Measure CPU Speed for your distros

[Anil]

It's a cute way to get CPU speed information. Super-pi is a free program (about 80 K Zip file)to calculate Pi to Millions of digits1) download Super_Pi from ftp://pi.super-computing.org/Linux/super_pi.tar.gz orhttp://superpi.ilbello.com/ (windows and Linux)2)Type in Terminal as user ./super_pi 20which runs the program to measure the time it takes to calculate pi to 1 million digits 3) Pi.dat file is created in superpi directory. which contains the million digits of PiHere are my resultsfor Mandriva 2008:pi to 1 million digitsDesktop Intel P4, 1 GB RAM, 2Ghz ---------------66 SecAcer Laptop Pentium duel Core 32bit 1GB ram ---30 SecI am curious to know whether different distros give different results ? Many ATL members are having multiple distros installed.(some have 2 dozen distros per hard disk !)

[cybormoron]

here's mine, mandriva 2008, p4-3Ghz, 1gb ram.

[steve@localhost Download]$ ./super_pi 20 Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with version 19981204 of f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minline=size:1000 -Mnoframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used for the compilation. ------ Started super_pi run : Fri Dec 14 10:59:19 EST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.494 Sec. I= 1 L= 0 Time= 1.570 Sec. I= 2 L= 0 Time= 1.809 Sec. I= 3 L= 1 Time= 1.813 Sec. I= 4 L= 2 Time= 1.810 Sec. I= 5 L= 5 Time= 1.805 Sec. I= 6 L= 10 Time= 1.805 Sec. I= 7 L= 21 Time= 1.798 Sec. I= 8 L= 43 Time= 1.801 Sec. I= 9 L= 87 Time= 1.802 Sec. I=10 L= 174 Time= 1.800 Sec. I=11 L= 349 Time= 1.806 Sec. I=12 L= 698 Time= 1.796 Sec. I=13 L= 1396 Time= 1.798 Sec. I=14 L= 2794 Time= 1.798 Sec. I=15 L= 5588 Time= 1.787 Sec. I=16 L= 11176 Time= 1.785 Sec. I=17 L= 22353 Time= 1.799 Sec. I=18 L= 44707 Time= 1.711 Sec. I=19 L= 89415 Time= 1.617 Sec. End of main loop End of calculation. Time= 35.628 Sec. End of data output. Time= 0.192 Sec. Total calculation(I/O) time= 35.820( 1.152) Sec. ------ Ended super_pi run : Fri Dec 14 10:59:55 EST 2007[steve@localhost Download]$

[securitybreach]

Here is mine on Arch Linux with CoreDuo T2250@1.73ghz with 2.066gb of ram:

[comhack@DarkStar ~]$ ./super_pi 20 Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with versi f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minlframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used compilation. ------ Started super_pi run : Fri Dec 14 11:22:30 EST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time=	   0.370 Sec. I= 1 L=	   0		Time=	   1.203 Sec. I= 2 L=	   0		Time=	   1.397 Sec. I= 3 L=	   1		Time=	   1.390 Sec. I= 4 L=	   2		Time=	   1.387 Sec. I= 5 L=	   5		Time=	   1.417 Sec. I= 6 L=	  10		Time=	   1.400 Sec. I= 7 L=	  21		Time=	   1.393 Sec. I= 8 L=	  43		Time=	   1.397 Sec. I= 9 L=	  87		Time=	   1.393 Sec. I=10 L=	 174		Time=	   1.400 Sec. I=11 L=	 349		Time=	   1.430 Sec. I=12 L=	 698		Time=	   1.393 Sec. I=13 L=	1396		Time=	   1.413 Sec. I=14 L=	2794		Time=	   1.397 Sec. I=15 L=	5588		Time=	   1.387 Sec. I=16 L=   11176		Time=	   1.377 Sec. I=17 L=   22353		Time=	   1.400 Sec. I=18 L=   44707		Time=	   1.317 Sec. I=19 L=   89415		Time=	   1.237 Sec. End of main loop End of calculation.	Time=	  27.585 Sec. End of data output.	Time=	   0.153 Sec. Total calculation(I/O) time=	  27.738(1.050) Sec. ------ Ended super_pi run : Fri Dec 14 11:22:59 EST 2007

27.738 seconds not too badThanks

[Anil]

