Date: Fri, 2 Feb 1996 04:13:26 -0600 (CST) From: mailing list account <lists@argus.flash.net> To: freebsd-hackers@freebsd.org Subject: pitest results on ALR P6 at 150MHz Message-ID: <199602021013.EAA01704@argus.flash.net>
next in thread | raw e-mail | index | archive | help
I thought someone out there would like to see this pitest run...
The essentials of the system:
ALR P6 150MHz, 64M,
AHA-2940W, Connor CFP4207W (4.28G),
and the standard 2.1-RELEASE kernel, as installed.
f77 options were for all tests:
-O2 -m486 -fomit-frame-pointer -pipe
All runs use Borwein's quartically convergent algorithm, and run on an unloaded
system.
I find the results to be impressive. The final test utilized a large amount
of swap space, and thus shows both the stability and efficiency of the vm
subsystem. I feel that the run time of the last test could have been improved
through the use of a higher performance drive [avg sequential ~= 3.5M/S], it
is a shame to see such a potentially classy machine being sold with a
substandard hard drive at the price they sold it for. Yes, the drive has a
fast/wide interface, although it is anything but. Let's say that for the same
price ALR charged, I could have gotten a Barracuda. The Connor drive also
failed to completely verify after a format, verify, verify, format, verify,
verify cycle, although they did offer to cover 2-day shipping for a free
exchange part. The format always did complete though, and it was possible to
run these benchmarks reliably, if there is any change in the drive's transfer
rates when we get the replacement, I'll post an update.
Overall, with the above noted exceptions, this ALR box is hot.
More benchamarks tomorrow [Less time/resource consuming ones].
-------------------------------------------------
The first run was completely in core, and is the only one to compare to
the documentation for pitest, it is for 24,570 places.
-------------------------------------------------
PI4 COMPUTATION TEST -- DP COMPLEX FFT MP VERSION
MW = 14 NW = 4096 ND = 24570
ITERATION 1 7
ITERATION 2 7
ITERATION 3 7
ITERATION 4 7
ITERATION 5 7
ITERATION 6 7
ITERATION 7 7
CPU TIME = .0000 SECONDS.
62.87 real 62.18 user 0.05 sys
1676 maximum resident set size
41 average shared memory size
1337 average unshared data size
164 average unshared stack size
338 page reclaims
0 page faults
0 swaps
0 block input operations
0 block output operations
0 messages sent
0 messages received
0 signals received
0 voluntary context switches
665 involuntary context switches
-------------------------------------------------
The next run was completely in core, and for 786,426 places of pi.
-------------------------------------------------
PI4 COMPUTATION TEST -- DP COMPLEX FFT MP VERSION
MW = 19 NW = 131072 ND = 786426
ITERATION 1 10
ITERATION 2 10
ITERATION 3 10
ITERATION 4 10
ITERATION 5 10
ITERATION 6 10
ITERATION 7 10
ITERATION 8 10
ITERATION 9 10
ITERATION 10 10
CPU TIME = .0000 SECONDS.
5022.98 real 720.11 user 0.21 sys
33200 maximum resident set size
285 average shared memory size
-7975 average unshared data size
1141 average unshared stack size
8242 page reclaims
0 page faults
0 swaps
1 block input operations
13 block output operations
0 messages sent
0 messages received
0 signals received
2 voluntary context switches
55949 involuntary context switches
-------------------------------------------------
The last run was to compute pi to 1,572,858 places, swapinfo shows that this
run used approximately 49,572k of swap space while running.
-------------------------------------------------
PI4 COMPUTATION TEST -- DP COMPLEX FFT MP VERSION
MW = 20 NW = 262144 ND = 1572858
ITERATION 1 10
ITERATION 2 10
ITERATION 3 10
ITERATION 4 10
ITERATION 5 10
ITERATION 6 10
ITERATION 7 10
ITERATION 8 10
ITERATION 9 10
ITERATION 10 10
CPU TIME = .0000 SECONDS.
11871.80 real 2976.27 user 1.72 sys
55080 maximum resident set size
238 average shared memory size
6434 average unshared data size
637 average unshared stack size
112812 page reclaims
22004 page faults
0 swaps
14 block input operations
24 block output operations
0 messages sent
0 messages received
0 signals received
22117 voluntary context switches
145809 involuntary context switches
-------------------------------------------------
You may obtain the pitest benchmark from
ftp://wcarchive.cdrom.com/pub/FreeBSD/incoming/pitest.tar.gz
Jim
--
All opinions expressed are mine, if you | "I will not be pushed, stamped,
think otherwise, then go jump into turbid | briefed, debriefed, indexed, or
radioactive waters and yell WAHOO !!! | numbered!" - #1, "The Prisoner"
jbryant@argus.flash.net - FlashNet Communications - Ft. Worth, Texas
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199602021013.EAA01704>