Difference between revisions of "Rocket.txt"
Jump to navigation
Jump to search
(cosmetics: moved the results to compiler lines for those that hadn't yet) |
|||
| (9 intermediate revisions by 4 users not shown) | |||
| Line 1: | Line 1: | ||
Easier to maintain here, so: | Easier to maintain here, so: | ||
| − | Some results of "rocket 0.00001 159.0": http:// | + | Some results of "rocket 0.00001 159.0": http://wtf.hijacked.us/~andrew/rocket.tar |
See gcc --target-help for some system specific flags. | See gcc --target-help for some system specific flags. | ||
| − | Note that this program wasn't really intended to be a benchmark, but it's a typical piece of code that seems CPU bound enough to provide a good idea of what the CPU can really do. If it reports a final altitude of 0.00m, you're doing it wrong. You should get something like 112317.72m. | + | Note that this program wasn't really intended to be a benchmark, but it's a typical piece of code that seems CPU bound enough to provide a good idea of what the CPU (or, more to the point, the FPU) can really do. If it reports a final altitude of 0.00m, you're doing it wrong. You should get something like 112317.72m. |
| − | Report normalized | + | Report a normalized value of your best result as well (clockrate * seconds). |
| − | + | Report the "real" value, not "sys". | |
| + | |||
| + | 1360 PC (mtl) | ||
| + | Intel Core i7-7700(kaby lake) 3.6GHz(4GHz turbo) running Linux 5.4.38-1 x86-64 | ||
| + | 0.34s gcc-9.3.0 -O3 -march=skylake -mtune=skylake -funroll-loops -funswitch-loops -funsafe-math-optimizations | ||
| + | |||
| + | 1750 PC (arachnotron after replacing the guts spring 2012; plan9) | ||
| + | Intel Core i7 3770K 3.5GHz running Linux 3.2.0 x86-64 | ||
| + | 0.50s gcc-4.6.3 -O3 -fomit-frame-pointer -finline -mtune=corei7 -funroll-loops | ||
2397 Sun Ultra 24 (Vagabond) | 2397 Sun Ultra 24 (Vagabond) | ||
| Line 16: | Line 24: | ||
0.87s gcc-3.4.6 (pkgsrc) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=i686 | 0.87s gcc-3.4.6 (pkgsrc) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=i686 | ||
1.29s gcc-3.4.3 (sun freeware) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math | 1.29s gcc-3.4.3 (sun freeware) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math | ||
| + | |||
| + | 2207 PC (mtl) | ||
| + | Intel Core i7 920 2.66Gz running Linux 3.3.4-2 64bit | ||
| + | 0.83s gcc-4.7.0 -Os -fomit-frame-pointer -finline -mtune=corei7 -funroll-loops -funswitch-loops -ffast-math -msse3 -mmmx -fgcse-after-reload | ||
2763 PC (SanityInFlux) | 2763 PC (SanityInFlux) | ||
| Line 30: | Line 42: | ||
1.21s gcc-4.3.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math | 1.21s gcc-4.3.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math | ||
| − | PC (plan9) | + | PC (arachnotron; plan9) |
