.\" $Header$
.\" groff -t -ms -X benchmark.tr
.\" groff -t -ms benchmark.tr | print-postscript
.\"
.\"  sort '-t       ' +7nr -8
.de BR
\fB\\$1\fR\\$2
..
.TL
The BRL-CAD Benchmark Suite
.AU
Michael John Muuss
.AU
Charles M. Kennedy
.AU
Lee A. Butler
.AI
The U.S. Army Research Laboratory
.PP
This document describes the BRL-CAD Benchmark Suite Methodology, and
the Ray-Tracing Figure of Merit (RTFM) used to judge the computational
performance of different computer systems.
.PP
The core of the test is the \fBrt\fR ray-tracing program.  The results
for each processor are determined by running a benchmark test of the
RT ray-tracing program on each of the sample databases provided.  The
\fBrt\fR program requires a binary database for input ("db/xxx.g"),
and produces a binary image as output ("bench/xxx.pix"), along with a
log file ("bench/xxx.log").  Each run will produce a 512x512x24 bit
color shaded image in a "bench/xxx.pix" file, and a run log in
"bench/xxx.log".  Reference copies of the images are provided as
"pix/xxx.pix" and "pix/xxx.log".  The sample databases provided are
processed in order of increasing difficulty, and each produces a
colorful image.  In each picture there are two light sources; the
primary source is located near the center of the model (the white
ball), and the secondary light source is located at the "eye"
position.
.NH 1
The Benchmark Models
.LP
moss.g -- This is a simple model, containing a yellow torus, a green
ellipsoid, a bluish cube, a pink truncated (distorted) cone, all
sitting on a blue plate.  Note the shadowing, and "specular splash".
.sp
world.g -- The same model as moss.g, but with various rendering
features enabled.  The model is enclosed in a hollow cloud sphere.
The plate is a mirror, the egg is crystal, the box has a debugging
texture map applied.
.sp
star.g -- A familiar sight to Star Trek fans; a low-detail exterior of
the ship "Enterprise".  The hull is white, with some other structures
in other colors.  The small red object in the foreground is the
shuttlecraft Galileo, to scale.
.sp
bldg391.g -- An exploded view of a two-story imaginary computer site,
including various walls, red stairways, green disk drives, etc.  The
lower floor is a mirror, the upper floor is glass.
.sp
m35.g -- A truck.
.sp
sphflake.g -- A sphereflake.
.NH 1
MACHINE INDEPENDENCE
.PP
To make the benchmark distribution machine independent, the databases
are provided in an ASCII format in the "db" directory, and
ascii-to-binary converters are provided in the "conv" directory.  The
setup command "\fBcake benchmark\fR" (see the document \fIInstalling
the BRL-CAD Package\fR) will automatically compile and run the
converters.
.PP
If you are running this benchmark on a non-UNIX machine that has
difficulty writing binary streams, change the "-o file.pix" flag to
"-O file.a_pix" and RT will write an ASCII form directly.  For more
discussion, consult the doc/pix.5 and conv/asc2pix.1 manual pages.
.PP
The reference images are distributed in binary BRL-CAD
.BR pix (5)
format.  Earlier releases distributed the reference images in the
portable ASCII form, and then converted them to the binary form.  This
consumed a significant amount of extra storage on UNIX machines, so
this practice was discontinued.  It is assumed that anyone attempting
to benchmark a system that can not read pure binary files will at
least have access to a UNIX system.  That UNIX system can be used to
convert the reference images to the portable ASCII form with the
.BR pix2asc (1)
tool, and then those images may be sent to the machine under test.
.PP
Information about modifying the benchmark to run on a particular
system may be found in the document \fIInstalling the BRL-CAD
Package\fR.  For a detailed explanation of various hardware specific
modifications necessary to take advantage of parallel processing, see
the document \fIWorkstations, Networking, Distributed Graphics, and
Parallel Processing\fR.  Implementations for several different types
of hardware in use at ARL are included.
.PP
If you encounter difficulty with running the benchmark, you might wish
to build the full BRL-CAD package, to take advantage of the image
tools in determining the nature of the problem.  For example, one
system suffered from a compiler bug that made TGC solids vanish, while
another system had difficulties with the light visibility
calculations.
.NH 1
OBTAINING CORRECT RESULTS
.PP
The benchmark timings are not considered valid unless the correct
results are given.  Make sure that the answers match the reference
files to within plus-or-minus one in each color (see pixdiff(1)).
Slight variations in the calculated pixels are generally attributable
to variations in floating point precision.  You may wish to compare
the execution times and log file remarks from your tests (in the
"bench/xxx.log" files) with the VAX-11/780 (with hardware FPA) times
which are given in the "pix/xxx.log" files.  The log file for sphflake
has yet to be run on the same system, causing weighted results.  It's
still useful and informative, however, for deriving performance
statistics.  The reference images are also located in the "pix/"
directory.
.PP
For example, if you are running the benchmark on a Cray XMP, the
world.pix and bldg391.pix images will have a single byte in the blue
channel off by 1, out in the cloud background.  This is not to be
considered an error.
.NH 1
BENCHMARK IMAGE DISPLAY
.PP
If you have a suitable framebuffer, you may wish to display the images
generated by the benchmark, and compare them to the reference images.
This section assumes that you have successfully compiled the full set
of BRL-CAD software, not just the benchmark portion.  Programs for
dealing with images are in the "util" directory.  To display a
\fBpix\fR(5) file on a framebuffer, set environment variable FB_FILE
(see \fBbrlcad\fR(1) for details), and run \fBpix-fb\fR.  Note that by
leaving FB_FILE unset, your default display will be used.
