GCG on DEC AXP info

Maggie smith at GCG.COM
Tue May 17 10:47:47 EST 1994

In response to David Mathog's request:

<Will someone at GCG please explain exactly what the benchmarks were?

<>Some recent Alpha AXP performance benchmarks that may be of interest are:

<>FASTA Benchmark (run with cooperation of MIT/Whitehead)
<>SPARC10/40			1.20 sec
<>AlphaAXP OSF/1 3000-600	 .37 sec
<>AlphaAXP OSF/1 3000-800	 .31 sec

<>BLAST (run with cooperation of NIH/NCBI)
<>SPARCstation/2 		88 sec
<>SPARC 10/41		  	43 sec
<>SGI R3000		  	65 sec
<>SGI Challenge/XLR4400	  	23 sec
<>AlphaAXP OSF/1 3000-600	23 sec
<>AlphaAXP OSF/1 3000-800	19 sec

The benchmarks were run using native FastA and BlastP, as opposed to FastA 
as it is found in the Wisconsin Package.  Here is the information that was 
supplied to me by Digital this morning when I asked for their help in 
answering this question:

The FASTA command was "fasta -Q -n -b 100 -d 100 alu1 test.seq", where alu1
is a 1002-residue sequence called HUMALURP7, and test.seq is a library
containing 22572 residues in 11 sequences. The options mean that HUMALURP7
will be treated as a DNA sequence, there will be no prompting for any other
input, and all scores and alignments will be output. 

The BLAST command was "blastp swissprot dxch.aa", where swissprot is the
SWISS-PROT database (Release 26.0, August 1993, 31,808 sequences; 10,875,091
total letters) and dxch.aa is "Gene X protein - Chicken (fragment)" (232

Maggie Smith
Genetics Computer Group
smith at gcg.com

More information about the Info-gcg mailing list

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