@ Steve -- 3 Gzh makes the difference ! 35 sec for P4 is one of the Topmost result on Net@ Securitybreach -- Is it a Laptop or Desktop ? Also if you multiboot, have you tried it with Windows ?Thanks

Edited December 16, 2007 by Anil

[Bruno]

Main system: 2188MHz, 2GB RAM:

$ uname -aLinux saturn 2.6.22.12-desktop-1mdv #1 SMP Tue Nov 20 08:38:27 EST 2007 i686 AMD Athlon(tm) XP 2700+ GNU/Linux$ free			 total	   used	   free	 shared	buffers	 cachedMem:	   2076160	 407148	1669012		  0	  19132	 204736-/+ buffers/cache:	 183280	1892880Swap:	  4618648		  0	4618648 ------ Started super_pi run : Fri Dec 14 19:09:11 CET 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time=	   0.615 Sec. I= 1 L=	   0		Time=	   1.939 Sec. I= 2 L=	   0		Time=	   2.219 Sec. I= 3 L=	   1		Time=	   2.206 Sec. I= 4 L=	   2		Time=	   2.225 Sec. I= 5 L=	   5		Time=	   2.209 Sec. I= 6 L=	  10		Time=	   2.205 Sec. I= 7 L=	  21		Time=	   2.226 Sec. I= 8 L=	  43		Time=	   2.201 Sec. I= 9 L=	  87		Time=	   2.221 Sec. I=10 L=	 174		Time=	   2.198 Sec. I=11 L=	 349		Time=	   2.223 Sec. I=12 L=	 698		Time=	   2.200 Sec. I=13 L=	1396		Time=	   2.223 Sec. I=14 L=	2794		Time=	   2.212 Sec. I=15 L=	5588		Time=	   2.210 Sec. I=16 L=   11176		Time=	   2.194 Sec. I=17 L=   22353		Time=	   2.165 Sec. I=18 L=   44707		Time=	   2.128 Sec. I=19 L=   89415		Time=	   1.942 Sec. End of main loop End of calculation.	Time=	  43.734 Sec. End of data output.	Time=	   0.183 Sec. Total calculation(I/O) time=	  43.917(	   1.453) Sec. ------ Ended super_pi run : Fri Dec 14 19:09:56 CET 2007

BrunoPS: Same result in runlevel 3

[mlangdn]

33.135 secondsSlackware 12 with Boinc client runningAMD 64x2 4400+2gb ram

[securitybreach]

@ Steve -- 3 Gzh makes the difference ! 35 sec for P4 is one of the Topmost result on Net@ Securitybreach -- Is it a Laptop or Desktop ? Also if you multiboot, have you tied it with Windows ?Thanks

Its a Toshiba Satellite S105-4114 Laptop. I do not use Windows at all.THanks

[Bruno]

Test System: AMD Athlon XP 1700+ = 1100MHz, 1GB RAM:Mandriva: 94.1 secSlackware: 93.9 secDebian: 93,8PCLos: 91.9 secVectorLinux: 94.2 secKubuntu: 90,7Fedora 94,8 secSuSE 90,4B) Bruno

[V.T. Eric Layton]

On my AMD Athlon 64 3800+ running Slackware 12:

------ Started super_pi run : Fri Dec 14 14:55:23 EST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.424 Sec. I= 1 L= 0 Time= 1.324 Sec. I= 2 L= 0 Time= 1.516 Sec. I= 3 L= 1 Time= 1.508 Sec. I= 4 L= 2 Time= 1.516 Sec. I= 5 L= 5 Time= 1.520 Sec. I= 6 L= 10 Time= 1.516 Sec. I= 7 L= 21 Time= 1.512 Sec. I= 8 L= 43 Time= 1.520 Sec. I= 9 L= 87 Time= 1.508 Sec. I=10 L= 174 Time= 1.516 Sec. I=11 L= 349 Time= 1.520 Sec. I=12 L= 698 Time= 1.432 Sec. I=13 L= 1396 Time= 1.416 Sec. I=14 L= 2794 Time= 1.504 Sec. I=15 L= 5588 Time= 1.504 Sec. I=16 L= 11176 Time= 1.508 Sec. I=17 L= 22353 Time= 1.488 Sec. I=18 L= 44707 Time= 1.444 Sec. I=19 L= 89415 Time= 1.340 Sec. End of main loop End of calculation. Time= 29.730 Sec. End of data output. Time= 0.160 Sec. Total calculation(I/O) time= 29.890( 0.896) Sec. ------ Ended super_pi run : Fri Dec 14 14:55:54 EST 2007