AMD64 X2 4200+ running Linux 2.6.22-14-generic 32 bit | AMD64 X2 4200+ running Linux 2.6.22-14-generic 32 bit | ||
1.37s gcc -O3 -funroll-loops -mmmx -msse -msse2 -msse3 -fomit-frame-pointer -ffast-math | 1.37s gcc -O3 -funroll-loops -mmmx -msse -msse2 -msse3 -fomit-frame-pointer -ffast-math | ||
| Line 103: | Line 115: | ||
Via C7 1.8ghz | Via C7 1.8ghz | ||
7.18s gcc-4.2.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=nocona | 7.18s gcc-4.2.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=nocona | ||
| + | |||
| + | 8004 HP 9000 C3600 (mancubus; plan9) | ||
| + | HP PA-8600 552 MHz running OpenBSD | ||
| + | 7.92s gcc-4.2.1 -O3 -funroll-loops (OpenBSD 5.0) | ||
| + | 14.5s gcc-(2.95?) -march=2.0 -mpa-risc-2-0 -O3 -funroll-loops (OpenBSD 4.3) | ||
6840 Weird rackmountable PC (jontow) | 6840 Weird rackmountable PC (jontow) | ||
| Line 138: | Line 155: | ||
Quad Sun UltraSparc-IIi 450 MHz running Solaris 10 | Quad Sun UltraSparc-IIi 450 MHz running Solaris 10 | ||
13.8s gcc -funroll-loops -mcpu=ultrasparc -O3 | 13.8s gcc -funroll-loops -mcpu=ultrasparc -O3 | ||
| − | |||
| − | |||
| − | |||
| − | |||
4850 Cobalt Raq2 (SanityInFlux) | 4850 Cobalt Raq2 (SanityInFlux) | ||
| Line 200: | Line 213: | ||
Sun Sun4m 70mhz running NetBSD 2.1 | Sun Sun4m 70mhz running NetBSD 2.1 | ||
4m 49s gcc -funroll-loops -mcpu=v7 -O3 | 4m 49s gcc -funroll-loops -mcpu=v7 -O3 | ||
| + | |||
| + | 21880 Sun SparcStation 2 (dragon_cz) | ||
| + | Sun Sun4c 40 MHz running OpenBSD 4.8 | ||
| + | 9m 7s gcc -funroll-loops -mcpu=v7 -O3 | ||
19140 NeXT NeXTStation Turbo Color (dragon_cz) | 19140 NeXT NeXTStation Turbo Color (dragon_cz) | ||
Motorola M68040 33 MHz running NeXTStep 3.3 | Motorola M68040 33 MHz running NeXTStep 3.3 | ||
9m 40s gcc-2.95.3 -O3 -ffast-math -m68040 | 9m 40s gcc-2.95.3 -O3 -ffast-math -m68040 | ||
Latest revision as of 14:12, 8 May 2020
Easier to maintain here, so:
Some results of "rocket 0.00001 159.0": http://wtf.hijacked.us/~andrew/rocket.tar
See gcc --target-help for some system specific flags.
Note that this program wasn't really intended to be a benchmark, but it's a typical piece of code that seems CPU bound enough to provide a good idea of what the CPU (or, more to the point, the FPU) can really do. If it reports a final altitude of 0.00m, you're doing it wrong. You should get something like 112317.72m.
Report a normalized value of your best result as well (clockrate * seconds).
Report the "real" value, not "sys".
1360 PC (mtl)
Intel Core i7-7700(kaby lake) 3.6GHz(4GHz turbo) running Linux 5.4.38-1 x86-64
0.34s gcc-9.3.0 -O3 -march=skylake -mtune=skylake -funroll-loops -funswitch-loops -funsafe-math-optimizations
1750 PC (arachnotron after replacing the guts spring 2012; plan9)
Intel Core i7 3770K 3.5GHz running Linux 3.2.0 x86-64
