X-pop3-spooler:	POP3MAIL 2.1.0 b 4 980420 -bs-
Date:	Wed, 8 Jul 1998 09:21:14 -0400 (EDT)
From:	Joshua Lamorie <jpl AT xiphos DOT ca>
X-Sender:	jpl AT ntsweb DOT com
To:	Marc Lehmann <pcg AT goof DOT com>
cc:	beastium <beastium-list AT Desk DOT nl>
Subject:	Re: please benchmark / MMX #2
In-Reply-To:	<19980708045815.20996@cerebro.laendle>
Message-ID:	<Pine.BSI.3.95.980708091020.26421A-100000@ntsweb.com>
MIME-Version:	1.0
Sender:	Marc Lehmann <pcg AT goof DOT com>
Status:	RO
X-Status:	A
Lines:	195

Gidday there,

This was my first ever benchmarking thingy.. so  it could all be
noise.. but these are my results. (with script, and actual results
appended)

Processor: P-II 233MHz (3.5 * 66MHz) 64Megs RAM - Xwindows (SuSE)

Test 1,
800k wav file

NOMMX
user: 2.236   (-0%)

MMX
user: 2.175   (-2.7%)

MMX-ONLY
user: 2.098   (-6.2%)

Test 2,
11000k wav file (11706412)

NOOMMX
user: 48.17   (-0%)

MMX
user: 47.74   (-0.9%)

MMX-ONLY
user: 46.41   (-3.6%)
                         
Percentage Calculation: 1-(MMX/NOMMX) i.e. 1 - (47.74/48.17)

Appendix A: Script

1. Command line:
../run_tests > results 2>&1

2. run_tests 
#!/bin/sh

SOURCE=$HOME/radio/test.wav
SOURCENAME=test.wav
OUTPUT=$HOME/pgcc/test/test.wav.bz2
BINDIR=$HOME/pgcc
TESTDIR=$HOME/pgcc/test

echo "No MMX"
for i in 1 2 3 4 5 6 7 8 9 10; do
        rm -rf $OUTPUT;
        cp $SOURCE $TESTDIR;
        time $BINDIR/bzip2-nommx $SOURCENAME;
done;

echo "MMX"
for i in 1 2 3 4 5 6 7 8 9 10; do
        rm -rf $OUTPUT;
        cp $SOURCE $TESTDIR;
        time $BINDIR/bzip2-mmx $SOURCENAME;
done;

echo "MMX-ONLY"
for i in 1 2 3 4 5 6 7 8 9 10; do
        rm -rf $OUTPUT;
        cp $SOURCE $TESTDIR;
        time $BINDIR/bzip2-mmxonly $SOURCENAME;
done;                                         

Appendix B:
Actual Output
No MMX

real    0m53.178s
user    0m47.630s
sys     0m0.450s

real    0m51.764s
user    0m46.810s
sys     0m0.560s

real    0m51.466s
user    0m47.400s
sys     0m0.520s

real    0m51.779s
user    0m48.540s
sys     0m0.510s

real    0m51.764s
user    0m48.740s
sys     0m0.690s

real    0m51.279s
user    0m48.930s
sys     0m0.620s

real    0m51.290s
user    0m48.750s
sys     0m0.480s

real    0m54.001s
user    0m47.180s
sys     0m0.540s

real    0m51.572s
user    0m49.000s
sys     0m0.580s

real    0m50.669s
user    0m48.760s
sys     0m0.530s
MMX

real    0m53.739s
user    0m48.970s
sys     0m0.630s

real    0m53.227s
user    0m46.800s
sys     0m0.550s

real    0m50.328s
user    0m47.520s
sys     0m0.510s

real    0m52.522s
user    0m49.830s
sys     0m0.650s

real    0m52.774s
user    0m46.660s
sys     0m0.470s

real    0m50.278s
user    0m46.900s
sys     0m0.650s

real    0m50.841s
user    0m47.270s
sys     0m0.560s

real    0m52.793s
user    0m49.030s
sys     0m0.590s

real    0m52.687s
user    0m46.770s
sys     0m0.550s

real    0m54.151s
user    0m47.650s
sys     0m0.660s
MMX-ONLY

real    0m51.707s
user    0m45.280s
sys     0m0.680s

real    0m53.375s
user    0m46.930s
sys     0m0.530s

real    0m51.358s
user    0m46.760s
sys     0m0.490s

real    0m50.136s
user    0m45.960s
sys     0m0.640s

real    0m52.324s
user    0m49.450s
sys     0m0.520s

real    0m51.960s
user    0m45.610s
sys     0m0.460s

real    0m48.882s
user    0m46.420s
sys     0m0.490s

real    0m47.972s
user    0m45.570s
sys     0m0.520s

real    0m48.200s
user    0m45.510s
sys     0m0.670s

real    0m48.470s
user    0m46.650s
sys     0m0.540s

- Raw text -