty string except for flexiblas. I assume this...

Hello, Would a less resource-intensive value, such as 1, be a safer default CPU value for confint? I noticed excessive CPU usage on a I.T. administrator-managed server which was being three-quarters used by another staff member when the confidence interval calculation in an R Markdown document suddenly changed from two seconds to ninety seconds because of competition for CPUs between users. Also,

Building R 4.5.0 from source on Fedora 42 hits problems because 42 ships with Tcl/Tk 9. tcl8-devel and tk8-devel are available and can be used, but I haven't yet found a viable incantation yet. Neither R released nor devel support Tcl/Tk 9 https://www.tcl-lang.com/software/tcltk/9.0.html. Building against external BLAS is also not self-evident, so may need guidance. Any suggestions welcome!

...mes undesired (e.g. your case). A single setting > is probably far from enough. > >> Does R even know whether a linked library is doing multi-processing? > > Unfortunately, no, there is no standard interface for that. Best I can > recommend is to link your R installation with FlexiBLAS and then use the > 'flexiblas' CRAN package to talk to it. [[alternative HTML version deleted]]

...tic.c(77): error: pointer to incomplete class type is not allowed exc->retval = 0.0; icc -std=c99 -std=gnu11 -I../../src/extra -I../../src/extra/xdr -I. -I../../src/include -I../../src/include -I/usr/local/include -I../../src/nmath -DHAVE_CONFIG_H -fopenmp -fpic -g -O3 -wd188 -ip -mp -c flexiblas.c -o flexiblas.o icc: command line remark #10148: option '-mp' not supported compilation aborted for arithmetic.c (code 2) make[3]: *** [../../Makeconf:129: arithmetic.o] Error 2 make[3]: *** Waiting for unfinished jobs.... icc: command line remark #10148: option '-mp' not supporte...

From the R Installation and Admin manual: "There is a tendency for re-distributors of R to use ?enhanced? linear algebra libraries without explaining their downsides.? There?s a downside not mentioned in the manual that caught and baffled me for a while. I was using all 64 cores of an AWS instance via parallel::mclapply() and doing matrix multiplications in the parallelized function. If

...re isn't much R can do about the behaviour of the BLAS, because > there is no standard interface to set the number of threads. Some BLASes > (like ATLAS) don't even offer it as a tunable number at all [*]. > > A system administrator could link the installation of R against > FlexiBLAS [**], provide safe defaults in the environment variables and > educate the users about its tunables [***], but that's a choice just > like it had been a choice to link R against a parallel variant of > OpenBLAS on a shared computer. This is described in R Installation and > Adminis...

...lowed >> ?????? exc->retval = 0.0; >> >> icc -std=c99 -std=gnu11 -I../../src/extra -I../../src/extra/xdr -I. >> -I../../src/include -I../../src/include? -I/usr/local/include >> -I../../src/nmath -DHAVE_CONFIG_H?? -fopenmp -fpic? -g -O3 -wd188 -ip >> -mp? -c flexiblas.c -o flexiblas.o >> icc: command line remark #10148: option '-mp' not supported >> compilation aborted for arithmetic.c (code 2) >> make[3]: *** [../../Makeconf:129: arithmetic.o] Error 2 >> make[3]: *** Waiting for unfinished jobs.... >> icc: command line rem...

...essing the same cache line. (descriptors already grew to 16 bytes, do we need more?) - was playing around with the idea of different ring types for different devices e.g. scsi, net. starting with generic information then comes protocol specific data. Ilya agrees. length of descriptor would be flexibla by adding a descriptor length field. How to continue / TODOs: - do benchmarking with bigger frame sizes on fast enough NICs - turn prototype code into a RFC series (work in progress) - more people interested to join monthly meetings Open questions: - Do we need an (optional) completion rin...

...essing the same cache line. (descriptors already grew to 16 bytes, do we need more?) - was playing around with the idea of different ring types for different devices e.g. scsi, net. starting with generic information then comes protocol specific data. Ilya agrees. length of descriptor would be flexibla by adding a descriptor length field. How to continue / TODOs: - do benchmarking with bigger frame sizes on fast enough NICs - turn prototype code into a RFC series (work in progress) - more people interested to join monthly meetings Open questions: - Do we need an (optional) completion rin...

...vector for argument what. A new generic, nameOfClass(), is called to resolve the class name from what. This supports experimenting with alternative object systems. Based on contributions by Tomasz Kalinowski in PR#18485. * Detection of BLAS/LAPACK in use (sessionInfo()) with FlexiBLAS now reports the current backend. * The "data.frame" method for subset() now warns about extraneous arguments, typically catching the use of = instead of == in the subset expression. * Calling a:b when numeric a or b is longer than one may now be made int...