0.50s gcc-4.6.3 -O3 -fomit-frame-pointer -finline -mtune=corei7 -funroll-loops
2397 Sun Ultra 24 (Vagabond)
Intel Core2 Duo E8600 3.33gHz running Solaris 10 10/08 i386
0.72s gcc-4.3.2 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=i686
0.87s gcc-3.4.6 (pkgsrc) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=i686
1.29s gcc-3.4.3 (sun freeware) -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math
2207 PC (mtl)
Intel Core i7 920 2.66Gz running Linux 3.3.4-2 64bit
0.83s gcc-4.7.0 -Os -fomit-frame-pointer -finline -mtune=corei7 -funroll-loops -funswitch-loops -ffast-math -msse3 -mmmx -fgcse-after-reload
2763 PC (SanityInFlux)
AMD Phenom II 940 overclocked to 3.33gHz running Linux 2.6.28
0.83s gcc-4.1.2 -march=i686 -O3 -msse3 -funroll-loops -funsafe-math-optimizations -mfpmath=sse
2234 PC (dragon_cz)
Intel Core2Duo E8200 2.66 GHz running Linux 2.6.24-23 64bit
0.84 gcc-4.1 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -mtune=nocona
0.87 icc-11.0 -fast -xSSE3 -msse3 -fp-model fast=2 -fp-speculation=fast
2250 IBM ThinkPad X200s (dragon_cz)
Intel Core2Duo SL9400 @ 1.86GHz
1.21s gcc-4.3.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math
PC (arachnotron; plan9)
AMD64 X2 4200+ running Linux 2.6.22-14-generic 32 bit
1.37s gcc -O3 -funroll-loops -mmmx -msse -msse2 -msse3 -fomit-frame-pointer -ffast-math
1.55s icc -fp-model fast=2 -msse2 -O3 -funroll-loops
PC (ghostn9e)
Intel Q6600 running Linux 2.6.24
1.38s gcc -O3 -funroll-loops -mmmx -msse -msse2 -msse3 -fomit-frame-pointer -funsafe-math-optimizations
-m128bit-long-double
3132 PC (dragon_cz)
AthlonXP 2500+ 1.8GHz running Linux 2.6.21.5-smp
1.74s gcc-4.3.3 -O3 -funroll-loops -m3dnow -pipe -fomit-frame-pointer -ffast-math -march=athlon-xp
2.16s gcc-4.1.2 -O3 -funroll-loops -m3dnow -pipe -fomit-frame-pointer -funsafe-math-optimizations
3536 PC (jontow)
Intel Core2duo 1.6ghz running Linux 2.6.25-gentoo-r7 32-bit
2.21s gcc-4.1.2 -O3 -pipe -funroll-loops -funsafe-math-optimizations -march=i686 -finline -msse2
4.33s gcc-3.3.6 -O3 -pipe -funroll-loops -funsafe-math-optimizations -march=i686 -finline -msse2
3291 IBM RS/6000 IntelliStation POWER 275 (9114-275) (dragon_cz)
POWER4 1.45GHz running AIX 6.1.0.0 64bit
2.27s gcc-4.2.0 -O3 -mtune=power4 -mcpu=power4 -maltivec -mhard-float -mpowerpc64 -maix64 -funroll-loops
-funsafe-math-optimizations -mmultiple -mmfcrf -lm -pipe
2.60s xlc-10.1 -lm -qtune=pwr4 -Q -qunroll=yes -O3
3648 PC (fernandotcl)
AMD Sempron(tm) Processor 2800+ @ 1.6GHz 2.6.24-16-generic 32-bit
2.28s gcc -03 -finline -funroll-loops -mmmx -msse -msse2 -msse3
4272 Embedded via board
Via nano 1.6ghz running FreeBSD 7.1
2.67s gcc-4.2.1 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=nocona
3397 Dell Latitude X1 (dragon_cz)
Intel Pentium M 1.1GHz ULV running Linux 2.6.21.5-smp