.sp
.nf
.ft B
	cd bench
	pix-fb moss.pix
.ft R
.fi
.sp
If the images computed on your machine do not match the reference
images, the program "util/pixdiff" will compute a pix file that will
highlight the differences for you, and report a summary of bytes
equal, off-by-1, and off-by-many, e.g.,
.sp
.ft B
	pixdiff moss.pix ../pix/moss.pix \(or pix-fb
.ft R
.fi
.sp
For a display of the relative magnitude of the differences at each
pixel, use "util/bwdiff" instead, e.g.,
.sp
.nf
.ft B
	bwdiff moss.pix ../pix/moss.pix \(or pix-fb
.ft R
or
.ft B
	bwdiff moss.pix ../pix/moss.pix \(or \\
	bwmod -s 128 -m 4 -a 128 \(or pix-fb
.ft R
.fi
.sp
and for statistics on the differences, use
.sp
.ft B
	bwdiff moss.pix ../pix/moss.pix \(or pixstat
.ft R
.ne 8i
.NH 1
THE COMPETITION
.PP
The full benchmark takes about 9 CPU hours on a VAX-11/780, so you
should scale your expectations accordingly.
.PP
Here is a sampling of some systems that have been tested, with as much
information about the configuration as could be obtained.  Some
systems have a great many options for small things that can affect
performance, such as memory interleaving.
.LP
.TS
center box;
l l.
System Name	Configuration
_	_
axposf	DEC Alpha 4000/710, 256M, OSF/1 V3.0 \(dg
ex-art	Pentium-P90, 16MB, 256K cache, BSD/OS 2.0,
hawks	486DX/33, 24MB, 256K cache, BSD/OS 2.0
orac	486/100, 20MB, No cache, BSD/OS 2.0 (laptop)
pk	Cray C90, C916/161024, UNICOS 8.0.2.4
vapor	SGI, 18* 75 MHz R8000 IP21, 1024M RAM 8-way, Irix 6.0
vblt	66 MHz 486DX2/66, 16M RAM, Linux 1.1.73
vgr	DEC VAX-11/780 w/FPA, 64M RAM, 4.3 BSD
vision	Sun-3/280 w/68881 FPA, 16M RAM, SunOS 4.1.1
voyage	SGI, 4* 25 MHz R3000 IP7, 56M RAM, Irix 4.0.1
whiz	SGI Ingido**2, 100 MHz IP22, 96M RAM, Irix 5.2
waffle	SGI, 4* 25 MHz R3000 IP7, 64M RAM, Irix 4.0.5 (heavy network use)
warp	Sun SPARCstation 5, 32M RAM, SunOS 4.1.4
wax	SGI, 24* 150 MHz R4400 IP19, 1024M RAM 4-way, Irix 5.2
wilson	SGI, 8* 100 MHz R4400 IP19, 1024M RAM 8-way, Irix 5.2
wilted	Sun-4/25 ELC, 16M RAM, SunOS 4.1.1
wimp	33 MHz 386DX/33 w/387, 256k cache, 16M RAM, BSDI 2.0beta2
wimpy	Sun SS10BSX-GX, 2* 50 MHz SuperSPARC, 64M RAM, SunOS 5.4
wizard	SGI Ingido**2, 150 MHz IP22, 96M RAM, Irix 5.3
.TE
\(dg The Alpha was heavily loaded while the benchmark was being run.
DEC requires us to show the load average at the time:
.br
03:28  up 18 days, 10:59,  7 users,  load average: 2.00, 2.00, 2.08
.NH 1
BENCHMARK RESULTS
.PP
In the tradition of Dongarra, the Rays/sec figure is considered the
"RT Figure of Merit" (RTFM).  Note that the RTFM can only be compared
between different runs on the \fIsame\fR database; it is not
meaningful to compare RTFM numbers between different databases. The
Rays/sec number for a multi-processor machine is for "aggregate CPU
cluster seconds", rather than rays/total CPU seconds (which remains
fairly constant on good parallel machines).
.PP
The statistics are recorded in an easily understood format, with all
results for a particular configuration listed on a single line.  The
numbers reported are the rays/sec (RTFM) figure.  Entries denoted with
a dagger ("\(dg") produced incorrect results.  Entries denoted with a
dash ("-") were not available and should be considered differently
than those entries with a dagger.  Generally the results are not
available because time constraints prevented running the complete set
of databases.
.fi
.br
.ne 8i
.NH 1
RELEASE 4.4 STATISTICS (4-Jan-1995)
.PP
There is very little difference in RTFM numbers between Release 4.0
and Release 4.4, so they can be usefully compared.  In some cases the
4.4 performance is slower due to the use of shared libraries.  For
example "vision" dropped from 1.33 to 1.22 due to the shared
libraries.  For most users reducing the disk consumption from 300
MBytes to 30 MBytes makes this a reasonable tradeoff.  However, for
the best performance, use non-shared libraries.
.LP
.TS H
expand box;
l | n n n n n | n | l.
