Hardware Information

On a lot of benchmark sites on the net, I miss detailed information about the hardware used. On this page, I try to be as complete as possible with all relevant hardware information about all the computers tested with RGBench.

NOTE: Due to the amount of data, this table is wider than normal!

Alpha Workstations

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

A01 DEC Alpha AXP 21164 @ 500+ MHz DEC-Unix? ??
??@??MHz
??@??MHz

A02 DEC Alpha AXP 21164 @ 300 MHz DEC-Unix? ??
??@??MHz
??@??MHz

Apple Macintosh

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

P01 PowerMac 7500/100 PowerPC 601 @ 100 MHz MacOS 8.0 1x32 kB
unified 256 kB@50MHz
Boardcache 32@50MHz
FPM interleaved reference machine

P01a PowerMac 7500/100 PowerPC 601 @ 100 MHz MkLinux DR2.1.5 1x32 kB
unified 256 kB@50MHz
Boardcache 32@50MHz
FPM interleaved compiler options used: -O3 -ffast-math-finline -functions -funroll-loops -fexpensive-optimizations

P02 PowerMac 7500/200 PowerPC 604e @ 200 MHz MacOS 2x32 kB
256 kB@50MHz
Boardcache 32@50MHz
FPM interleaved new CPU-card installed (Maccelerate by Phase 5)

P03 PowerMac G3 PowerPC 750 @ 266 MHz MacOS 2x32 kB
512 kB@133MHz
Backsidecache 64@66MHz
SDRAM

P04 PowerBook 3400 PowerPC 603v @ 200 MHz MacOS 2x16 kB
256 kB@40MHz
Boardcache ??@40MHz

P05 Umax Pulsar PowerPC 604e @ 250 MHz MacOS 2x32 kB
512 kB@50MHz
Boardcache 80@50MHz
FPM part. interl.

P05a Umax Pulsar PowerPC 604e @ 250 MHz MacOS 2x32 kB
512 kB@50MHz
Boardcache 272@50MHz
FPM interleaved

P06 PowerMac 8500/180 PowerPC 604e @ 180 MHz MacOS 2x32 kB
256 kB@50MHz
Boardcache ??@45MHz
FPM interleaved?

P06a PowerMac 8500/180 PowerPC 604e @ 180 MHz MacOS 2x32 kB
1024 kB@50MHz
Boardcache ??@45MHz
FPM interleaved?

P07 PowerMac 8500/G3 PowerPC 750 @ 266 MHz MacOS 2x32 kB
1024 kB@133MHz
Backsidecache ??@48MHz
FPM interleaved? new CPU-card installed (Sonnet Crescendo G3)

P07a PowerMac 8500/G3 PowerPC 750 @ 266 MHz MacOS 2x32 kB
1024 kB@50MHz
Boardcache ??@45MHz
FPM interleaved? new CPU-card installed (Sonnet Crescendo G3), BS-cache disabled

P07b PowerMac 8500/G3 PowerPC 750 @ 266 MHz MacOS 2x32 kB
0@0MHz
??@45MHz
FPM interleaved? new CPU-card installed (Sonnet Crescendo G3), L2- and BS-cache disabled

P08 PowerMac G3 /233 PowerPC 750 @ 233 MHz MacOS 2x32 kB
512 kB@116.6MHz
Backsidecache ??@66MHz
SD-RAM standard PowerMac G3/233

P08a PowerMac G3 /233 PowerPC 750 @ 233 MHz MacOS 2x32 kB
0@0MHz
??@66MHz
SD-RAM PowerMac G3/233 with BS-Cache disabled

P08b PowerMac G3 /233 PowerPC 750 @ 233 MHz MacOS 2x32 kB
512 kB@155.5MHz
Backsidecache ??@66MHz
SD-RAM PowerMac G3/233 but with the BS-clock-ratio tuned to 1:1.5

P08c PowerMac G3 /233 PowerPC 750 @ 263 MHz MacOS 2x32 kB
512 kB@131.6MHz
Backsidecache ??@75MHz
SD-RAM PowerMac G3/233 with tuned systembus and CPU

P08d PowerMac G3 /233 PowerPC 750 @ 266 MHz MacOS 2x32 kB
512 kB@89.1MHz
Backsidecache ??@66MHz
SD-RAM PowerMac G3/233 but tuned CPU while the BS-clock-ratio is reduced to 1:3

