home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!cs.utexas.edu!zaphod.mps.ohio-state.edu!darwin.sura.net!sgiblab!sgigate!sgi!fido!zola!zuni!anchor!olson
- From: olson@anchor.esd.sgi.com (Dave Olson)
- Subject: Re: Need Specmarks of R3K & R4K indigo
- Message-ID: <sj6bdf4@zuni.esd.sgi.com>
- Sender: news@zuni.esd.sgi.com (Net News)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <GREENDOG.92Nov19122209@insti.physics.sunysb.edu>
- Date: Fri, 20 Nov 92 04:40:58 GMT
- Lines: 152
-
- In <GREENDOG.92Nov19122209@insti.physics.sunysb.edu> greendog@insti.physics.sunysb.edu (Michael D. Fischer) writes:
-
- | Hi! I need the SPECfp92 and SPECint92 ratings of the R3 and R4 indigos...
- | Can anyone mail me these ASAP? Thanks...
-
- Well, since they have been published now, I can send them to you,
- I guess. I'd prefer that you don't repost them or mail them
- to somebody else. These aren't quite in the official format,
- but are close.
- ====================
-
- ---------------------------------------------------------------
- Benchmark Ref Actual SPECRatio
- ---------------------------------------------------------------
- 013.spice2g6 24000 541.5 44.3
- 015.doduc 1860 36.3 51.2
- 034.mdljdp2 7090 88.0 80.6
- 039.wave5 3700 89.1 41.5
- 047.tomcatv 2650 40.8 65.0
- 048.ora 7420 101.6 73.0
- 052.alvinn 7690 101.5 75.8
- 056.ear 25500 272.7 93.5
- 077.mdljsp2 3350 76.8 43.6
- 078.swm256 12700 359.2 35.4
- 089.su2cor 12900 176.7 73.0
- 090.hydro2d 13700 182.0 75.3
- 093.nasa7 16800 230.8 72.8
- 094.fpppp 9200 171.5 53.6
- Geometric Mean: SPECfp92 60.3
-
-
- System: pre-MR 50 Mhz Indigo R4000 with 64 Mbytes of RAM, and
- IBM 1.2 Gbyte disk.
- (Results may change slightly with final hardware configuration)
-
- Performance changes: I don't know; I don't have the original source
- to compare against. These are the versions I got from Mike
- Humphrey.
-
- Compiler options used (SGI MR'ed 3.10 compilers were used):
-
- 013.spice2g6
- -O3 -mips2 -Olimit 1000 -G 8200 -noinline -noisam
- -DUNDERBAR=1 -cckr -j -jmpopt -no_mp
-
- 015.doduc
- -O3 -mips2 -Olimit 2000 -G 10216 -w1
- -sopt,-inl,-ind=10,-inll=10,-inff=.,-ur=1,-r=3,-chs=8 -noisam
- -j -jmpopt -lF77 -lfastm -lm -lU77 -lI77 -no_mpc
-
- 034.mdljdp2
- -O3 -mips2
- -sopt,-ro=3,-o=5,-ur=8,-ur2=3000,-limit=10000,-arclm=10000,-so=3,-chs=8
- -w -Olimit 1200 -G 8200 -j -lfastm -lm43
-
- 039.wave5
- -O3 -mips2 -G 8200 -sopt,-so=3,-r=3,-ur=8,-chs=8 -j -jmpopt
- -lfastm
-
- 047.tomcatv
- -O3 -mips2 -Olimit 2000 -G 8200 -static -ddopt -j
-
- 048.ora
- -O3 -mips2 -Olimit 2000 -G 5716 -Wo,-loopunroll,8
- -sopt,-r=3,-so=3,-ur=8,-ur2=77000,-inl=ABC,-inll=3,-ind=3,-ag=a,-chs=8
- -K -j -jmpopt -lfastm
-
- 052.alvinn
- -O3 -mips2 -Olimit 2000 -G 10304 -Wo,-loopunroll,2 -j -jmpopt
- -lm
-
- 056.ear
- -O3 -mips2 -Olimit 2000 -G 8200 -Wo,-loopunroll,8 -j -jmpopt
- -lfastm -lm43
-
- 077.mdljsp2
- -O3 -mips2 -Olimit 2000 -G 4800 -j -jmpopt -lfastm -lm43
-
- 078.swm256
- -O3 -mips2 -Olimit 2000 -G 4800 -ddopt -K
- -sopt,-so=3,-r=3,-inl,-chs=8 -j -jmpopt
-
- 089.su2cor
- -O2 -mips2 -Olimit 1060 -G 16384 -Wo,-loopunroll,4 -ddopt
- -sopt,-inline=matmat,-ind=10,-inll=4,-so=3,-r=3,-chs=8,-ur=8
- -j -jmpopt -lfastm -lm43
-
- 090.hydro2d
- -O3 -mips2 -Olimit 2000 -G 4800 -sopt,-so=3,-r=3,-ur=6,-chs=8
- -static -j -jmpopt -lfastm -lm43
-
- 093.nasa7
- -O3 -mips2 -inline_to 1 -G 8200 -DUNDERBAR -DTIMES
- -sopt,-inl=mxm,-ag=a,-o=5 -noisam -j -jmpopt -lF77 -lfastm
- -lm -no_mpc
-
- 094.fpppp
- -O3 -mips2 -Olimit 1000 -G 14432 -Wo,-loopunroll,0 -Nc40
- -sopt,-inl,-ind=10,-inll=10,-inff=.,-ur=4,-r=3,-chs=8 -noisam
- -j -jmpopt -lF77 -lfastm -lm -no_mpc
-
- ==================
-
- ---------------------------------------------------------------
- Benchmark Ref Actual SPECRatio
- ---------------------------------------------------------------
- 008.espresso 2270 41.1 55.2
- 022.li 6210 95.0 65.4
- 023.eqntott 1100 14.4 76.4
- 026.compress 2770 66.9 41.4
- 072.sc 4530 61.9 73.2
- 085.gcc 5460 124.8 43.8
- Geometric Mean: SPECint92 57.6
-
-
- System: pre-MR 50 Mhz Indigo R4000 with 64 Mbytes of RAM, and
- IBM 1.2 Gbyte disk.
- (Results may change slightly with final hardware configuration)
-
- Portability changes: changed /bin/cc to plain cc for 023.eqntott
-
- Performance changes: I don't know; I don't have the original source
- to compare against. These are the versions I got from Mike
- Humphrey.
-
- Compiler options used (SGI MR'ed 3.10 compilers were used):
-
- 008.espresso
- -O3 -mips2 -G 8200 -sopt -Wb,-peep,-symregs,-swpipe,-xbb
- -cckr -j -jmpopt
-
- 022.li
- -O3 -mips2 -G 1200 -noinline -cckr -j -jmpopt
-
- 023.eqntott
- -O3 -mips2 -Olimit 1200 -G 8200 -ddopt
- -sopt,-r=3,-stdio,-ur=8,-ur2=400,-chs=8 -cckr -j -DSYSV
-
- 026.compress
- -O3 -mips2 -Olimit 2000 -G 8200 -cckr -j -jmpopt -lfastm
- -lm43 -lm
-
- 072.sc
- -O2 -Olimit 2000 -G 8200 -c -DSYSV3 -DSIGVOID -jmpopt
- -lcurses -lPW -lm -lblas
-
- 085.gcc
- -O3 -mips2 -Olimit 1200 -I./config -cckr -j -jmpopt
- --
- Let no one tell me that silence gives consent, | Dave Olson
- because whoever is silent dissents. | Silicon Graphics, Inc.
- Maria Isabel Barreno | olson@sgi.com
-