IUBio

GCG Benchmarks wanted

Chad Price cprice at inetnebr.com
Fri Apr 9 14:43:02 EST 1999


I'm in process of purchasing a new system to run GCG and  am looking at
all 3 possible vendors: SGI, Sun, and an Alpha system.

Systems under consideration are:
    SGI Origin 2000, 4 250Mhz processors, 1GB RAM, and 18GB drives
    E3500, 4 400Mhz processors, 1GB RAM, and 18GB drives
    Alpha 4100, 2 600Mhz processors, 1GB RAM, 18GB drives.

No RAID is under consideration.  I want speed and RAM rather than the
RAID uptime.

It seems that the most commonly used really long-running job here is
fasta.  I've tested it on a
couple of systems to which I have access.  Running
    Hypersparc, 2 90Mhz processors and 160MB RAM and GCG
10.0: cpu=1:18:54
    Alpha4100, 2 466Mhz processors, .5GB RAM and GCG 9.1: cpu=0:27:35

If anyone can provide some further tests of fasta, I'd appreciate it.
I've included the sequence used below. It was run agains the most
recently released GenEmble (the 99-1 database release from GCG) with all
default
options except MAXPROCESSORS, which was set to 2 rather than the default
64.

I'd particularly like to see results of tests on the systems I'm
considering (see above). CC's of any results directly to me as well as
the newsgroup would be appreciated as I have to use a pretty convoluted
method to get to the newsgroups.  My University no longer carries them
and my ISP has out-sourced news and I cannot connect directly from
work..


!!NA_SEQUENCE 1.0

   core3  Length: 526  April 8, 1999 19:39  Type: N  Check: 8545  ..

       1  TAgCTGGGCC CAGGCCTTGT ACGGTGCAGT CAGTACTGTC ACTATCACAC

      51  TTGCCACAGT GACATTCAGT GGCTACTGGA TATGTATACA GGGAATCTGC

     101  CTGGTGAGCA CAGCCGGGCA CTTTCACTGT CTCGTACACC AGCTCTTTGA

     151  AGGTACATGT TTTCTGGTTG TTGGGCCTGG CTGGGTCCTT GTACACTAGA

     201  TCCTGAGGAG TTTAGAAAGG AGAGGATTAG GTTAAAATAT GCTGGAAGTT

     251  TCTATGTAGC ATAGTGAAAA TGAGATGGAA ACTTAATATT GCTTGTTCTA

     301  ACATTTACCC ATTAAAGCTC ACTTAAGGAA TTCCTAACCA AGGCAAATAT

     351  GATCTTATAA ATATCCTATT TCCTGGTTCT GGAAATCATA TAgAAAAGAT

     401  AAGTCATTCC TTTTCCTACT CCTTATGTCA ATATCAAATC TTAAATATTT

     451  CTCAAACTGT AGAgAAATTT TAAATTTCnA TATAGAAAAA GATgAAAAGT

     501  GAAATATAAC AAATTTTAAG TTGATT








More information about the Info-gcg mailing list

Send comments to us at biosci-help [At] net.bio.net