P08e PowerMac G3 /233 PowerPC 750 @ 266 MHz MacOS 2x32 kB
512 kB@133MHz
Backsidecache ??@66MHz
SD-RAM PowerMac G3/233 with tuned CPU

P08f PowerMac G3 /233 PowerPC 750 @ 266 MHz MacOS 2x32 kB
512 kB@133MHz
Backsidecache ??@66MHz
SD-RAM PowerMac G3/233 with tuned CPU and libmoto

P09 PowerBook G3(neu) PowerPC 750 @ 250 MHz MacOS 2x32 kB
1024 kB@125MHz
Backsidecache ??@83MHz
SD-RAM model of the new PowerBook G3 line (as of Maz 98)

P10 PowerMac 8500/G3 PowerPC 750 @ 300 MHz MacOS 2x32 kB
1024 kB@125MHz
Backsidecache ??@40MHz
FPM interleaved? PowerMac 8500 with PowerLogoix PowerForce 300/300 G3-card

P10a PowerMac 8500/G3 PowerPC 750 @ 180 MHz MacOS 2x32 kB
1024 kB@90MHz
Backsidecache ??@45MHz
FPM interleaved? PowerMac 8500 with PowerLogoix PowerForce 300/300 G3-card, but reduced clock

P10b PowerMac 8500/G3 PowerPC 750 @ 180 MHz MacOS 2x32 kB
1024 kB@180MHz
Backsidecache ??@45MHz
FPM interleaved? PowerMac 8500 with PowerLogoix PowerForce 300/300 G3-card, but reduced clock

P10c PowerMac 8500/G3 PowerPC 750 @ 266 MHz MacOS 2x32 kB
1024 kB@266MHz
Backsidecache ??@44.3MHz
FPM interleaved? PowerMac 8500 with PowerLogoix PowerForce 300/300 G3-card, but reduced clock

P10d PowerMac 8500/G3 PowerPC 750 @ 333 MHz MacOS 2x32 kB
1024 kB@222MHz
Backsidecache ??@44.3MHz
FPM interleaved? PowerMac 8500 with PowerLogoix PowerForce 300/300 G3-card, but tuned CPU and reduced Cache-clock

P11 Umax Pulsar PowerPC 604e @ 240 MHz MacOS 2x32 kB
512 kB@60MHz
Boardcache 208@60MHz
FPM, interleaved, tuned RAM-Access tuned to x-2-2-2

P12 PowerMac 7500/G3 PowerPC 750 @ 310 MHz MacOS 2x32 kB
512 kB@124MHz
Backsidecache ??@??MHz
&nsbsp; PowerMac with PowerLogix 220/110 tuned to 310MHz/124MHz

P12a PowerMac 7500/G3 PowerPC 750 @ 310 MHz MacOS 2x32 kB
512 kB@124MHz
Backsidecache ??@??MHz
&nsbsp; PowerMac with PowerLogix 220/110 tuned to 310MHz but BS-Cache disabled

P13 PowerMac B&W G3/350 PowerPC 750 @ 350 MHz MacOS 2x32 kB
1024 kB@175MHz
Backsidecache 192@100MHz
SDRAM Mac OS 8.5.1

P13a PowerMac B&W G3/350 PowerPC 750 @ 350 MHz Linux 2x32 kB
1024 kB@175MHz
Backsidecache 192@100MHz
SDRAM compiled using gcc 2.95, running LinuxPPC 1999 with 2.2.10 kernel

P14 PowerMac 7500/G3 300 PowerPC 750 @ 300 MHz MacOS 2x32 kB
1024 kB@200MHz
Backsidecache 96@40MHz
FPM interleaved MacOS 8.1; Same PowerMac as in P1 but with MacTell 300/200 Mhz G3-card (Zif-Carrier)

P14a PowerMac 7500/G3 300 PowerPC 750 @ 300 MHz Linux 2x32 kB
1024 kB@200MHz
Backsidecache 96@40MHz
FPM interleaved compiled using egcs 1.1.1, running LinuxPPC R4 with 2.2.6 kernel, same machine as P14

P14b PowerMac 7500/G3 300 PowerPC 750 @ 300 MHz MacOS 2x32 kB
1024 kB@200MHz
Backsidecache 96@50MHz
FPM non-interleaved same machine as P14

P14c PowerMac 7500/G3 300 PowerPC 750 @ 300 MHz MacOS 2x32 kB
1024 kB@150MHz
Backsidecache 96@40MHz
FPM interleaved same machine as P14