[sunrat]

Here's mine. Core2Duo E6400 (OC to 2.44GHz) 2Gb RAM Debian Testing/Sid Sunrat Special brand roger@brain:~/bin/super_pi$ ./super_pi 20 Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with version 19981204 of f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minline=size:1000 -Mnoframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used for the compilation. ------ Started super_pi run : Sat Dec 15 07:05:31 EST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.260 Sec. I= 1 L= 0 Time= 0.852 Sec. I= 2 L= 0 Time= 1.012 Sec. I= 3 L= 1 Time= 1.032 Sec. I= 4 L= 2 Time= 1.004 Sec. I= 5 L= 5 Time= 1.016 Sec. I= 6 L= 10 Time= 0.992 Sec. I= 7 L= 21 Time= 0.984 Sec. I= 8 L= 43 Time= 0.988 Sec. I= 9 L= 87 Time= 1.012 Sec. I=10 L= 174 Time= 1.028 Sec. I=11 L= 349 Time= 0.992 Sec. I=12 L= 698 Time= 1.012 Sec. I=13 L= 1396 Time= 0.984 Sec. I=14 L= 2794 Time= 0.960 Sec. I=15 L= 5588 Time= 0.976 Sec. I=16 L= 11176 Time= 0.976 Sec. I=17 L= 22353 Time= 0.956 Sec. I=18 L= 44707 Time= 0.936 Sec. I=19 L= 89415 Time= 0.884 Sec. End of main loop End of calculation. Time= 19.613 Sec. End of data output. Time= 0.104 Sec. Total calculation(I/O) time= 19.717( 0.612) Sec. ------ Ended super_pi run : Sat Dec 15 07:05:51 EST 2007roger@brain:~/bin/super_pi$

[V.T. Eric Layton]

Well, SecurityBreach (27.738) and I (29.890) were both looking pretty good till Sunrat (19.717) posted.

[réjean]

sunrat is definitely flying.Mine on a Intel Core2 duo 3 Ghz with 2 Gb of DDR2;

End of main loop End of calculation.	Time=	  30.838 Sec. End of data output.	Time=	   0.178 Sec. Total calculation(I/O) time=	  31.016(	   1.070) Sec. ------ Ended super_pi run : Fri Dec 14 17:42:51 AST 2007[rejean@localhost super_pi]$

Not bad, even if a bit slower than Eric and securitybreach.Oops! I forgot to say that I was in Mandriva. Now from Linux Mint; a tad faster:

 End of main loop End of calculation.	Time=	  30.254 Sec. End of data output.	Time=	   0.180 Sec. Total calculation(I/O) time=	  30.434(	   0.984) Sec.

much slower in Fedora Core 6

 End of main loop End of calculation.	Time=	  39.478 Sec. End of data output.	Time=	   0.273 Sec. Total calculation(I/O) time=	  39.751(	   1.610) Sec.

so far PCLinuxOS 2007 is the winner;

End of main loop End of calculation.	Time=	  29.618 Sec. End of data output.	Time=	   0.180 Sec. Total calculation(I/O) time=	  29.798(	   0.828) Sec.

on my machine that is.and in Slackware 12;

End of main loop End of calculation.	Time=	  29.570 Sec. End of data output.	Time=	   0.180 Sec. Total calculation(I/O) time=	  29.750(	   0.908) Sec.

Edited December 16, 2007 by réjean

[BarryB]