3.069s gcc-4.1.2 -O3 -funsafe-math-optimizations -funroll-loops -msse2 -fomit-frame-pointer -mfpmath=sse -m128bit-long-double -pipe
7800 Sun Fire x4100 (jontow)
AMD Opteron(tm) 2216 (2 * dual-core 2.4GHz) 32 bit
3.25s gcc -O3 -funroll-loops -mmmx -msse -msse2 -msse3 -funsafe-math-optimizations
3810 PC (SanityInFlux)
Pentium III (Coppermine) 1gHz Linux 2.6.24.4
3.81s gcc-4.1.2 -march=pentium3 -O3 -pipe -funroll-loops -funsafe-math-optimizations
5.34s icc -fp-model fast=2
7.45s sunstudio12/bin/cc -lm -fast -fstore
7.72s gcc-3.3.6 -march=pentium3 -O3 -pipe -funroll-loops -finline -fomit-frame-pointer -lm
3950 HP C8000 (Vagabond)
HP PA-RISC PA-8800 1Ghz (dual core) running HP-UX 11.11
3.95s gcc 4.4.0 -march=2.0 -mfast-indirect-calls -mlinker-opt -O3 -funroll-loops -fomit-frame-pointer -ffast-math -mpa-risc-2-0 -mschedule=8000
3255 HP Visualize 9000 C3700
HP PA-RISC PA-8700 750Mhz running HP-UX 11.11
4.34s gcc-4.2.3 -march=2.0 -mfast-indirect-calls -mlinker-opt -O3 -funroll-loops -fomit-frame-pointer -ffast-math
5.96s cc +O3 +Oaggressive +Onolimit +Opipeline +Ovectorize +Ofastaccess +Onofltacc +Olibcalls +Oloop_unroll +Ostatic_prediction +Oprofile=use
3357 Asus Eee 701 (rob_ert)
Intel Celeron-M ULV 900MHz clocked at 630MHz, running Linux 2.6.26-1 i686
5.33s gcc-4.3.2 -O3 -funroll-loops -msse2 -fomit-frame-pointer -ffast-math -mtune=pentium4m
5500 Apple PowerBook G4 12" (Vagabond)
Apple G4 1.0 GHz running MacOS X 10.4.11
5.5s gcc -funroll-loops -mtune=powerpc -m32 -O3
2356 IBM RS/6000 44P-170 (plan9)
POWER3 400 MHz running AIX 5.3 64 bit
5.89s gcc -mtune=power3 -mcpu=power3 -O3 -funroll-loops -funsafe-math-optimizations -fomit-frame-pointer -mmultiple -mmfcrf
12924 Epia SN-18000g
Via C7 1.8ghz
7.18s gcc-4.2.3 -O3 -funroll-loops -msse3 -fomit-frame-pointer -ffast-math -march=nocona
8004 HP 9000 C3600 (mancubus; plan9)
HP PA-8600 552 MHz running OpenBSD
7.92s gcc-4.2.1 -O3 -funroll-loops (OpenBSD 5.0)
14.5s gcc-(2.95?) -march=2.0 -mpa-risc-2-0 -O3 -funroll-loops (OpenBSD 4.3)
6840 Weird rackmountable PC (jontow)
Pentium 3/855 running FreeBSD 6.2
8.0s gcc -O3 -mtune=i686 -funroll-loops -finline -mmmx -msse
5665 Dell PowerEdge somethingorother tower (jontow)
Pentium 3/550mhz running FreeBSD 6.3
10.3s gcc -O3 -mtune=i686 -funroll-loops -finline -mmmx -msse
5285 Digital Personal Workstation 500au (Vagabond)
DEC Alpha EV56 500 MHz running NetBSD 4.0
10.57s gcc -O3 -mcpu=ev56 -funroll-loops -mno-soft-float
4815 Sun Ultra 60 (plan9)
Dual Sun UltraSPARC-IIi 450 MHz running Solaris 10
10.7s suncc -m32 -O4 -fns -fsingle -fsimple=2 -xprofile=use:rocket3 -xlinkopt=2 -xautopar -xpagesize=8K
11.3s suncc -m32 -O5 -Xc -xprofile=use:rocket3 -xlinkopt
13.4s gcc -O3 -mcpu=ultrasparc -m64 -funroll-loops -mvis -mrelax
3240 SGI Octane (Vagabond)
Dual MIPS R12000 300 MHz running IRIX64 6.5.29
10.8s gcc ...