P14d PowerMac 7500/G3 300 PowerPC 750 @ 320 MHz MacOS 2x32 kB
1024 kB@213MHz
Backsidecache 96@40MHz
FPM interleaved same machine as P14

P15 Umax Pulsar G3 PowerPC 750 @ 400 MHz MacOS 2x32 kB
1024 kB@200MHz
Backsidecache 208@50MHz
FPM interleaved same machine as P14

P16 PowerMac G4 400 (Yikes) PowerPC 7400 @ 400 MHz MacOS 2x32 kB
1024 kB@200MHz
Backsidecache 64@100MHz
SDRAM same machine as P14

P17 Umax Pulsar G3 PowerPC 750 @ 325 MHz MacOS 2x32 kB
1024 kB@163MHz
Backsidecache 272@??MHz
FPM interleaved same machine as P14

P17a Umax Pulsar G3 PowerPC 750 @ 325 MHz MacOS 2x32 kB
0 kB@0MHz
disabled 272@??MHz
FPM interleaved same machine as P14

P18 PowerBook G3/333 PowerPC 750 @ 333 MHz MacOS 2x32 kB
512 kB@133MHz
Backsidecache ???@66MHz
SDRAM

P19 Apple iBook PowerPC 750 @ 300 MHz Linux 2x32 kB
512 kB@150MHz
Backsidecache 64@66MHz
SDRAM LinuxPPC 2000; gcc 2.95.2; 2.2.15.pre3

P20 Apple PowerMac G4/450 PowerPC 7400 @ 500 MHz MacOS 2x32 kB
1024 kB@250MHz
Backsidecache 256@100MHz
SDRAM (CL2) Sawthooth-Board, overclock to 500 MHz, VM & Speculative Processing on, Mac 9.04

P20a Apple PowerMac G4/450 PowerPC 7400 @ 466 MHz MacOS 2x32 kB
1024 kB@233MHz
Backsidecache 256@133MHz
SDRAM (CL2) Sawthooth-Board, overclock to 466 MHz, bus overclocked to 133MHz, VM & Speculative Processing on, Mac 9.04

P21 Apple iMac G5 PowerPC 970 @ 1800 MHz MacOS X, 10.3 32 kB Data +64kB Intr.
512k kB@1800MHz
Backsidecache on Die 256@166MHz
DDR333 compiled with gcc 3.3 and -mcpu=G5 -mtune=G5; only Single-Channel RAM-Access (only one DIMM)

P22 PowerMac 8600 PowerPC 604e @ 250 MHz MacOS 2x32 kB
1024 kB@100MHz
Inline-Cache ??@50MHz
FPM interleaved?

P23 Apple iBook PowerPC 750CXe @ 500 MHz MacOS X, 10.3 2x32 kB
512 kB@500MHz
Backsidecache On Die 384@66MHz
SDRAM compiled with gcc 3.3 and -mcpu=750 -mtune=750; MacOS X 10.3.5

P24 Apple Mac mini PowerPC 74xx @ 1420 MHz MacOS X 10.3.9 2x32 kB
512 kB@1420MHz
Backsidecache on Die 1024@166MHz
DDR-RAM compiled with gcc 3.3 and -mcpu=7450 -mtune=7450; MacOS X 10.3.9

P25 Apple PowerMac Dual G5 2.3GHz PowerPC 970 @ 2.3 GHz MHz Linux 64bit 32kB+64kB
512kB@on DieMHz
2300 ???@200MHzMHz
DDR-RAM compiled with -g -O3 -mcpu=970 -mtune=970 -maltivec -fexpensive-optimizations -funroll-loops -fexpensive-optimizations

P25a Apple PowerMac Dual G5 2.3GHz PowerPC 970 @ 2.3 GHz MHz Linux 64bit 32kB+64kB
512kB@on DieMHz
2300 ???@200MHzMHz
DDR-RAM compiled with -g -O3 -mcpu=970 -mtune=970 -maltivec -fexpensive-optimizations -funroll-loops -fexpensive-optimizations -pthread -mpowerpc64 -m64 -mabi=altivec -fthread-jumps

P26 IBM p630 Power4+ @ 1.45GHz MHz Linux 64bit 32kB+64kB
`1.5MB' 1200 'L2-Cache@sharedMHz
by two@CoresMHz
+ Cache

P27 IBM S390 ??? @ ??? MHz Linux 64bit ???
???@???MHz
???@???MHz
gcc

P28 Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 1 core used

P28a Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64 -maltivec; 1 core used

