Date: Fri, 13 Oct 1995 12:26:05 +0100 (MET) From: Didier Derny <didier@omnix.fr.org> To: David Greenman <davidg@root.com> Cc: hardware@FreeBSD.org, dyson@root.com, batie@agora.rdrop.com Subject: Re: benchmark results Message-ID: <Pine.BSF.3.91.951013122444.3092D-100000@zapata.omnix.fr.org> In-Reply-To: <199510121256.FAA00163@corbin.Root.COM>
next in thread | previous in thread | raw e-mail | index | archive | help
is it possible to find the programs used to make these benchmarks. I've recently tested the AMD 486DX4-120 with dhrystone 1.1 and I get 141000 dhrystones / secondes. +---------------------+ | Didier Derny | | didier@omnix.fr.org | +---------------------+ On Thu, 12 Oct 1995, David Greenman wrote: > Attached are the results of some benchmark comparisons between similar > systems. All of the machines have Triton chipset based motherboards, so the > only significant difference is async/pipeline-burst cache and the CPU speed. > > -DG > > 90Mhz-async: > ASUS PCI/I-P54TP4, 32MB RAM, 256k async cache, Pentium 90Mhz > FreeBSD-2.1-stable (Oct. 9th) > > 90Mhz: > ASUS P/I-P55TP4XE, 64MB RAM, 256K pipeline-burst cache, Pentium 90Mhz > FreeBSD-current (Oct. 10th) > > 133Mhz: > ASUS P/I-P55TP4XE, 64MB RAM, 256K pipeline-burst cache, Pentium 133Mhz > FreeBSD-2.0.5-RELEASE > > Note that on the 90Mhz machine, the source being compiled is 2.1-stable, > which should be very similar (in time) to the source in 2.0.5-RELEASE. > -pipe was used in all cases. All tests done twice (and the second time > was used) to make sure that everything was cached. > > "CORBIN" kernel make depend: > 46.100u 18.674s 1:05.98 98.1% 105+657k 7+51io 0pf+0w (90Mhz-async) > 44.291u 11.232s 0:56.38 98.4% 101+535k 7+49io 0pf+0w (90Mhz) > 29.265u 9.772s 0:39.94 97.7% 104+659k 3+41io 0pf+0w (133Mhz) > > "CORBIN" kernel make: > 432.518u 53.907s 8:14.79 98.3% 1046+1271k 13+505io 0pf+0w (90Mhz-async) > 347.405u 28.275s 6:23.66 97.9% 1006+1293k 14+497io 0pf+0w (90Mhz) > 260.303u 27.605s 5:00.20 95.9% 1043+1264k 52+480io 0pf+0w (133Mhz) > > For the following, the same (-static) binary was used within each test set. > > dry: > (-O -finline-functions -fomit-frame-pointer) > Dhrystone(1.1) time for 500000 passes = 3 (90Mhz-async) > This machine benchmarks at 157248 dhrystones/second > 3.184u 0.015s 0:03.31 96.3% 48+213k 0+0io 0pf+0w > > Dhrystone(1.1) time for 500000 passes = 3 (90Mhz) > This machine benchmarks at 159203 dhrystones/second > 3.159u 0.007s 0:03.16 99.6% 46+206k 0+0io 0pf+0w > > Dhrystone(1.1) time for 500000 passes = 2 (133Mhz) > This machine benchmarks at 230215 dhrystones/second > 2.175u 0.007s 0:02.18 99.5% 47+207k 0+0io 0pf+0w > > ttcp: > [implode:davidg] ttcp -p9 -n16384 -t localhost (90Mhz-async) > ttcp-t: buflen=8192, nbuf=16384, align=16384/+0, port=9 tcp -> localhost > ttcp-t: socket > ttcp-t: connect > ttcp-t: 134217728 bytes in 16.76 real seconds = 7822.28 KB/sec +++ > ttcp-t: 16384 I/O calls, msec/call = 1.05, calls/sec = 977.79 > ttcp-t: 0.1user 7.0sys 0:16real 42% 40i+247d 68maxrss 0+2pf 7244+1594csw > 0.120u 7.045s 0:16.77 42.6% 40+247k 0+0io 0pf+0w > > [corbin:davidg] ttcp -p9 -n16384 -t localhost (90Mhz) > ttcp-t: buflen=8192, nbuf=16384, align=16384/+0, port=9 tcp -> localhost > ttcp-t: socket > ttcp-t: connect > ttcp-t: 134217728 bytes in 12.49 real seconds = 10495.37 KB/sec +++ > ttcp-t: 16384 I/O calls, msec/call = 0.78, calls/sec = 1311.92 > ttcp-t: 0.1user 5.6sys 0:12real 46% 42i+275d 98maxrss 0+2pf 5443+4471csw > 0.175u 5.693s 0:12.50 46.8% 42+275k 0+0io 0pf+0w > > [pyrex:davidg] ttcp -p9 -n16384 -t localhost (133Mhz) > ttcp-t: buflen=8192, nbuf=16384, align=16384/+0, port=9 tcp -> localhost > ttcp-t: socket > ttcp-t: connect > ttcp-t: 134217728 bytes in 9.31 real seconds = 14071.71 KB/sec +++ > ttcp-t: 16384 I/O calls, msec/call = 0.58, calls/sec = 1758.96 > ttcp-t: 0.0user 4.4sys 0:09real 49% 41i+267d 46maxrss 0+2pf 4445+6338csw > > iozone: > (We're only interested in the cache read performance) > MB reclen bytes/sec written bytes/sec read (90Mhz-async) > 4 512 2085319 12761286 > 4 8192 2127731 24353765 > > MB reclen bytes/sec written bytes/sec read (90Mhz) > 4 512 2235863 15093178 > 4 8192 2455878 31840138 > > MB reclen bytes/sec written bytes/sec read (133Mhz) > 4 512 1635305 19576771 > 4 8192 1656868 34723076 >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.951013122444.3092D-100000>