GIV matrix - additional information

From: vivi noviati <vivi_pmt34_at_YAHOO.COM>
Date: Wed, 6 Jun 2007 22:56:27 -0700

Dear All,
  I sent you the subset data (data file, as file, giv file, and asr file) so you can take a look and please tell me if you find something wrong. Because it is a small job and 2GB memory should be enough to run it, so I suspect I did something wrong.
  For the larger data set, the data (BLUEs) has been centered (substract the trial mean), so there is no variability in trials (ie. occ).
  Thank you.

vivi noviati <> wrote:
  Date: Wed, 6 Jun 2007 21:42:48 -0700 (PDT)
From: vivi noviati <>
Subject: GIV matrix

  Dear All,
  I need some assitance for running ASREML job. I tried to fit GIV matrix into ASREML. However, it always said that it failed to form G structure and I should increase memory. Currently, my computer only able to allocate 2048MB. The GIV matrix is a lower triangular matrix from 686 x 686 entry. It has three column: row, colomn, and value.
  I also try to run the job under unix, it can allocate 4096MB (s9), but when it allocate that memory, it failed to read my data (it able to read my data when I used less memory).
  This is the sample of my as file:
  ESWYT yield combine data
 tid 1 !A
 yea 25 !I
 occ 208 !I
 loc 386 !A
 ent 50 !I
 gen 686 !I
C:\DATA\YLD\stage_2\GIV_browse.giv !skip 1
C:\DATA\YLD\stage_2\yld_cent.txt !skip 1 !AISING
  BLUE !WT wt ~ mu !r gen yea.occ.gen
  1 1 1
0 0 0 !s2== 0.443036672875
  gen 1
gen 0 GIV
  I also try to subset the data and only run with 30 genotype and 17 location. But it still failed to form G scores (memory allocated 2048MB).
  However, if I use giv(gen) instead of fitted G structure model, it will converge. Does giv() give the same result as fitted G structure?
  Thank you.
  Vivi Arief
  University of Queensland
  Get the free Yahoo! toolbar and rest assured with the added security of spyware protection.

Park yourself in front of a world of choices in alternative vehicles.
Visit the Yahoo! Auto Green Center.

Received on Sun Jun 06 2007 - 22:56:27 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)