P28b Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28c Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28d Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28e Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM YellowDogLinux 4.1; compiled using 4.0 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28f Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 1 core used

P28g Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64 -maltivec; 1 core used

P28h Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28i Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28j Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28k Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28l Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 2 core used

P28m Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 4 core used

P28n Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 8 processes on 4 cores

P28o Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 4.0 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 16 processes on 4 cores

P28p Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 1 core used

P28q Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64 -maltivec; 1 core used

P28r Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28s Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28t Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32; 1 core used

P28u Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.0 with -O3 -mcpu=7450 -mtune=7450 -fexpensive-optimizations -funroll-loops -m32 -maltivec; 1 core used

P28v Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 2 core used

P28w Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 4 core used

P28x Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 8 process on 4 cores

P28y Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz Linux 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM OpenSuSE 10.1; compiled using 4.1 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops -m64; 16 processes on 4 cores

P28z Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 3.3 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops; 1 core used

P28A Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM compiled using 3.3 with -O3 -mcpu=970 -mtune=970 -fexpensive-optimizations -funroll-loops; 4 core used

P29 Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM only 1 core used

P29a Apple PowerMac Quad G5 2xPPC 970MP @ 2500 MHz MacOS X 32 kB Data +64kB Intr.
1024 kB@on DieMHz
2500 2.5GB@266MHzMHz
DDR-RAM 4 cores used

HP Workstations

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

H01 HP 715 HP PA 7100LC @ 100 MHz HPUX no
L1-Cache 256 kB@100?MHz
Boardcache 128@100?MHz
The Hp PA 7xxx-CPU

H02 HP 735 HP PA 7150 @ 125 MHz HPUX no
L1-Cache 256 kB@100?MHz
Boardcache 144@125?MHz
for details about the Cache see remarks at H01

H03 HP K 210 HP PA 7200 @ 120 MHz HPUX no
L1-Cache 256 kB@120?MHz
Boardcache 512@120?MHz
4-processor-machine with system bus joined by all four CPUs; for details about the Cache see remarks at H01

H04 HP C 160 HP PA 8000 @ 160 MHz HPUX ??
??@??MHz
128@??MHz

SGI Workstations

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

R01 SGI Indy MIPS R 4400 @ 150 MHz IRIX 6.2 ??
1024 kB@75?MHz
64@75MHz

R02 SGI Indy MIPS R 4400 @ 175 MHz IRIX 6.2 ??
1024 kB@87?MHz
64@87MHz

R03 SGI Indigo 2 MIPS R 4400 @ 200 MHz IRIX 5.3 ??
1024 kB@100+MHz
128@100+MHz

R04 SGI Indigo R10000 MIPS R 10000 @ 175 MHz IRIX 6.2 32kB+32kB
1024 kB@87+MHz
192@87+MHz

R05 SGI Origin 2000 MIPS R 10000 @ 195 MHz IRIX 6.4 32kB+32kB
4096 kB@97+MHz
1536@??MHz
via Crossbarswitch 8-processor-machine; CPUs are connected with the RAM-banks via a crossbarswitch rather than a common bus.

R06 SGI Octane MIPS R 10000 @ 175 MHz IRIX 6.4 32kB+32kB
1024 kB@97+MHz
192@??MHz
via Crossbarswitch CPU, IO-borad, graphics-board and RAM is connected via crossbarswitch rather than a common bus.

R07 SGI Octane IP30 MIPS R12000 @ 400 MHz IRIX 6.5 32kB+32kB
2 MB@??MHz
1GB@??MHz
gcc 3.3

R07a SGI Octane IP30 MIPS R12000 @ 400 MHz IRIX 6.5 32kB+32kB
2 MB@??MHz
1GB@??MHz
cc -O3 -mips4 -r10000 -64

R08 SGI Octane IP30 MIPS R10000 @ 250 MHz IRIX 6.5 32kB+32kB
2 MB@??MHz
256MB@??MHz
cc -O3 -mips4 -r10000 -64

SUN Workstations

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

S01 SUN Ultra2 UltarSparcII @ 300 MHz Solaris 2x16 kB?
2048 kB@??MHz
??@??MHz

S01a SUN Ultra2 UltarSparcII @ 300 MHz Solaris/gcc 2x16 kB?
2048 kB@??MHz
??@??MHz

S02 SUN Ultra30 UltraSparcII @ 250 MHz Solaris 2x16 kB
1024 kB@??MHz
??@??MHz

S02a SUN Ultra30 UltarSparcII @ 250 MHz Solaris/gcc 2x16 kB
1024 kB@??MHz
??@??MHz