H/W	Moss	World	Star	Bldg	M35	Mean	S/W Notes
_
.TH
vapor	810.91	1091.39	1012.63	1027.25	1092.85	1007	-P18 static
vapor	773.8	1038.7	979.49	970.91	1036.62	959.9	-P17
vapor	732.85	1021.69	968.93	922.62	1024.15	934.04	-P18 (SW lock)
wax	769.38	940.83	953.36	924.7	1050.14	927.68	-P24
vapor	730.17	994.97	934.43	917.1	987.83	912.9	-P16
wax	725.26	957.68	927.61	932.67	992.75	907.19	-P24
wax	756.95	915.96	940.43	903.45	966.09	896.57	-P23
vapor	686.8	939.63	897.72	884.39	927.16	867.14	-P15
wax	691.18	853.75	887.17	859.32	933.41	844.96	-P21
wax	724.73	683.08	890.87	900.82	939.31	827.76	-P22
vapor	645.32	881.18	843.23	801.41	859.26	806.08	-P14
wax	637.28	757.74	820.21	871.37	877.6	792.84	-P20
vapor	608.12	835.84	795.36	759.95	806.32	761.11	-P13
_
wax	606.98	784.3	754.46	799.95	794.52	748.04	-P19
wax	590.16	758.23	794.06	794.52	757.79	738.95	-P18
wax	579.29	722.35	787.55	720.26	754.65	712.82	-P17
vapor	571.14	783.72	740.47	697.24	746.72	707.85	-P12
wax	553.72	697.1	744.21	663.61	723.03	676.33	-P16
vapor	534.85	716.58	685.87	656.25	685.77	655.86	-P11
wax	519.01	655.57	690.95	651.9	663.92	636.27	-P15
wax	475.88	637.33	665.38	608.31	653.99	608.17	-P14
vapor	499.46	663.26	633.48	614.58	624.34	607.02	-P10
wax	449.16	591.54	616.21	590.59	614.25	572.35	-P13
vapor	463.3	598.34	575.07	555.62	560.89	550.64	-P9
wax	424.71	556.8	595.02	521.63	538.78	527.38	-P12
vapor	424.89	534.5	509.85	497.63	503.99	494.17	-P8
wax	391.66	490.63	545.28	488.57	508.96	485.02	-P11
wax	315.45	472.03	525.47	442.37	461.44	443.35	-P10
vapor	381.1	467.67	440.13	425.57	442.04	431.3	-P7
wax	309.35	426.37	386.11	417.85	434.96	394.92	-P9
vapor	336.55	392.2	390.72	373.92	380.03	374.68	-P6
_
wax	307.6	361.32	396.07	380.17	380.74	365.18	-P8
wax	245.87	332.33	362.5	338.62	331.39	322.14	-P7
vapor	288.45	327.78	328.76	314.54	320.16	315.93	-P5
wax	254.82	255.29	299.93	302.08	256.54	273.73	-P6
wilson	223.01	265.17	287.8	262.19	269.62	261.55	-P8
vapor	240.6	269.93	266.46	253.99	259.05	258	-P4
wax	215.96	231.65	221.78	248.63	221.45	227.89	-P5
wax	177.94	185.27	213.74	201.64	197.27	195.17	-P4
vapor	187.58	198.84	202.28	191.94	194.74	195.07	-P3
wax	120.4	138.01	162.12	159.36	141.33	144.24	-P3
vapor	130.61	140.04	134.88	129.38	129.78	132.93	-P2
wax	100.83	102.09	111.12	87.26	98.62	99.98	-P2
wimpy	99.10	95.78	92.15	97.51	96.65	96.23	-P2
wimpy	82.14	93.60	92.17	103.59	82.22	90.74	-P2
_
axposf	81.05	82.27	86.72	84.36	79.28	82.73	-P1
vapor	79.68	77.7	73.89	71.38	70.06	74.54	-P1 static
vapor	69.34	70.14	67.52	65.55	66.69	67.84	-P1 shared
wizard	63.8	55.86	68.97	73.21	61.56	64.68
wax	62.28	58.58	65.24	67.39	59.31	62.56	-P1
wax	59.08	56.06	61.45	64.4	56.11	59.42	-P1
voyage	52.78	55.44	58.19	58.03	52.41	55.37	-P4
pk	64.89	55.71	48.01	49.50	53.43	54.30	-P1
waffle	38.56	47.63	62.36	59.62	51.20	51.87	-P4
wimpy	50.17	48.82	50.46	55.12	49.27	50.76	-P1
wimpy	47.80	43.23	51.37	54.10	46.23	48.54	-P1
weber	46.53	43.45	50.88	50.33	46.03	47.44	-P1
wilson	41.16	39.54	45.98	42.93	34.82	40.88	-P1
whiz	36.25	31.08	37.21	39.62	36.98	36.22
warp	19.88	21.33	23.71	24.19	22.60	22.34
ex-art	16.89	18.54	22.98	26.97	25.57	22.19
orac	7.57	8.23	10.72	12.43	11.30	10.05
wilted	9.58	9.53	10.66	10.53	9.87	10.03
hawks	3.64	4.17	5.71	6.90	6.23	5.33
vblt	4.85	4.89	5.29	5.24	4.90	5.03
wimp	1.16	1.34	1.82	2.25	2.06	1.72
vision	1.07	1.23	1.23	1.31	1.26	1.22	-P1
vgr	1.0	1.0	1.0	1.0	1.0	1.0
.TE
.br
.ne 8i
.NH 1
THE OLD COMPETITION
.PP
These are the definitions for systems measured in previous tests.
.LP
.TS
center box;
l l.