LOL...well it's a Gateway!Started super_pi run : Fri Dec 14 17:50:21 EST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.331 Sec. I= 1 L= 0 Time= 1.020 Sec. I= 2 L= 0 Time= 1.169 Sec. I= 3 L= 1 Time= 1.167 Sec. I= 4 L= 2 Time= 1.168 Sec. I= 5 L= 5 Time= 1.166 Sec. I= 6 L= 10 Time= 1.165 Sec. I= 7 L= 21 Time= 1.174 Sec. I= 8 L= 43 Time= 1.170 Sec. I= 9 L= 87 Time= 1.174 Sec. I=10 L= 174 Time= 1.174 Sec. I=11 L= 349 Time= 1.167 Sec. I=12 L= 698 Time= 1.163 Sec. I=13 L= 1396 Time= 1.165 Sec. I=14 L= 2794 Time= 1.163 Sec. I=15 L= 5588 Time= 1.161 Sec. I=16 L= 11176 Time= 1.156 Sec. I=17 L= 22353 Time= 1.139 Sec. I=18 L= 44707 Time= 1.102 Sec. I=19 L= 89415 Time= 1.042 Sec. End of main loop End of calculation. Time= 23.035 Sec. End of data output. Time= 0.133 Sec. Total calculation(I/O) time= 23.168( 0.681) Sec. ------ Ended super_pi run : Fri Dec 14 17:50:45 EST 2007[barryb@Core2BoxHave to try a few more systems

[Anil]

Well, SecurityBreach (27.738) and I (29.890) were both looking pretty good till Sunrat (19.717) posted.

Yes It's quite impressive!But I think your results are without Overclocking Eric ?

[Anil]

Test System: AMD Athlon XP 1700+ = 1100MHz, 1GB RAM:Mandriva: 94.1 secSlackware: 93.9 secDebian: 93,8PCLos: 91.9 secVectorLinux: 94.2 secKubuntu: 90,7Fedora 94,8 secSuSE 90,4B) Bruno

Thanks Bruno ! I knew PCLos gives better timing than Mandriva but Suse is a surprise for me.

[réjean]

Funny that my OpenSUSE is slower than Mandy, Linux Mint and PCLos but faster than Fedora;

 End of main loop End of calculation.	Time=	  33.666 Sec. End of data output.	Time=	   0.200 Sec. Total calculation(I/O) time=	  33.866(	   1.344) Sec.

still on the same machine.Even slower in Debian /etch

 End of calculation.	Time=	  37.682 Sec. End of data output.	Time=	   0.184 Sec. Total calculation(I/O) time=	  37.866(	   1.068) Sec.

Edited December 16, 2007 by réjean

[Anil]

Thanks Rejean. Puppy Live CD, which runs in RAM, gives better timing on my machine than Mandriva2008.

Edited December 15, 2007 by Anil

[sunrat]

OK, finally got round to testing my other installed distros. 64Studio is 64bit, others are 32bit. I expected the 2 studio distros to be fast, but the Sidux kernel, also the latest, was a pleasant surprise at just over a full second faster than lenny!Debian lenny/sidroger@brain:~$ uname -aLinux brain 2.6.22-3-686 #1 SMP Mon Nov 12 08:32:57 UTC 2007 i686 GNU/LinuxTotal calculation(I/O) time= 19.717( 0.612) Sec.Studio64 2.0Linux brain 2.6.21-1-multimedia-amd64 #1 SMP PREEMPT RT Fri Jun 22 00:00:33 UTC 2007 x86_64 GNU/Linux Total calculation(I/O) time= 19.004( 0.761) Sec.UbuntuStudio 7.10Linux brain 2.6.22-14-rt #1 SMP PREEMPT RT Mon Oct 15 01:05:51 GMT 2007 i686 GNU/LinuxTotal calculation(I/O) time= 18.782( 0.771) Sec.Sidux "eros"Linux brain-sidux 2.6.23.9-rc1-slh-smp-1 #1 SMP PREEMPT Tue Nov 20 19:33:34 UTC 2007 i686 GNU/LinuxTotal calculation(I/O) time= 18.693( 0.568) Sec.

[V.T. Eric Layton]

Yes It's quite impressive!But I think your results are without Overclocking Eric ?

That's correct, Anil. Clock speed on my system is 2411Mhz (stock speed, no tweaks).

[Anil]

OK, finally got round to testing my other installed distros. 64Studio is 64bit, others are 32bit. I expected the 2 studio distros to be fast, but the Sidux kernel, also the latest, was a pleasant surprise at just over a full second faster than lenny!Sidux "eros"Linux brain-sidux 2.6.23.9-rc1-slh-smp-1 #1 SMP PREEMPT Tue Nov 20 19:33:34 UTC 2007 i686 GNU/LinuxTotal calculation(I/O) time= 18.693( 0.568) Sec.