Intel/AMD PC's

ID Computermodel CPU Operating- Cache RAM Remarks

system L1 L2

L01 Quad PentiumPro 4xPentiumPro @ 200 MHz Linux 2x16 kB
256 kB@200MHz
on CPU 256@66MHz
Quadruple CPU machine running Linux

L02 Dual Pentium II 2xPentium II @ 300 MHz Linux 2x16 kB
512 kB@150MHz
Backsidecache 256@66MHz
Dual CPU machine running Linux

L03 DualPentium II 2xPentium II @ 266 MHz Linux/gcc 2x16 kB
512 kB@133MHz
Backsidecache 256@66MHz
Dual CPU machine running Linux

L03a DualPentium II 2xPentium II @ 266 MHz Linux/egcs 2x16 kB
512 kB@133MHz
Backsidecache 256@66MHz
Dual CPU machine running Linux, but egcs 1.01 is used as compiler instead of gcc.

L04 Pentium II Pentium II @ 350 MHz Linux/egcs 2x16 kB
512 kB@175MHz
Backsidecache 128@100MHz
SuSe 6.2, Kernel 2.2.10, egcs 1.1.2

L05 DualPentium II Pentium II @ 450 MHz Linux/egcs 2x16 kB
512 kB@225MHz
Backsidecache 256@100MHz
SuSe 6.2, Kernel 2.2.10, egcs 1.1.2

L06 DualPentium III Pentium III @ 556 MHz Linux/egcs 2x16 kB
512 kB@278MHz
Backsidecache 256@112MHz
SDRAM SuSe 6.1; FSB overclocked (112 MHz)

L07 Transtec Celeron Celeron @ 400 MHz Linux 2x16 kB
128 kB@400MHz
Backsidecache 56@66MHz
SDRAM/UMA SuSe 6.2; cheap-PC priced 1050 DM (aka $480 w/o Tax); UMA-RAM-Design

L08 AMD Athlon K7 750 Athlon K7 @ 750 MHz Linux 2x64 kB
512 kB@300MHz
Backsidecache 256@100MHz
SDRAM RedHat 6.1

L09 LiAMD AMD K6/2 @ 333 MHz Linux 2x32 kB ' '
kB' 66 'Boardcache (Pipelienburst)' 64 66 'EDO+FPM,@jeMHz
32MB' 'SuSE Linux@6.1;MHz
gcc 2.2.14' 'SuSE

L10 Sony Vaio PCG-F190 mobile Pentium II @ 366 MHz Linux 2x16 kB
256 kB@366MHz
Backsidecache 128@66MHz
SDRAM Sony Notebook running SuSe 6.4

L11 AMD Athlon K7 700 Athlon K7 @ 700 MHz Linux 2x64 kB
512 kB@350MHz
Backsidecache 128@100MHz
SDRAM SuSE 6.4

L12 Dual Pentium III Pentium III @ 600 MHz Linux 2x16 kB
512 kB@300MHz
Backsidecache 512@100MHz
SDRAM SuSE 6.4

L13 Dual Pentium III Pentium III @ 700 MHz Linux 2x16 kB
256 kB@700MHz
Backsidecache 512@100MHz
SDRAM SuSE 6.4

L14 Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 512@100MHz
SDRAM SuSE 6.4

L14a Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 512@100MHz
SDRAM SuSE 6.4

L14b Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 1024@100MHz
SDRAM SuSE 7.3, compiled on SuSE 7.3 with gcc 2.95

L14c Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 1024@100MHz
SDRAM Debian-Linux, compiled on SuSE 7.3 with gcc 2.95

L15 Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 1024@100MHz
SDRAM SuSE 7.3 compiled with beta of intel compiler

L16 Dual Pentium III Pentium III @ 1000 MHz Linux 2x16 kB
256 kB@1000MHz
Backsidecache (on Die) 1024@100MHz
SDRAM SuSE 7.3 compiled with gcc2.95

L16a Dual Pentium III Pentium III @ 1000 MHz Linux 2x16 kB
256 kB@1000MHz
Backsidecache (on Die) 1024@100MHz
SDRAM SuSE 7.3 compiled with gcc2.95

L17 Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 1024@100MHz
SDRAM Debian-Linux compiled with gcc 3.2.2

L17a Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache 1024@100MHz
SDRAM SuSE 7.3, compiled on Debian-Linuxwith gcc 3.2.2