System Name	Configuration
_	_
alliant	Alliant FX/8 (8 CEs, 9 IPs, 64 Mbytes), Concentrix 3.0
amber	HP 9000/720, 32 MB memory
amsaa-seer	Gould PN 9080, w/MACCs, UTX 2.0, 4x4Mb mem boards
ardec-3	Pyramid 90Mx, Dual-CPU, OSx 2.5, 16 Mbytes
bob	Cray-2, SN 2009, 4.3ns clock
cor3	Pyramid MIS-12T/3, OSx 192
patton	Cray X-MP/48, SN213, UNICOS 5.0, 8.5ns clock
uy1	Cray Y-MP8/2128, UNICOS 5.1.11
shpcrc2	IBM RS/6000, AIX 3.1
slc1	Macintosh II, 68020, 68881, AUX 1.1, GCC 1.39
spark	Gould PN 9050, no MACC, UTX 2.0, 2x4Mb mem boards
sws2	Convex C120, 16 Mbytes, Convex Unix 6.2.32.2
vector	Alliant FX/80 (8 ACEs, 6 IPs, 32 Mbytes), Concentrix
venom	Alliant FX/8 (8 CEs, 6 IPs, 32 Mbytes), Concentrix
vgr	VAX 780, FPA, 4.3 BSD
virus	Sun-3/50, 15 Mhz clock, 12 Mhz 68881, Sun UNIX 3.2
vista	SGI 3030, w/FP chip, UNIX release 3.5
vmb	Gould PN 9080, no MACCs, UTX 2.0, 12x1Mb mem boards
vhs	Silicon Graphics 4D/60T, 12.5 Mhz clock 16 MB memory
ovoyage	Silicon Graphics 4D/120, 16.7 Mhz clock 16 MB memory
taylor	Silicon Graphics 4D/220, 25 Mhz clock 16 MB memory
crim	Silicon Graphics 4DCRIMS 50Mhz clock 128 MB memory
wiltse	Silicon Graphics 4D R4000 50Mhz clock
c1east	T{
Convex C1 XP (2 IOPs, 4 Multibus, 64 Mbytes),
S/W 6.0.1.12
T}
elxsi-gnu	Elxsi 6420, BSD 4.2 16 MB
elxsi-gnuy	Elxsi 6420, Sys Vr2, 16MB
elxsi-m1	Elxsi 6410, BSD 4.2 16 MB
hep	Denelcor HEP, 4 PEMs
indigo	SGI Iris 3030, FPA, GL2-W3.5
mseries	T{
MIPS Mseries, Release 0.6, UMIPS 2.1, 16 MB of R2350 RAM.
R2000 CPU rev 5.0, R2010 FP rev 2.0
T}
multiflow	Multiflow Trace 7/200 (PRELIMINARY)
nova	T{
Sun SPARCserver 490, 32 Mb, 33 MHz clock, 33 MHz TI 8847 FPU, SunOS 4.1.1B
T}
rh2	Tektronix, Motorola 88000
snm2	Cray 1-M, SN2, UNICOS 2.0
sws1	Convex C1, ConvexOS 9.1 48MB memory
tek4132	Tektronix 4132, 32082 fpp, UTek 2.3
utah-gr	VAX 785, FPA, 4.3 BSD
utah-cs	VAX 8600, FPA, 4.3 BSD
veto	Sun 4/260 32MB memory Sun UNIX 4.0.3
violet	VaxStationII GPX, Ultrix 1.2
vision	Sun-3/280, 16 MB memory
vortac	Sun-3/160, 16.67 Mhz clock, 12 Mhz 68881, Sun 3.2
walrus	Silicon Graphics 4D/280 25 Mhz clock 64 MB memory Irix 3.3.1
whiz	Silicon Graphics 4D/240 25 Mhz clock 56 MB memory Irix 3.3.1
whisper	Sun SPARCStation 1+
wilted	Sun SPARCstation ELC , Diskless, 12MB memory Sun UNIX 4.1.1
worm	Silicon Graphics 4D/280 25 Mhz clock 64 MB memory Irix 3.3.1
.TE
.br
.ne 8i
.NH 1
OLD BENCHMARK RESULTS
.PP
These are the results of previous benchmarks.
.NH 1
RELEASE 1.20 STATISTICS (12-Feb-1987)
.LP
.TS H
expand box;
l l | n n n n | n | l.
A/*	H/W	Moss	World	Star	Bldg	Mean	S/W Notes
_
.TH
Rays	tek4132	80.5	44.0	33.9	31.4	?	UTek 2.3
*vgr	tek4132	0.72	0.69	0.62	0.62	?
_
Rays	violet	107.7	57.3	44.7	44.1	?	Ultrix 1.2
*vgr	violet	0.96	0.90	0.81	0.87	?
_
Rays	violet	119.0	63.0	50.0	46.3	?	Ultrix 1.2,
*vgr	violet	1.06	0.99	0.91	0.91	?	w/fast_sqrt
_
Rays	vgr	112.1	63.6	55.0	50.7	?	BSD 4.3
*vgr	vgr	1.0	1.0	1.0	1.0	1.0
_
Rays	indigo	115.9	85.3	\(dg	101.8	?	GL2-W3.5
*vgr	indigo	1.03	1.34	\(dg	2.01	?
_
Rays	vista	116.4	86.0	\(dg	102.3	?	SGI UNIX 3.5
*vgr	vista	1.04	1.35	\(dg	2.02	?
_
Rays	virus	127.3	69.9	57.0	52.4	?	Sun UNIX 3.2
*vgr	virus	1.14	1.10	1.04	1.03	?
_
Rays	vortac	148.2	81.0	65.8	60.7	?	Sun UNIX 3.2
*vgr	vortac	1.32	1.27	1.20	1.20	?
_
Rays	utah-gr	191.8	105.8	89.9	86.5	?	4.3 BSD
*vgr	utah-gr	1.71	1.66	1.64	1.71	?
_
Rays	ardec-3	-	68.4	57.2	-	?	OSx 2.5
*vgr	ardec-3	-	1.08	1.04	-	?
_
Rays	elxsim1	380.5	232.2	204.8	189.3	?	BSD 4.2
*vgr	elxsim1	3.39	3.65	3.72	3.73	?
_
Rays	spark	413.4	232.0	213.8	211.9	?	UTX 2.0
*vgr	spark	3.69	3.65	3.89	4.18	?
_
Rays	vmb	413.9	233.4	210.6	212.6	?	UTX 2.0
*vgr	vmb	3.69	3.67	3.83	4.19	?
_
Rays	c1east	454.8	252.4	205.6	192.3	?	Convex UNIX 6.0.1.12,
*vgr	c1east	4.06	3.97	3.74	3.80	?	vanilla cc
_
Rays	seer	460.7	263.2	246.8	241.0	?	UTX 2.0
*vgr	seer	4.11	4.14	4.49	4.75	?