5080 Sun Ultra 30 (dragon_cz)
Sun UltraSparc-IIi 400MHz running Solaris 10
12.7s gcc-3.4.3 -mrelax -mhard-float -O3 -mvis -mcpu=ultrasparc -funroll-loops -fomit-frame-pointer
-funsafe-math-optimizations -pipe
4555 Apple iMac G3 (revision D)
PowerPC G3 333mhz running NetBSD 4.0.1
13.68 gcc-4.1.2 -O3 -funroll-loops -ffast-math -mcpu=G3 -mtune=G3
6210 Sun Ultra 80 (Vagabond)
Quad Sun UltraSparc-IIi 450 MHz running Solaris 10
13.8s gcc -funroll-loops -mcpu=ultrasparc -O3
4850 Cobalt Raq2 (SanityInFlux)
MIPS RM5231 250 MHz running Linux 2.6.28.5-mipsgit-20090215-cobalt
19.4s gcc-4.1.2 -O2 -mips4 -mabi=32 -pipe -ffast-math -funroll-loops -funsafe-math-optimizations
8580 Sun Netra T1/105 (jontow)
Sun UltraSPARC-IIi 440 MHz running OpenBSD 4.2 64 bit
19.5s gcc -O3 -mcpu=v9 -mtune=ultrasparc -funroll-loops
8250 PC (stormblast) (plan9)
Intel Pentium II 300 MHz running OpenBSD 3.9
27.5s gcc -O3 -fomit-frame-pointer -funroll-loops -funsafe-math-optimizations -march=i686
8700 Sun Ultra 10 (jontow)
Sun UltraSPARC-IIi 300 MHz running OpenBSD 4.2 64 bit
29.0s gcc-3.3.5 -O3 -mcpu=v9 -mtune=ultrasparc -funroll-loops -funsafe-math-optimizations -finline
31.3s gcc-3.3.5 -O3 -mcpu=v9 -mtune=ultrasparc -funroll-loops
9210 DEC Alphaserver 1000 5/300 (andrew)
DEC Alpha ev56 300mhz running OpenBSD 4.1
30.7 gcc-3.3.5 -O3 -mcpu=ev56 -funroll-loops -mno-soft-float
5634 SGI Indy
MIPS R5000 180 MHz running IRIX 6.5.22
31.3s gcc-3.4.6 -O3 -mips4 -mtune=vr5000 -mhard-float -ffast-math -funroll-loops -funsafe-math-optimizations
38.4s (MIPSpro 7.2.1)c89 -Ofast
7972 PC (SanityInFlux)
Pentium 133mHz running Linux 2.6.26.3
59.94s gcc-4.1.2 -march=i386 -O -funroll-loops -finline -ffast-math
65.17s gcc-3.4.6 -march=pentium -O3 -pipe -funroll-loops -ffast-math
74.79s (ICC 10.0 20070809)icc -fp-model fast=2 -fpic -align -no-complex-limited-range
102.94s (Sun C 5.9)cc -lm -fast -xarch=386 -fstore
4866 Sun SparcStation 10 (jontow)
Sun Sun4m(??) 60 MHz running NetBSD 4.0
81.1s gcc -funroll-loops -mcpu=v7 -O3
7280 HP 9000 712/80 (rob_ert)
PA-7100SC PA-RISC 1.1c 80MHz running OpenBSD 4.6
1m 31s gcc-3.3.5 -Wall -O3
11700 SGI Indy (dragon_cz)
MIPS R4600 100MHz running IRIX 5.3
1m 57s gcc-3.4.6 -mhard-float -O3
2m 7s cc -mips2 -O3
18480 Sun SparcStation 5 (dragon_cz)
Sun Sun4m 110mhz running OpenBSD 4.4
2m 48s gcc -funroll-loops -mcpu=v7 -O3
8450 HP (Apollo) 9000/710
PA-7000 1.1a 50 MHz running HP-UX 10.20
2m 49s gcc -O3 -funroll-loops -fomit-frame-pointer (gcc 2.95.2)
20230 Sun SparcStation 5 (jontow)
Sun Sun4m 70mhz running NetBSD 2.1
4m 49s gcc -funroll-loops -mcpu=v7 -O3
21880 Sun SparcStation 2 (dragon_cz)
Sun Sun4c 40 MHz running OpenBSD 4.8
9m 7s gcc -funroll-loops -mcpu=v7 -O3
19140 NeXT NeXTStation Turbo Color (dragon_cz)
Motorola M68040 33 MHz running NeXTStep 3.3
9m 40s gcc-2.95.3 -O3 -ffast-math -m68040