Great Timings Sunrat

[Anil]

Results so far :P.S: Sorry, forgot to add Mlangdn's Result: AMD 64x2 4400+/2gb ram/33.135 seconds/Slackware 12

Edited December 16, 2007 by Anil

[securitybreach]

Very nice chart. Helps to compare the results better. Quick question--Do you have all the other distros you listed the Time(sec) on? If so you have quite a list of distros you tried it on. Wow I cannot believe the 90+ second ones.THanks

Edited December 16, 2007 by securitybreach

[Anil]

Very nice chart. Helps to compare the results better. Quick question--Do you have all the other distros you listed the Time(sec) on? If so you have quite a list of distros you tried it on. Wow I cannot believe the 90+ second ones.THanks

LOL..No I just have Mandriva ! Poor me As the chart tries to show-- Bruno tested for 8 distros, Sunrat 4 and Rejean 5. The difference in speed will be noticeable if we run ./super_pi 25 . It will do the calculation for 32 million digits of Pi

[Frank Golden]

Intel Core2Duo 2000MHz, 2GB DDR ram, PCLinuxOS 2007

[frank@localhost ~]$ ./super_pi 20 Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with version 19981204 of f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minline=size:1000 -Mnoframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used for the compilation. ------ Started super_pi run : Sun Dec 16 08:43:21 PST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.292 Sec. I= 1 L= 0 Time= 0.848 Sec. I= 2 L= 0 Time= 0.968 Sec. I= 3 L= 1 Time= 0.968 Sec. I= 4 L= 2 Time= 0.972 Sec. I= 5 L= 5 Time= 0.964 Sec. I= 6 L= 10 Time= 0.964 Sec. I= 7 L= 21 Time= 0.968 Sec. I= 8 L= 43 Time= 0.968 Sec. I= 9 L= 87 Time= 0.968 Sec. I=10 L= 174 Time= 0.968 Sec. I=11 L= 349 Time= 0.964 Sec. I=12 L= 698 Time= 0.964 Sec. I=13 L= 1396 Time= 0.968 Sec. I=14 L= 2794 Time= 0.964 Sec. I=15 L= 5588 Time= 0.960 Sec. I=16 L= 11176 Time= 0.952 Sec. I=17 L= 22353 Time= 0.936 Sec. I=18 L= 44707 Time= 0.908 Sec. I=19 L= 89415 Time= 0.844 Sec. End of main loop End of calculation. Time= 19.053 Sec. End of data output. Time= 0.124 Sec. Total calculation(I/O) time= 19.177( 0.744) Sec. ------ Ended super_pi run : Sun Dec 16 08:43:40 PST 2007[frank@localhost ~]$

Same processor etc., Gutsy Gibbon Ubuntu (7.10)

Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with version 19981204 of f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minline=size:1000 -Mnoframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used for the compilation. ------ Started super_pi run : Sun Dec 16 08:33:29 PST 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.308 Sec. I= 1 L= 0 Time= 0.896 Sec. I= 2 L= 0 Time= 1.064 Sec. I= 3 L= 1 Time= 1.096 Sec. I= 4 L= 2 Time= 1.024 Sec. I= 5 L= 5 Time= 1.016 Sec. I= 6 L= 10 Time= 0.984 Sec. I= 7 L= 21 Time= 1.064 Sec. I= 8 L= 43 Time= 1.092 Sec. I= 9 L= 87 Time= 1.072 Sec. I=10 L= 174 Time= 1.060 Sec. I=11 L= 349 Time= 1.016 Sec. I=12 L= 698 Time= 1.020 Sec. I=13 L= 1396 Time= 1.020 Sec. I=14 L= 2794 Time= 1.012 Sec. I=15 L= 5588 Time= 1.020 Sec. I=16 L= 11176 Time= 1.004 Sec. I=17 L= 22353 Time= 0.948 Sec. I=18 L= 44707 Time= 0.960 Sec. I=19 L= 89415 Time= 0.852 Sec. End of main loop End of calculation. Time= 20.325 Sec. End of data output. Time= 0.132 Sec. Total calculation(I/O) time= 20.457( 0.792) Sec. ------ Ended super_pi run : Sun Dec 16 08:33:50 PST 2007frank@frank-laptop:~$

This may or may not be of interest but I ran Super Pi (win version) on XP pro,same proc etc. Got 26 sec!! I always knew Linux was faster. Edited December 16, 2007 by Frank Golden

[Anil]

This may or may not be of interest but I ran Super Pi (win version) on XP pro,same proc etc. Got 26 sec!! I always knew Linux was faster. hysterical.gif hysterical.gif

Thanks Frank for posting XP result. On some other forum, running the same thread , certain results were banned because of having Unfair Advantage(Read Linux OS) Edited December 16, 2007 by Anil

[Bruno]

Those Core2Duo processors really do make a difference don't they ? Really impressive !I am a bit worried about Réjeans results . . . . it looks like only one core of the C2D is giving the horsepower shown . . . . . ( ?? )

Wow I cannot believe the 90+ second ones.