_
Rays	venom	492.9	228.0	180.0	157.1	?	Concentrix 2.0
*vgr	venom	4.40	3.58	3.27	3.10	?
_
Rays	elxgnu	520.6	315.4	264.4	242.0	?	BSD 4.2
*vgr	elxgnu	4.64	4.96	4.81	4.77	?
_
Rays	utah-cs	521.1	292.1	237.4	216.2	?	BSD 4.3
*vgr	utah-cs	4.64	4.59	4.32	4.27	?
_
Rays	c1east	521.6	285.3	230.5	215.1	?	Convex UNIX 6.0.1.12,
*vgr	c1east	4.66	4.49	4.20	4.24	?	vc -O1 (scalar)
_
Rays	c1east	527.7	287.4	228.7	210.0	?	Convex UNIX 6.0.1.12,
*vgr	c1east	4.71	4.52	4.16	4.14	?	vc -O2 (vector)
_
Rays	elxgnuy	644.7	349.3	264.6	242.1	?	System Vr2
*vgr	elxgnuy	5.75	5.49	4.81	4.77	?
_
Rays	mflow	845.2	439.5	313.1	338.9	?	(preliminary)
*vgr	mflow	7.54	6.91	5.69	6.68	?
_
Rays	venom	904.9	424.6	349.3	312.6	?	Concentrix 2.0,
*vgr	venom	8.07	6.68	6.35	6.17	?	2 CEs, no vectors
_
Rays	venom	1375.5	650.0	523.4	459.9	?	Concentrix 2.0,
*vgr	venom	12.27	10.22	9.52	9.07	?	3 CEs, no vectors
_
Rays	venom	1813.5	845.5	686.0	600.7	?	Concentrix 2.0,
*vgr	venom	16.17	13.29	12.47	11.85	?	4 CEs, no vectors
_
Rays	venom	2364.1	1104.4	870.2	775.9	?	Concentrix 2.0,
*vgr	venom	21.08	17.36	15.82	15.3	?	5 CEs, no vectors
_
Rays	snm2	2492.9	-	-	-	?	Unicos 2.0,
*vgr	snm2	22.24	-	-	-	?	no vectors, no optim
_
Rays	hep	2502.0	-	-	-	?	1 PEM, npsw=10
*vgr	hep	22.32	-	-	-	?
_
Rays	venom	2811.4	1319.0	1051.6	918.0	?	Concentrix 2.0,
*vgr	venom	25.08	20.74	19.12	18.1	?	6 CEs, no vectors
_
Rays	venom	3248.8	1533.0	1232.3	1065.9	?	Concentrix 2.0,
*vgr	venom	28.98	24.10	22.41	21.02	?	7 CEs, no vectors
_
Rays	patton	3453.3	1514.3	1271.2	1027.3	?	COS V115BF2
*vgr	patton	30.81	23.81	23.11	20.26	?	1 CPU, no vectors
_
Rays	venom	3677.3	1724.2	1375.4	1208.4	?	Concentrix 2.0,
*vgr	venom	32.80	27.11	25.01	23.83	?	8 CEs, no vectors
_
Rays	alliant	3972.9	1840.0	1457.2	1266.4	?	Concentrix 3.0,
*vgr	alliant	35.44	28.93	26.49	24.98	?	8 CEs, no vectors
_
Rays	hep	4055.9	-	-	-	?	1 PEM, npsw=40
*vgr	hep	36.18	-	-	-	?
_
Rays	patton	6856.6	-	2522.2	-	?	COS V115BF2
*vgr	patton	61.16	-	45.86	-	?	2 CPUs, no vectors
_
Rays	patton	10205.1	-	3749.9	-	?	COS V115BF2
*vgr	patton	91.04	-	68.18	-	?	3 CPUs, no vectors
_
Rays	patton	13320.2	-	4955.6	-	?	COS V115BF2
*vgr	patton	118.82	-	90.10	-	?	4 CPUs, no vectors
.TE
.NH 1
RELEASE 2.3 STATISTICS (2-Nov-1987)
.LP
.TS H
expand box;
l l | n n n n | n | l.
A/*	H/W	Moss	World	Star	Bldg	Mean	S/W Notes
_
.TH
Rays	vgr	118.8	64.9	36.5	34.2	?	BSD 4.3
*vgr	vgr	1.0	1.0	1.0	1.0	1.0
_
Rays	vmb	375.4	218.6	186.4	196.2	?	UTX 2.0
*vgr	vmb	3.16	3.37	5.11	5.74	?
_
Rays	venom	468.2	221.7	166.1	144.9	?	Concentrix 3.0,
*vgr	venom	3.94	3.42	4.55	4.24	?	1 CE, no vectors
_
Rays	sws2	471.9	258.4	202.5	186.4	?	Convex Unix 6.2.32.2
*vgr	sws2	3.97	3.98	5.55	5.45	?	vanilla cc -O
_
Rays	vector	656.6	301.6	221.1	195.0	?	Concentrix 3.0,
*vgr	vector	5.53	4.65	6.06	5.70	?	1 ACE, no vectors
_
Rays	venom	827.3	424.0	327.1	287.9	?	Concentrix 3.0,
*vgr	venom	6.96	6.53	8.96	8.42	?	2 CEs, no vectors
_
Rays	vhs	958.3	545.6	447.0	414.1	?	-O2 optimization
*vgr	vhs	8.07	9.41	12.25	12.11	?
_
Rays	mseries	988.3	603.5	515.6	491.6	?
*vgr	mseries	8.32	9.29	14.13	14.37	?
