similar to: default min-v/nsize parameters

On Thu, Jan 15, 2015 at 3:55 PM, Michael Lawrence <lawrence.michael at gene.com> wrote: > Just wanted to start a discussion on whether R could ship with more > appropriate GC parameters. I've been doing a number of similar measurements, and have come to the same conclusion. R is currently very conservative about memory usage, and this leads to unnecessarily poor performance on

>>>>> Peter Haverty <haverty.peter at gene.com> >>>>> on Mon, 19 Jan 2015 08:50:08 -0800 writes: > Hi All, This is a very important issue. It would be very > sad to leave most users unaware of a free speedup of this > size. These options don't appear in the R --help > output. They really should be added there. Indeed,

Hi All, This is a very important issue. It would be very sad to leave most users unaware of a free speedup of this size. These options don't appear in the R --help output. They really should be added there. Additionally, if the garbage collector is working very hard, might it emit a note about better setting for these variables? It's not really my place to comment on design philosophy,

Martin M has suggested I widen this discussion to R-devel, and > I agree that we should increase them, > but I'm not sure at all about the amount. > > The default could even depend on the architecture (via "./configure").. Views, please. ------------- Begin Forwarded Message ------------- Is is not time we increased the defaults a bit? As the base gets bigger I hit

Hi all, I'm having a very minor problem. Most every time I use the pager (I think) I get the following error about character 32. It works fine, but the error is a bit annoying. > ?version sh: invalid character 32 in exportstr for export R_NSIZE sh: invalid character 32 in exportstr for export R_VSIZE R.Version package:base R Documentation (and then the

When I use the -v I can modify the size of the heap, as assessed by gc(), but the -n key seems to be without effect ? On a machine with 48 mB ram I can load the libraries without problem, but on my own 36 mB ram machine, I get dynload problems with the larger, eg survival4. Any suggestions ? Troels -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list

When I use the -v I can modify the size of the heap, as assessed by gc(), but the -n key seems to be without effect ? On a machine with 48 mB ram I can load the libraries without problem, but on my own 36 mB ram machine, I get dynload problems with the larger, eg survival4. Any suggestions ? Troels -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list

This example from the INSTALL help seems to be broken in R 0.90.1 (on Solaris): gilp/dse : R CMD check --help Usage: R CMD check [options] [-l lib] pkg_1 ... pkg_n I'm trying to figure out how to request more nsize and vsize when using R CMD check. Paul Gilbert -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read

[RFC = Request for Comments] { Yes, the documentation for .Renviron is really not there (but the FAQ...); the rest is in ?Startup } In R's Startup (on Unix only??) {unless --no-environ is specified} ~/.Renviron (if there) is read as an 'sh' script before R is called, then R looks ((for the site-wide Rprofile and then)) for .Rprofile in the current directory and then for

For some of my packages I am getting a segmentation fault in Linux when I use R CMD check. (Using R 1.0.1) The segmentation fault does not happen in Solaris and in some cases it does not happen in Linux if I set R_NSIZE and R_VSIZE much higher than I need in Solaris. Should I expect a segmentation fault if there is not enough memory for R CMD check, or are these unrelated? Paul Gilbert

Dear R users we're having a problem reading a largish data file using read.table(). The file consists of 175000 lines of 4 floating pt numbers. Here's what happens: > dat_read.table('sst.dat') Error: memory exhausted (This is line 358 of src/main/memory.c). Cutting down the file to around 15000 lines allows read.table() to work OK. I edited the memory limits in Platform.h

Full_Name: Laurent Gautier Version: 0.65.0 OS: mips SGI-Irix 6.5 Submission from: (NULL) (195.110.4.98) Using R through emacs with ess5.1.8, I cannot set R workspace (--vsize and --nsize). So far I was using R0.64.2 without such a problem. I am aware my bug report is a bit light, but just let me know if anything I could do with my R and emacs would be of any help for\ specifying better what is

Hi, The Gnome version now compiles, and it should also be working (at least as much as it ever has). I've changed Makefile.in to the new system, which is very cool. What I want to work on now is: - Graphics. I want to move to the Gnome canvas for this, which should be reasonably easy. This will give us rotated text (which I never got going properly before) and the option for

Full_Name: J Utans Version: 1.1.1 OS: NT4 (SP6) Submission from: (NULL) (155.140.123.250) After grep is called many times (> 250k), with constant strings as patterns, it complains with "invalid regular expression" on calls that worked before (with same pattern and x). At the same time copying to the clipboard does no longer work with "out of memory" error (i.e. when trying

Dear R-ers, While browsing the R sources, I found the following piece of code in src\main\memory.c: static void reset_pp_stack(void *data) { R_size_t *poldpps = data; R_PPStackSize = *poldpps; } To me, it looks like the poldpps pointer is a nuissance; can't you just cast the data pointer and derefer it at once? Say, static void reset_pp_stack(void *data) { R_PPStackSize = *

Hi Folks, I loaded a couple of quite large data sets into an R session and then quit (after saving the image). Now I get: Error: a read error occured Fatal error: unable to restore saved data (remove .RData or increase memory) after trying to start my R session using something like: R --vsize XXX --nsize 1000000 For any value of XXX (I went up to 300 or 400, which is as high as I could go.

Dear R people, I wonder if some kind person can tell me the correct syntax to set the R_VSIZE environmental variable. I tried R_VSIZE = 10M in ~/.Renviron and also export R_VSIZE = 10M. These don't seem to work. I scrounged around looking for details about this, couldn't find any, got fed up. Faheem.

Dear All, I'm linking R from another application and embedding it as described in the R-exts manual, i.e. with initialization done via Rf_initEmbeddedR. While everything works the same as in standalone R for Linux, under Windows I found a difference in the default memory.limit, which is fixed to 2GB (both win32 and win64) - compared to a limit in standalone R of 3.5GB for win32 and 16GB on

Dear R developers, The slow performance of serializing to a raw vector on Windows is an issue that has appeared in this list before. It appears to be due to the frequent use of realloc from the resize_buffer method in serialize.c. I suggest a more granular, but still incremental, re-allocation of memory. For example change near the top of resize_buffer to: R_size_t newsize = needed + 65536 -

When doing repeated regressions on large data sets, I'm finding that the time spent on garbage collection often exceeds the time spent on the regression itself. Consider this test program which I'm running on an Intel Haswell i7-4470 processor under Linux 3.13 using R 3.1.2 compiled with ICPC 14.1: nate at haswell:~$ cat > gc.R library(speedglm) createData <- function(n) {