L18 Dual Pentium III Pentium III @ 800 MHz Linux 2x16 kB
256 kB@800MHz
Backsidecache (on Die) 1024@100MHz
SDRAM SuSE 9.2 compiled with gcc 3.3

L19 Pentium 4 HT Pentium 4 HT @ 3200 MHz Linux 16 kB + 16kμOps
1024 kB@3200MHz
Backsidecache (on Die) 2048@4x200MHz
DDR400 SuSE 9.2 compiled with gcc 3.3

L19a Pentium 4 HT Pentium 4 HT @ 3200 MHz Linux 16 kB + 16kμOps
1024 kB@3200MHz
Backsidecache (on Die) 2048@4x200MHz
DDR400 SuSE 9.2 compiled with gcc 3.3, two processes run in parralel on onw CPU (using Hyperthreading), result are the average of both runs

L20 Xeon 4 DP HT Xeon 4 DP HT @ 3066 MHz Linux 16 kB + 16kμOps
512 kB@3066MHz
Backsidecache (on Die) 512@4x133MHz
DDR266 SuSE 9.2 compiled with gcc 3.3

L20a Xeon 4 DP HT Xeon 4 DP HT @ 3066 MHz Linux 16 kB + 16kμOps
512 kB@3066MHz
Backsidecache (on Die) 512@4x133MHz
DDR266 SuSE 9.2 compiled with gcc 3.3, two processes run in parralel on onw CPU (using Hyperthreading), result are the average of both runs

L21 Notebook, Athon XP Athlon XP 2800+ @ 2083 MHz Linux 2x64kB
512 kB@2083MHz
Backsidecache (on Die) 512@2x166MHz
DDR333 SuSE 9.2 compiled with gcc 3.3

L22 Athlon XP 2400 Athlon XP 2400 @ 2000 MHz FreeBSD 2x64kB
256 kB@2000MHz
Backsidecache (on Die) 512@2x133MHz
DDR266 FreeBSD 5.3-STABLE, compiled with gcc version 3.4.2

L22a Athlon XP 2400 Athlon XP 2400 @ 2000 MHz FreeBSD 2x64kB
256 kB@2000MHz
Backsidecache (on Die) 512@2x133MHz
DDR266 FreeBSD 5.3-STABLE, compiled with gcc version 3.3

L23 Dual Opteron Server Opteron 244 @ 1800 MHz Linux 2x64kB
1024 kB@1800MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? running on Knoppix-Linux with binaries compiled and optimised for K8 on a SuSE 9.2 box.

L24 Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 9.2(x86-64), complied with -m32 -march=k8

L24a Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 9.2(x86-64), complied with -m32 -march=k8 -msse2

L24b Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 9.2(x86-64), complied with -m64 -march=k8

L24c Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 9.2(x86-64), complied with gcc 4.02 and -m64 -march=k8 -msse2

L24d Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m32 -march=k8

L24e Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m32 -march=k8 -msse2

L24f Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 10.0(x86-64), complied with gcc 4.02 -m64 -march=k8

L24g Athlon 64 Athlon 64 @ 2400 MHz Linux 2x64kB
1024 kB@2400MHz
Backsidecache (on Die) ???@800MHz HTMHz
???? Clawhammer-Coe, running on SuSE 10.0(x86-64), complied with -m64 -march=k8 -msse2

L25 Dual Pentium III Pentium III @ 550 MHz Linux/Debian 2x16kB
512 kB@275MHz
Backsidecache ??@100 MHzMHz
gcc 3.3.5

L25a Dual Pentium III Pentium III @ 550 MHz Linux/Debian 2x16kB
512 kB@275MHz
Backsidecache ??@100 MHzMHz
gcc 3.3.5 with mit -mcpu=pentium3 -O3 -fomit-frame-pointer -malign-jumps=2 -malign-functions=2 -malign-loops=2

L25b Dual Pentium III Pentium III @ 550 MHz Linux/Debian 2x16kB
512 kB@275MHz
Backsidecache ??@100 MHzMHz
gcc 3.3.5 with mit -mcpu=pentium3 -O3 -fomit-frame-pointer -ffast-math -malign-jumps=2 -malign-functions=2 -malign-loops=2

L26 Dual Opteron 242 Opteron 242 @ 1600 MHz Linux 2x64kB ' '
kB 1600@Backsidecache (on Die)MHz
??? 800MHz HT@????MHz
compiliert mit -mcpu=k8 -m64 -O3 -fomit-frame-pointer -ffast-math -funroll-loops -malign-jumps=2 -malign-functions=2 -malign-loops=2 gcc 3.3.5 with mit -mcpu=pentium3 -O3 -fomit-frame-pointer -ffast-math -malign-jumps=2 -malign-functions=2 -malign-loops=2