_
Rays	vector	1280.9	596.8	436.7	389.2	?	Concentrix 3.0,
*vgr	vector	10.78	9.19	11.96	11.38	?	2 ACEs, no vectors
_
Rays	venom	1288.7	630.6	486.4	428.0	?	Concentrix 3.0,
*vgr	venom	10.85	9.72	13.33	12.51	?	3 CEs, no vectors
_
Rays	venom	1732.9	853.2	645.4	560.1	?	Concentrix 3.0,
*vgr	venom	14.59	13.15	17.68	16.38	?	4 CEs, no vectors
_
Rays	bob	1856.7	801.9	574.1	500.3	?	1 CPU, UNICOS 3.0,
*vgr	bob	15.63	12.36	15.73	15.62	?	CC -O
_
Rays	vector	1952.1	881.7	647.6	578.8	?	Concentrix 3.0,
*vgr	vector	16.43	13.58	17.74	16.92	?	3 ACEs, no vectors
_
Rays	venom	2158.4	1031.3	797.3	703.1	?	Concentrix 3.0,
*vgr	venom	18.17	15.89	21.84	20.56	?	5 CEs, no vectors
_
Rays	bob	2346.1	1127.4	751.0	677.2	?	1 CPU UNICOS 4.0
*vgr	bob	19.75	17.37	20.57	19.80	?	CC -O
_
Rays	bob	2423.5	1045.7	719.3	609.4	?	1 CPU, UNICOS 3.0,
*vgr	bob	20.40	16.11	19.71	17.82	?	VCC -O
_
Rays	vector	2546.2	1170.0	867.2	758.7	?	Concentrix 3.0,
*vgr	vector	21.43	18.03	23.76	22.18	?	4 ACEs, no vectors
_
Rays	venom	2633.7	1232.7	943.4	839.5	?	Concentrix 3.0,
*vgr	venom	22.17	18.99	25.85	24.55	?	6 CEs, no vectors
_
Rays	venom	2769.4	1393.3	1098.1	971.6	?	Concentrix 3.0,
*vgr	venom	23.31	21.47	30.08	28.41	?	7 CEs, no vectors
_
Rays	vector	3131.2	1453.8	1065.8	945.4	?	Concentrix 3.0,
*vgr	vector	26.36	22.40	29.20	27.64	?	5 ACEs, no vectors
_
Rays	venom	3368.7	1605.3	1231.3	1095.11	?	Concentrix 3.0,
*vgr	venom	28.36	24.73	33.73	32.02	?	8 CEs, no vectors
_
Rays	patton	3456.5	1691.8	1169.0	1043.6	?	1 CPU, UNICOS 2.1
*vgr	patton	29.10	26.07	32.03	30.51	?	no vectors
_
Rays	vector	3686.4	1688.6	1259.1	1129.0	?	Concentrix 3.0,
*vgr	vector	31.03	26.02	34.49	33.01	?	6 ACEs, no vectors
_
Rays	vector	4140.3	1953.0	1463.3	1299.7	?	Concentrix 3.0,
*vgr	vector	34.85	30.09	40.09	38.00	?	7 ACEs, no vectors
_
Rays	bob	4641.7	2243.4	1498.2	1358.1	?	2 CPUs, UNICOS 4.0
*vgr	bob	39.07	34.57	41.05	39.71	?	CC -O
_
Rays	vector	4712.2	2226.3	1649.0	1469.7	?	Concentrix 3.0,
*vgr	vector	39.66	34.30	45.18	42.97	?	8 ACEs, no vectors
_
Rays	bob	6884.9	3422.4	2267.7	2034.3	?	3 CPUs, UNICOS 4.0
*vgr	bob	57.95	52.73	62.13	59.48	?	CC -O
_
Rays	bob	9477.1	4484.7	2947.0	2832.9	?	4 CPUs, UNICOS 4.0
*vgr	bob	79.77	69.10	80.76	82.83	?	CC -O
.TE
.NH 1
RELEASE 3.0 STATISTICS (10-Oct-1988)
.LP
.TS H
expand box;
l l | n n n n | n | l.
A/*	H/W	Moss	World	Star	Bldg	Mean	S/W Notes
_
.TH
Abs	vgr	138.85	67.15	54.48	49.11	77.39	BSD 4.3
*vgr	vgr	1.00	1.00	1.00	1.00	1.00	w/FPA
_
Abs	sdm	163.56	76.98	59.79	52.51	88.21	SunOS 3.4
*vgr	sdm	1.17	1.14	1.09	1.06	1.11	w/68881
_
Abs	ardec-3	417.95	212.43	177.01	160.73	242.03	9820, OSx 4.1
*vgr	ardec-3	3.01	3.16	3.24	3.27	3.17	-P 1
_
Abs	spark	443.84	216.92	190.42	192.82	261.00	UTX 2.0
*vgr	spark	3.19	3.23	3.49	3.92	3.37	no MACC
_
Abs	vmb	482.28	226.33	191.52	193.04	273.29	UTX 2.0
*vgr	vmb	3.47	3.37	3.51	3.93	3.57	no MACC
_
Abs	sws2	546.23	255.45	183.10	163.85	287.15	Convex 6.2.32.2
*vgr	sws2	3.93	3.80	3.36	3.33	3.60	/bin/cc
_
Abs	sun4	767.7	373.7	315.5	284.5	435.6	Unix 3.5
*vgr	sun4	5.52	5.56	5.79	5.79	5.62	-O2 optim
_
Abs	video	940.33	469.30	399.21	367.64	544.12	UNIX 2.0
*vgr	video	6.77	6.98	7.32	7.48	7.13	-O2
_
Abs	vhs	954.81	471.90	409.78	392.55	557.26	IRix 3.1
*vgr	vhs	6.87	7.02	7.52	7.99	7.45	Parallel, -P1
_
Abs	bob	2602.68	1078.25	712.39	642.42	1258.93	UNICOS 4.0, /bin/cc
*vgr	bob	21.22	18.17	15.39	13.98	17.19	-P 1
_
Abs	vector	3811.68	1469.80	1060.50	921.21	1815.79	Alliant 3.0
*vgr	vector	27.45	21.88	19.46	18.75	21.88	-P 5
_
Abs	vector	4515.09	1751.94	1266.49	1090.85	2156.09	Alliant 3.0
