similar to: FLAME

Dear R-help, Here's a posting to the most recent NA-digest: From: Robert van de Geijn <rvdg at cs.utexas.edu> Date: Fri, 13 Dec 2002 11:15:23 -0600 Subject: Fast BLAS Libraries for Current Architectures Recent research by Kazushige Goto, Visiting Scientist at UT-Austin, has resulted in high-performance BLAS libraries for the Intel (R) Pentium (R) III and 4 processors, the

Sorry for not responding earlier. On 05/20/2016 03:05 PM, Roman Gareev wrote: > Thank you very much for the advices! I could probably try to avoid > using of nonhardware prefetching in the project, if Tobias doesn’t > disagree with it. My understanding is that prefetching isn’t used > explicitly in [1] and, according to [2], in some cases 90% of the > turbo boost peak of the

2016-05-19 21:45 GMT+05:00 4lbert C0hen <4lbert.h.c0hen at gmail.com>: > One short note. I would advise against spending time on prefetching for x86. > Recent hardware prefetchers are amazingly good at strided accesses in > single-threaded code. Caution: this is not based on objective/published > data, but on personal experience. > > There are open challenges in

Hi Tobias, according to [1], we can expect 90% of the turbo boost peak of the processor with a C version of Matrix-Matrix Multiplication that is similar to the one presented in [1]. In case of Intel Core i7-3820 SandyBridge, the theoretical maximal performance of the machine is 28.8 gflops and hence the expected number is 25,92 gflops. However, in case of, for example, n = m = 1056 and k = 1024

I don't think it is necessary to put binary versions of the Darwin/X11 version on CRAN any more. The current sources compile without problems on both Mac OS X 10.1 and 10.2, using either gcc/g77 2.95.2 or gcc/g77 3.1, or f2c. It seems to me that if you are using R from the command line, using X11, that you are sophisticated enough to install the tools below. And, if you want to install and

On 10/11/2017 02:44 PM, Lamar Owen wrote: > On 10/10/2017 11:22 AM, Mark Haney wrote: >> >> We have this discussion on every list I've ever been, or currently >> are on about every 6 months or so.? I do my best to contribute to the >> list as often as I can, but I can't help people when they are deadset >> on doing dangerous things.? Posts like his, and

On 10/10/2017 11:03 AM, Andrew Holway wrote: > Hiya everyone, > > Is there a way to disable a thread that has degenerated into flaming? The > recent "discussion" on /var/run descended into some quite nasty places and > perhaps a lid should have been put on it. This seems to happen every few > weeks and is somewhat embarrassing when I'm trying to persuade people of

homals-0.9.0 is on CRAN -- by Jan de Leeuw and Patrick Mair This package implements the methods discussed in Gifi, Nonlinear Multivariate Analysis, Wiley, 1990. In the Gifi terminology it covers homals, princals, canals, morals, criminals, and overals. The R implementation fills several gaps in Gifi, adding multiple ordinal, numerical, and polynomial data transformations. Differences with

homals-0.9.0 is on CRAN -- by Jan de Leeuw and Patrick Mair This package implements the methods discussed in Gifi, Nonlinear Multivariate Analysis, Wiley, 1990. In the Gifi terminology it covers homals, princals, canals, morals, criminals, and overals. The R implementation fills several gaps in Gifi, adding multiple ordinal, numerical, and polynomial data transformations. Differences with

This is probably only interesting to zero or one or two persons. gcc-3.3 from Apple's CVS does not handle the -bundle flag properly. I had to hack configure.ac, which now has shlib_ldflags="-nostdlib /usr/lib/bundle1.o -Wl,-bundle -Wl,-bundle_loader,/usr/local/lib/R/bin/R.bin -lgcc" ;; LIBR_LDFLAGS="-dynamiclib -install_name \$(Rexeclibdir)/libR.dylib -framework

Over the last couple of years I have written quite a few R programs for various "psychometric" techniques, and I am regularly updating and expanding what is there. I now have (wholly or partially), or have planned -- gifi package (update to homals on CRAN). Code for multiple correspondence analysis, nonlinear principal component analysis, nonlinear multiset canonical

Byron I meant there is no native (Aqua) version of GTK yet (there is wxWindows and Qt and Tk, so possibly GTK will follow). The X11 version of RGtk works for me too. I agree with your diagnosis of the GUI problem. Best -- Jan On Monday, August 19, 2002, at 12:08 AM, Byron Ellis wrote: > > On Monday, August 19, 2002, at 12:09 AM, Jan de Leeuw wrote: >> > > I accidentally

Pretty soon, hopefully, the RAqua version will make the Darwin/X11 version unnecessary. In the meantime, for convergence, I'll modify the Gifi version in various ways. -- It's 1.7.1 -- It no longer supports gnome -- It still uses Tcl/Tk for X11 (using 8.5 from CVS) !! It no longer uses anything from fink (readline and dlcompat as on Stefano's site, jpeg and png and teTeX from

Thank you for looking into this so quickly. As you correctly surmise, I was using the Carbon version of R-1.6.1 on Mac OS 10.2.2 (Jaguar) when I got the "wrong" answers. One other observation: The right censoring seems to work fine. Thanks again, Tim On Thursday, November 14, 2002, at 11:09 AM, Jan de Leeuw wrote: > I take that back. I now get the "correct" result

I have build R-1.6.0 on Jaguar (actually on 10.2.2), using gcc 3.1 from Apple, g77 3.1 from fink, ATLAS from fink. It passes all checks OK. I also re-compiled the 233 packages listed below with the same setup. I have NOT tested all these packages in any detail. A tar.gz file with the whole /usr/local/lib/R will be on ftp://gifi.stat.ucla.edu on Tuesday. It will NOT include anything from fink

I have made a regular R install, so that /Library/Frameworks/R.frameworks/Resources/bin/R.bin is now a Carbon APPL, using Rez in the way described before. We now have basically two R's -- one from Terminal.app and one from R.app. The first one works fine, except for the small glitch of not starting the default device properly. The second one does not work properly with Aqua Tcl/Tk,

Many new things at http://www.cuddyvalley.org/psychoR/ The scalassoc package, which fits exponential distance association models to indicator matrices, it now at version 1.0.0. It seems to be robust and can analyze large examples easily. It is a major improvement (in speed and robustness) over the distassoc package, which is at the same site. scalassoc does something neat (if you like

1. R-devel on OS X aborts when the code below is run after spitting out the message Loading required package: MASS Parameter 5 to routine DSTEIN was incorrect Mac OS BLAS parameter error in DSTEIN, parameter #0, (unavailable), is 0 2. I compiled R-patched using the included BLAS and Lapack, although Brian says this is impossible on OS X, because they are in FORTRAN (??). This version of R

I was obviously fascinated by Duncan's earlier message on RXLisp. I managed to build a working version of RXLisp for OS X, by some unabashed and rather uninformed hacking. I compiled XLISP-STAT using the gcc flags -fno-common and --no-cpp-precomp (not sure if they are necessary or not, they are part of an older hack). I "make libxlisp.so", using Duncan's replacement for the

On 10/10/2017 11:22 AM, Mark Haney wrote: > > We have this discussion on every list I've ever been, or currently are > on about every 6 months or so.? I do my best to contribute to the list > as often as I can, but I can't help people when they are deadset on > doing dangerous things.? Posts like his, and posts like yours make it > harder for me to bother trying to