L27 Athlon 64X2 Athlon 64X2 @ 2200 MHz Linux 2x64kB
2x1024 kB@2200MHz
Backsidecache (on Die) ???@1000MHz HTMHz
???? running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m32 -march=k8

L27a Athlon 64X2 Athlon 64X2 @ 2200 MHz Linux 2x64kB
2x1024 kB@2200MHz
Backsidecache (on Die) ???@1000MHz HTMHz
???? running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m32 -march=k8 -msse3

L27b Athlon 64X2 Athlon 64X2 @ 2200 MHz Linux 2x64kB
2x1024 kB@2200MHz
Backsidecache (on Die) ???@1000MHz HTMHz
???? running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m64 -march=k8

L27c Athlon 64X2 Athlon 64X2 @ 2200 MHz Linux 2x64kB
2x1024 kB@2200MHz
Backsidecache (on Die) ???@1000MHz HTMHz
???? running on SuSE 10.0(x86-64), complied with gcc 4.02 and -m64 -march=k8 -msse3

L28 Apple iMac CoreDuo Intel CoreDuo @ 2000 MHz MacOS X 2x32kB
2024 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta Universal binary, both cores used

L28a Apple iMac CoreDuo Intel CoreDuo @ 2000 MHz MacOS X 2x32kB
2024 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta Universal binary, only 1 core used

L29 Apple iMac CoreDuo Intel CoreDuo @ 2000 MHz MacOS X 2x32kB
2024 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta Universal binary, both cores used

L29a Apple iMac CoreDuo Intel CoreDuo @ 2000 MHz MacOS X 2x32kB
2024 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta Universal binary, only 1 core used

L30 Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, 4 cores used, freshly compiled for x86-64

L30a Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, only 1 core used, freshly compiled for x86-64

L30b Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, 16 processes on 4 cores, freshly compiled for x86-64

L30c Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, 4 cores used, freshly compiled for x86-32

L30d Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, only 1 core used, freshly compiled for x86-32

L30e Apple MacPro 2x Intel CoreDuo 2 @ 2000 MHz MacOS X 2x32kB
4096 kB@2000MHz
unified cache for two cores ???@667MHzMHz
???? RGBbench 1.6beta2, 16 processes on 4 cores, freshly compiled for x86-32

L31 Athlon64X2 4600+ Athlon64X2 4600+ @ 2400 MHz Linux 2x64kB
2x512 kB@2400MHz
Backsidecache (on Die) 2 GB@1000MHz HTMHz
DDR2 800 Benchmark run on one core

L31a Athlon64X2 4600+ Athlon64X2 4600+ @ 2400 MHz Linux 2x64kB
2x512 kB@2400MHz
Backsidecache (on Die) 2 GB@1000MHz HTMHz
DDR2 800 Benchmark run on both core

L32 AthlonXP AthlonXP @ 1725 MHz Linux 2x64kB
256 kB@1725MHz
Backsidecache (on Die) 512@2x133MHz
DDR266 complield with gcc-4.1.1 and -O3 -mcpu=athlon-xp -mtune=athlon-xp -fexpensive-optimizations -funroll-loops -msse -mmmx -m3dnow -m32

L33 Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core, compilied with gcc 4.1.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33a Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on both cores, compilied with gcc 4.1.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33b Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 4 processes on two cores, compilied with gcc 4.1.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33c Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33d Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on both cores, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33e Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 4 processes on two cores, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for x86-64

L33f Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for IA32

L33g Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on both cores, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for IA32

L33h Apple MacBook Pro Intel Core2Duo @ 2167 MHz Linux 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 4 processes on two cores, compilied with gcc 4.0.2 and -fexpensive-optimizations -funroll-loops for IA32

L33i Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for x86-64

L33j Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on both core, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for x86-64

L33k Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 4 processes on two cores, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for x86-64

L33l Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for IA32

L33m Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on both core, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for IA32

L33n Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 4 processes on two cores, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for IA32

L33o Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on one core using Rosetta, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for x86-64

L33p Apple MacBook Pro Intel Core2Duo @ 2167 MHz MacOS X 2x32kB
4096 kB@2167MHz
unified cache for two cores 1GB@667MHzMHz
DDR2 667 Benchmark run on two cores using Rosetta, compilied with gcc 4.0.1 and -fexpensive-optimizations -funroll-loops for x86-64