*vgr	vector	32.51	26.08	23.24	22.21	26.01	-P 6
_
Abs	patton	4550.67	1920.15	1298.61	1153.02	2230.61	UNICOS 3.0, /bin/cc
*vgr	patton	37.10	32.36	28.05	25.09	30.65	-P 1
_
Abs	vector	5130.74	2004.49	1438.97	1261.36	2458.89	Alliant 3.0
*vgr	vector	36.95	29.85	26.41	25.68	29.72	-P 7
_
Abs	bob	5186.24	2132.17	1409.10	1273.72	2500.30	UNICOS 4.0, /bin/cc
*vgr	bob	42.28	35.93	30.44	27.71	34.09	-P 2
_
Abs	vector	5813.76	2260.61	1638.31	1422.82	2783.87	Alliant 3.0
*vgr	vector	41.87	33.66	30.07	28.97	33.64	-P 8
_
Abs	amber	5901.29	2702.65	2096.35	1893.41	3148.42
*vgr	amber	42.50	40.24	38.47	38.55	39.94	-P 1
_
Abs	bob	7934.12	3392.98	2119.11	1929.74	3843.98	UNICOS 4.0, /bin/cc
*vgr	bob	64.69	57.18	45.78	41.99	52.41	-P 3
_
Abs	patton	8898.79	3782.55	2559.48	2292.13	4383.23	UNICOS 3.0, /bin/cc
*vgr	patton	72.56	63.75	55.30	49.88	60.37	-P 2
_
Abs	bob	10734.52	4436.61	2883.53	2617.47	5168.03	UNICOS 4.0, /bin/cc
*vgr	bob	87.52	74.77	62.30	56.96	70.38	-P 4
_
Abs	patton	13078.79	5631.95	3832.88	3416.90	6490.13	UNICOS 3.0, /bin/cc
*vgr	patton	106.64	94.92	82.81	74.36	89.68	-P 3
_
Abs	patton	17157.78	7437.55	5073.58	4531.25	8550.04	UNICOS 3.0, /bin/cc
*vgr	patton	139.90	125.35	109.62	98.61	118.37	-P 4
.TE
.bp
.NH 1
RELEASE 3.7 STATISTICS (19-June-1989)
.LP
.TS H
expand box;
l l | n n n n n | n | l.
A/*	H/W	Moss	World	Star	Bldg	M35	Mean	CPU
_
.TH
Abs	vgr	128.64	54.47	45.17	40.35	49.02	63.53	1
*vgr	vgr	1.00	1.00	1.00	1.00	1.00	1.00
_
Abs	rh2	1704.56	791.31	599.81	549.74	589.01	846.88	1
*vgr	rh2	13.25	14.52	13.27	13.62	12.01	13.33
_
Abs	spark	435.69	208.59	176.76	177.69	174.67	234.68	1
*vgr	spark	3.53	3.61	4.17	4.29	3.40	3.80
_
Abs	vhs	973.11	475.00	384.62	351.12	343.44	505.45	1
*vgr	vhs	7.56	8.72	8.51	8.70	7.00	8.09
_
Abs	whisper	1118.35	561.81	473.52	439.76	509.26	620.54	1
*vgr	whisper	8.69	10.31	10.48	10.89	10.38	10.15
_
Abs	nova	2024.80	976.01	790.17	732.35	?	1130.83	1
*vgr	nova	15.74	17.92	17.49	18.15	?	17.32
_
Abs	ovoyage	1969.57	1001.21	862.77	809.61	786.11	1085.85	2
*vgr	ovoyage	15.31	18.38	19.10	20.06	16.03	17.77
_
Abs	taylor	2446.98	1076.24	912.03	837.56	799.58	1214.47	1
*vgr	taylor	19.02	19.75	20.19	20.75	16.31	19.20
_
Abs	worm	2671.46	1218.49	966.76	859.61	932.18	1329.70	1
*vgr	worm	20.76	22.36	21.40	21.30	19.01	20.96
_
Abs	viper	3255.73	1368.96	1102.71	1039.37	1008.15	1554.98	8
*vgr	viper	25.30	25.13	24.41	25.75	20.56	24.23
_
Abs	vector	3956.86	1625.23	1234.03	1137.36	1121.41	1814.97	8
*vgr	vector	30.75	29.83	27.31	28.18	22.87	27.78
_
Abs	taylor	4730.05	2225.56	1742.09	1590.57	1551.78	2368.01	2
*vgr	taylor	36.76	40.85	38.56	39.41	31.65	37.44
_
Abs	worm	5017.85	2309.86	1894.92	1708.27	1791.11	2544.40	2
*vgr	worm	39.00	42.40	41.95	42.33	36.53	40.44
_
Abs	worm	7082.79	3359.99	2740.50	2502.03	2594.37	3655.93	3
*vgr	worm	55.05	61.68	60.67	62.00	52.92	58.46
_
Abs	worm	8664.78	4162.53	3323.46	3115.36	3378.75	4528.97	4
*vgr	worm	67.35	76.41	73.57	77.20	68.92	72.69
_
Abs	worm	10072.1	5016.57	4032.60	3855.90	4043.11	5404.06	5
*vgr	worm	78.29	92.09	89.27	95.56	82.47	87.53
_
Abs	worm	11678.4	5752.20	4758.80	4493.16	4700.52	6276.60	6
*vgr	worm	90.78	105.60	105.35	111.35	95.88	101.79
_
Abs	worm	12282.3	6021.31	5394.36	4908.98	5141.71	6749.72	7
*vgr	worm	95.47	110.54	119.42	121.65	104.89	110.39
_
Abs	worm	12158.3	5913.50	5784.67	5327.66	5505.71	6937.96	8
*vgr	worm	94.51	108.56	128.06	132.03	112.31	115.09
_
Abs	patton	16028.8	7289.7	4827.9	4392.5	20829.2	53368.1	4
*vgr	patton	124.6	133.83	106.88	108.85	106.23	116.07
.TE
.bp
.NH 1
RELEASE 4.0 STATISTICS
.LP
.TS H
expand box;
l l | n n n n n | n | l.
