Last updates:
Fri Dec 27 09:18:32 2002
Fri Sep 24 08:28:02 2004
Thu Nov 4 06:19:14 2004
Index of files in /pub/benchmarks
-
00dir.cmd
-
FTP command file in alphabetical order
-
00dir.lst
-
verbose alphabetical directory listing
-
00tdir.cmd
-
FTP command file in reverse-time order
-
00tdir.lst
-
MD5SUM
MD5 128-bit checksums
-
MD5SUM.asc
digitally-signed MD5 128-bit checksums
-
SHA1SUM
SHA1 160-bit checksums
-
SHA1SUM.asc
digitally-signed SHA1 160-bit checksums
-
verbose reverse-time order directory listing
-
dec-alpha.ps.gz
-
Analysis of DEC Alpha 3000/500 results extracted from
matmlt report. See also pub/dec-alpha on
ftp.math.utah.edu for a large collection of DEC Alpha
performance reports and a pointer to where to get more.
-
laserdoctor.ps
-
PostScript file to display PostScript device resources,
and perform a simple timing benchmark
-
laserdoctor.timing
-
Benchmark results for several systems for
laserdoctor.ps.
-
local
-
Large collection of benchmark results from local
research programs in the University of Utah Mathematics
Department.
-
matmlt.ps.001-050.gz
-
Gzipped PostScript files containing
-
matmlt.ps.051-100.gz
-
typeset matrix multiply benchmark
-
matmlt.ps.101-150.gz
-
report. Each file should be uncompressed
-
matmlt.ps.151-200.gz
-
and printed separately. The complete
-
matmlt.ps.201-250.gz
-
indexed report is 626 pages long.
-
matmlt.ps.251-300.gz
-
Revision date: [30-Nov-1992].
-
matmlt.ps.301-350.gz
-
"" "" ""
-
matmlt.ps.351-400.gz
-
"" "" ""
-
matmlt.ps.401-450.gz
-
"" "" ""
-
matmlt.ps.451-500.gz
-
"" "" ""
-
matmlt.ps.501-550.gz
-
"" "" ""
-
matmlt.ps.551-600.gz
-
"" "" ""
-
matmlt.ps.601-650.gz
-
"" "" ""
-
matmlt.ps.651-700.gz
-
"" "" ""
-
matrix-multiply.report.910208.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
08-Feb-1991
-
matrix-multiply.report.911127.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
27-Nov-1991
-
matrix-multiply.report.920208.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
08-Feb-1992
-
matrix-multiply.report.920406.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
06-Apr-1992
-
matrix-multiply.report.920523.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
23-May-1992
-
matrix-multiply.report.920709.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
09-Jul-1992
-
matrix-multiply.report.920718.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
18-Jul-1992
-
matrix-multiply.report.921130.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
30-Nov-1992
-
matrix-multiply.report.930305.gz
-
Matrix multiply benchmark report in ASCII form (gzipped)
05-Mar-1993
-
matrix-multiply.dvi.930305.gz
-
gzipped TeX DVI file of matrix multiply benchmarks (this
can be processed by any TeX DVI driver) 05-Mar-1993
-
perf-upd
-
LINPACK update summary
-
performance.ps.gz
performance.pdf
-
LINPACK summary (PostScript) (gzipped)
-
random-io.report.901130.gz
-
Random I/O benchmark report in ASCII form (gzipped)
30-Nov-1990
-
random-io.report.911127.gz
-
Random I/O benchmark report in ASCII form (gzipped)
27-Nov-1992
-
random-io.report.920403.gz
-
Random I/O benchmark report in ASCII form (gzipped)
03-Apr-1992
-
random-io.report.930113.gz
-
Random I/O benchmark report in ASCII form (gzipped)
13-Jan-1993
-
texbook.timing
-
Benchmark of typesetting the TeXbook
-
slalom
-
Subdirectory with NASA Ames Slalom benchmark results
-
usirep.pdf
usirep.ps.gz
-
Utah Supercomputing Institute Report #3 on
High-Performance Matrix Multiplication
-
x11perf
-
Subdirectory with X Window System benchmarks
Other machines known to provide anonymous ftp access to benchmark
data include:
ftp://ftp.uu.net/published/byte/benchmarking/dos
ftp://ftp.uu.net/published/byte/benchmarking/macintosh
ftp://ftp.uu.net/published/byte/benchmarking/unix
ftp://ftp.uu.net/published/byte/benchmarking/windows
http://bitmover.com/lmbench/whatis_lmbench.html
The lmbench collection contains results for several different UNIX
systems, and covers things like process spawing, interprocess
communication, system latency, and memory latency; these are rarely
covered by benchmark collections that concentrate on
number-crunching performance.