L34 Sunfire X2200 M2 2xAMD Opteron 2218 (DualCore) @ 2600 MHz Linux 2x64kB
2x1024 kB@2600MHz
Backsidecache (on Die) 8GB@667MHzMHz
DDR2667

L35 Dell PowerEdge 2xIntel Core2Duo 5160 @ 3000 MHz Linux 232kB
4096 kB@3000MHz
unified cache for two cores 8 GB@667MhzMHz
DDR667, quad channel running SLES 10 on one core, compiled with -fexpensive-optimizations -funroll-loops -m64

L35a Dell PowerEdge 2xIntel Core2Duo 5160 @ 3000 MHz Linux 232kB
4096 kB@3000MHz
unified cache for two cores 8 GB@667MhzMHz
DDR667, quad channel running SLES 10 on two cores, compiled with -fexpensive-optimizations -funroll-loops -m64

L35b Dell PowerEdge 2xIntel Core2Duo 5160 @ 3000 MHz Linux 232kB
4096 kB@3000MHz
unified cache for two cores 8 GB@667MhzMHz
DDR667, quad channel running SLES 10 on four cores, compiled with -fexpensive-optimizations -funroll-loops -m64

L35c Dell PowerEdge 2xIntel Core2Duo 5160 @ 3000 MHz Linux 232kB
4096 kB@3000MHz
unified cache for two cores 8 GB@667MhzMHz
DDR667, quad channel running SLES 10 with 16 processes on four cores, compiled with -fexpensive-optimizations -funroll-loops -m64

L35d Dell PowerEdge 2xIntel Core2Duo 5160 @ 3000 MHz Linux 232kB
4096 kB@3000MHz
unified cache for two cores 8 GB@667MhzMHz
DDR667, quad channel running SLES 10 with 16 processes on four cores, compiled with -fexpensive-optimizations -funroll-loops -m64

Last Update : 05/21/2007

Some remarks on the columns:

The ID is meant to associate the rows in all the tables around these benchmarks pages.
There are at least four different ways of running L2 cache. Unfortunately I have no information about the kind of cache used in the Workstations.
1. Boardc ache: This is the traditional way of having L2 cache. It's located on the motherboard and runs at the clock speed of the system bus.
2. Inline cache: CPU clock speeds became faster and faster while the system bus clock speed stayed almost the same. Hence the cache became more and more ineffective. One way to get around this problem is to put the cache onto an accelerated part of the system bus right before the CPU. This part of the system bus runs at some multiple of the normal bus speed, typically twice as fast. Examples are the PowerMac 8600/9600 and the CacheDoubler for the Mac-clones by UMAX (APUS-Series, that is the SuperMac J and C-Series in the USA, afaik).
3. Backside cache: The most common form of L2 cache with modern Intel and PowerPC CPUs. These CPUs have an extra bus especially for the cache. The clock speed of this bus is always fast. It might be in a fixed ratio to the CPU clock speed (e.g. Pentium II: cache speed =1/2 CPU speed) or variable (e.g. PowerPC 750: cache speed:CPU speed ratios: 1:3, 1:2.5, 1:2, 1:1.5 and 1:1). Since the cache controller is integrated into the CPU, this is kind of cache is the fastest mentioned so far.
4. On-chip cache: In principle a variant of the backside cache. However, the cache sits on the CPU itself and is run at the clock speed of the CPU (e.g. Pentium Pro). Although this is very fast, it has the disadvantage of being very expensive.
Some processors (e.g. DEC Alpha 21164) have tiny integrated L2 caches. However, in this case, this is more a split L1 cache. Thus you find huge L3 caches on the motherboards with DEC-Alpha CPUs (up to 8MB).
If I have the information, I included the kind of RAM used. FPM stands here for Fast Page Mode (this is the traditional kind of RAM), EDO for Enhanced Data Out. and SDRAM for synchronous RAM.
On cheap computers (like Transtec Celeron) one finds UMA-main-memory design (UMA=united memory access). In this case the video-RAM is located within the RAM. This make the computer cheaper, but slow due to interfering accesses from the graphicsprocessor ans the CPU.
Some SGI-machine have another variant of accessing the RAM. On multiprocessor machines, you normally have a fight for the system bus, slowing down performance. With SGI's crossbar switch, the CPUs are connected directly to the RAM boards. All CPUs can be connected to the RAM at the same time as long as they access different parts of the RAM. It's like having e.g. 8 buses connecting each of the CPUs to the requested part of RAM.