A/*	H/W	Moss	World	Star	Bldg	M35	Mean	CPU
_
.TH
Abs	vgr	137.82	67.23	56.39	53.91	69.25	76.92	1
*vgr	vgr	1.00	1.00	1.00	1.00	1.00	1.00
_
Abs	slc1	153.63	80.95	62.94	60.48	?	71.60	1
*vgr	slc1	1.12	1.20	1.12	1.13	?	.91
_
Abs	vision	165.85	89.78	76.44	75.78	96.18	100.80	1
*vgr	vision	1.21	1.33	1.36	1.42	1.36	1.33
_
Abs	vmb	381.92	195.11	178.78	186.62	214.31	231.34	1
*vgr	vmb	2.78	2.90	3.18	3.49	3.03	3.07
_
Abs	sws1	636.26	288.76	212.99	210.18	271.02	323.84	1
*vgr	sws1	4.64	4.30	3.79	3.93	3.83	4.09
_
Abs	veto	827.25	436.38	361.89	348.99	422.19	479.34	1
*vgr	veto	6.03	6.50	6.45	6.54	5.97	6.29
_
Abs	vhs	1040.44	500.52	436.90	415.15	496.30	577.86	1
*vgr	vhs	7.59	7.46	7.79	7.78	7.02	7.52
_
Abs	cor3	1203.14	553.22	445.45	398.79	534.18	626.95	1
*vgr	cor3	8.78	8.24	7.94	7.47	7.55	7.99
_
Abs	wilted	1556.66	800.18	722.95	708.21	858.90	929.38	1
*vgr	wilted	11.36	11.93	12.89	13.27	12.15	12.32
_
Abs	shpcrc2	2274.36	1041.79	915.99	854.30	1110.23	1239.33	1
*vgr	shpcrc2	16.59	15.53	16.33	16.01	15.70	16.03
_
Abs	walrus	2291.38	1094.04	960.07	892.67	1066.31	1260.89	1
*vgr	walrus	16.72	16.31	17.12	16.72	15.08	16.39
_
Abs	walrus	4010.55	2019.34	1767.75	1685.19	2029.78	2302.52	2
*vgr	walrus	29.27	30.11	31.52	31.58	28.71	30.23
_
Abs	patton	5235.87	2108.02	1509.85	1507.81	2132.01	2498.71	1
*vgr	patton	38.21	31.43	26.92	28.25	30.15	30.99
_
Abs	crim	4679.97	2246.04	2143.07	1936.85	2224.65	2646.11	1
*vgr	crim	34.15	33.49	38.22	36.29	31.47	34.72
_
Abs	uy1	6220.79	2474.97	1742.33	1748.73	2464.90	2930.34	1
*vgr	uy1	45.40	36.90	31.07	32.77	34.86	36.20
_
Abs	wiltse	5785.83	2737.06	2473.66	2310.58	2699.06	3201.23	1
*vgr	wiltse	42.22	40.81	44.11	43.30	38.18	41.72	-mips1
_
Abs	walrus	5743.85	2938.73	2513.24	2446.85	2858.84	3300.30	3
*vgr	walrus	41.92	43.82	44.82	45.85	40.44	43.37
_
Abs	wiltse	7093.51	3458.89	2887.77	2730.62	3199.08	3873.97	1
*vgr	wiltse	51.77	51.57	51.50	51.17	45.25	50.25	-mips2
_
Abs	walrus	7131.48	3613.02	3262.01	3048.77	3594.35	4129.92	4
*vgr	walrus	52.05	53.87	58.17	57.13	50.84	54.41
_
Abs	whiz	7159.01	3655.65	3355.85	3132.36	3655.02	4191.57	4
*vgr	whiz	52.25	54.51	59.85	58.70	51.70	55.40
_
Abs	walrus	7698.43	4119.06	3889.37	3690.27	4244.30	4728.28	5
*vgr	walrus	56.18	61.42	69.36	69.15	60.04	63.23
_
Abs	walrus	8718.08	4625.56	4469.49	4295.53	4765.40	5374.81	6
*vgr	walrus	63.63	68.97	79.71	80.50	67.41	72.04
_
Abs	walrus	8937.71	4882.66	4911.99	4601.09	5335.14	5733.71	7
*vgr	walrus	65.23	72.81	87.60	86.22	75.47	77.46
_
Abs	walrus	9300.31	5136.17	5243.14	5112.23	5318.45	6022.06	8
*vgr	walrus	67.88	76.59	93.51	95.80	75.23	81.80
_
.TE
.bp
.NH	1
ADDITIONAL STATISTICS
.PP
It would be greatly appreciated if you would e-mail any statistics
that you gather to the BRL-CAD developers, including: the ``summary''
file produced by the benchmark script, which BRL-CAD release was used,
the manufacturer of the machine, hardware model name/number, software
version numbers (OS and compiler), presence or absence of floating
point hardware, processor cache sizes, bus and memory speeds, and the
local host name (preferably an Internet or UUCP host name) of the
specific machine used.