The 90+ second ones are on my test-system that has only 1 GHz CPU and 1 GB RAM . . . it is the system with the most distros installed so I could compare the difference between distros in a fair way.My faster 2.2 GHz CPU and 2 GB RAM system only has Mandriva installed ( and the 23 other distros running in VirtualBox ) so I was not able to compare the difference between distros on that one.

because of having Unfair Advantage

Goes to show how Windows makes use of ( wastes ) your CPU cycles Bruno

[V.T. Eric Layton]

Nice chart, Anil! :)By the way, my processor (AMD Athlon 64 3800+) is a single core processor. Also, I'm running 2Gig of RAM on my system.

[striker]

Acer 6465WLMi laptop - Core 2 Duo T7200 2 GHz - 2gig RAM - Mandriva 2008 64 bit./super_pi 20 Version 2.0 of the super_pi for Linux OS Fortran source program was translated into C program with version 19981204 of f2c, then generated C source program was optimized manually. pgcc 3.2-3 with compile option of "-fast -tp px -Mbuiltin -Minline=size:1000 -Mnoframe -Mnobounds -Mcache_align -Mdalign -Mnoreentrant" was used for the compilation. ------ Started super_pi run : Sun Dec 16 21:24:10 CET 2007 Start of PI calculation up to 1048576 decimal digits End of initialization. Time= 0.323 Sec. I= 1 L= 0 Time= 0.857 Sec. I= 2 L= 0 Time= 0.980 Sec. I= 3 L= 1 Time= 0.973 Sec. I= 4 L= 2 Time= 0.977 Sec. I= 5 L= 5 Time= 0.973 Sec. I= 6 L= 10 Time= 0.977 Sec. I= 7 L= 21 Time= 0.970 Sec. I= 8 L= 43 Time= 0.983 Sec. I= 9 L= 87 Time= 0.980 Sec. I=10 L= 174 Time= 0.970 Sec. I=11 L= 349 Time= 0.980 Sec. I=12 L= 698 Time= 0.980 Sec. I=13 L= 1396 Time= 0.977 Sec. I=14 L= 2794 Time= 0.963 Sec. I=15 L= 5588 Time= 0.963 Sec. I=16 L= 11176 Time= 0.960 Sec. I=17 L= 22353 Time= 0.937 Sec. I=18 L= 44707 Time= 0.903 Sec. I=19 L= 89415 Time= 0.843 Sec. End of main loop End of calculation. Time= 19.209 Sec. End of data output. Time= 0.130 Sec. Total calculation(I/O) time= 19.339( 0.697) Sec. ------ Ended super_pi run : Sun Dec 16 21:24:30 CET 2007

[sunrat]

Those Core2Duo processors really do make a difference don't they ? Really impressive !I am a bit worried about Réjeans results . . . . it looks like only one core of the C2D is giving the horsepower shown . . . . . ( ?? )

I ran gkrellm at the same time as super_pi, and the CPU use was 100% for one core only with the other almost idle. So super_pi runs only a single thread anyway.Maybe Rejean's comp has frequency scaling active. I would have expected to see times similar to Frank's. Instead it seems more like a 1.5GHz CPU. Exactly which model CPU is it, Rejean?I also ran superpi on WinXP and got 24seconds. Thats about 30% slower! Not sure how similar the Win version is to Linux version though, could be a different program.Might try it in VBox. Will post results soon.

[sunrat]

I'm not going to speculate, but what do you make of these figures from Sidux in VirtualBox?

Total calculation(I/O) time=	  12.020(	   0.605) Sec. ------ Ended super_pi run : Tue Dec 18 01:15:04 EST 2007roger@siduxbrain:~/super_pi$