Fedora Super Pi Results

Elektrey · #1 5th August 2008, 02:35 PM

Post your results for super pi to 1M digits
Post yout total time and your processor and speed.

time = 33.102
Processor: Intel Dual Core Pentium 1.46Ghz

Once I get 2GB it will go faster

LDC · #2 5th August 2008, 05:08 PM

how do I do that?

sideways · #3 5th August 2008, 09:27 PM

Quote:

Originally Posted by LDC

how do I do that?

You download Super PI and run it under Wine.

27 secs for 1 million digits, Core2 Duo T7250 @2Ghz, 2GB ram

Beat that

Elektrey · #4 5th August 2008, 10:51 PM

sideways: There IS a linux one, wine would take up a bunch of ram so it would be quicker in linux.

RupertPupkin · #5 5th August 2008, 11:48 PM

You can get the Linux version here: ftp://pi.super-computing.org/Linux/

I couldn't figure out how to get exactly 1 million digits; the super_pi program seems to take a parameter N and then calculates pi to 2^N decimal places. So I used N=20, which gives slightly more than 1 million decimal places (1,048,576 to be exact):

Code:

$ ./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 : Tue Aug 5 18:38:40 EDT 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.502 Sec.
 I= 1 L=       0        Time=       1.671 Sec.
 I= 2 L=       0        Time=       1.932 Sec.
 I= 3 L=       1        Time=       1.936 Sec.
 I= 4 L=       2        Time=       1.935 Sec.
 I= 5 L=       5        Time=       1.938 Sec.
 I= 6 L=      10        Time=       1.932 Sec.
 I= 7 L=      21        Time=       1.940 Sec.
 I= 8 L=      43        Time=       1.927 Sec.
 I= 9 L=      87        Time=       1.933 Sec.
 I=10 L=     174        Time=       1.937 Sec.
 I=11 L=     349        Time=       1.939 Sec.
 I=12 L=     698        Time=       1.933 Sec.
 I=13 L=    1396        Time=       1.938 Sec.
 I=14 L=    2794        Time=       1.929 Sec.
 I=15 L=    5588        Time=       1.931 Sec.
 I=16 L=   11176        Time=       1.922 Sec.
 I=17 L=   22353        Time=       1.897 Sec.
 I=18 L=   44707        Time=       1.868 Sec.
 I=19 L=   89415        Time=       1.753 Sec.
 End of main loop
 End of calculation.    Time=      38.232 Sec.
 End of data output.    Time=       0.172 Sec.
 Total calculation(I/O) time=      38.404(       0.966) Sec.
 ------ Ended super_pi run : Tue Aug 5 18:39:20 EDT 2008

My machine is an AMD64 3700+ @2.1GHz (not dual-core) with 1GB DDR RAM. So I guess 38.4 sec isn't too bad for a single-core CPU.

algej · #6 6th August 2008, 12:21 AM

Code:

$ ./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 : Tue Aug 5 18:58:58 EDT 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.265 Sec.
 I= 1 L=       0        Time=       0.753 Sec.
 I= 2 L=       0        Time=       0.853 Sec.
 I= 3 L=       1        Time=       0.853 Sec.
 I= 4 L=       2        Time=       0.855 Sec.
 I= 5 L=       5        Time=       0.854 Sec.
 I= 6 L=      10        Time=       0.852 Sec.
 I= 7 L=      21        Time=       0.854 Sec.
 I= 8 L=      43        Time=       0.853 Sec.
 I= 9 L=      87        Time=       0.881 Sec.
 I=10 L=     174        Time=       0.860 Sec.
 I=11 L=     349        Time=       0.855 Sec.
 I=12 L=     698        Time=       0.858 Sec.
 I=13 L=    1396        Time=       0.853 Sec.
 I=14 L=    2794        Time=       0.848 Sec.
 I=15 L=    5588        Time=       0.859 Sec.
 I=16 L=   11176        Time=       0.841 Sec.
 I=17 L=   22353        Time=       0.825 Sec.
 I=18 L=   44707        Time=       0.798 Sec.
 I=19 L=   89415        Time=       0.743 Sec.
 End of main loop
 End of calculation.    Time=      16.861 Sec.
 End of data output.    Time=       0.105 Sec.
 Total calculation(I/O) time=      16.966(       0.821) Sec.
 ------ Ended super_pi run : Tue Aug 5 18:59:15 EDT 2008

My machine has 2 x E5345 @ 2.33Ghz Xeon (8 cores, 8 GB RAM). Too bad it's single-threaded.

sideways · #7 6th August 2008, 12:35 AM

ok, cool, I didn't know there was a lInux version (slightly unusual to convert the fortran to c and then optimise, wouldn't it be easier to just write it in c)

I wouldn't think the wine overhead would be a big deal, but this one is quicker (21.8 secs don't beat algej though)

Code:

]$ ./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 : Wed Aug 6 00:30:47 BST 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.318 Sec.
 I= 1 L=       0        Time=       0.963 Sec.
 I= 2 L=       0        Time=       1.102 Sec.
 I= 3 L=       1        Time=       1.101 Sec.
 I= 4 L=       2        Time=       1.102 Sec.
 I= 5 L=       5        Time=       1.101 Sec.
 I= 6 L=      10        Time=       1.100 Sec.
 I= 7 L=      21        Time=       1.104 Sec.
 I= 8 L=      43        Time=       1.103 Sec.
 I= 9 L=      87        Time=       1.104 Sec.
 I=10 L=     174        Time=       1.101 Sec.
 I=11 L=     349        Time=       1.099 Sec.
 I=12 L=     698        Time=       1.103 Sec.
 I=13 L=    1396        Time=       1.100 Sec.
 I=14 L=    2794        Time=       1.100 Sec.
 I=15 L=    5588        Time=       1.097 Sec.
 I=16 L=   11176        Time=       1.091 Sec.
 I=17 L=   22353        Time=       1.074 Sec.
 I=18 L=   44707        Time=       1.043 Sec.
 I=19 L=   89415        Time=       0.980 Sec.
 End of main loop
 End of calculation.    Time=      21.742 Sec.
 End of data output.    Time=       0.133 Sec.
 Total calculation(I/O) time=      21.875(       1.023) Sec.
 ------ Ended super_pi run : Wed Aug 6 00:31:10 BST 2008

Super Pi is single threaded, to really benchmark you should launch as many instances as you have cores, but this is just for a bit of fun right?

pete_1967 · #8 6th August 2008, 12:55 AM

Code:

[00:53:49 @mainlinux Pi]$ ./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 : Wed Aug 6 00:54:01 BST 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.306 Sec.
 I= 1 L=       0        Time=       0.998 Sec.
 I= 2 L=       0        Time=       1.159 Sec.
 I= 3 L=       1        Time=       1.154 Sec.
 I= 4 L=       2        Time=       1.156 Sec.
 I= 5 L=       5        Time=       1.155 Sec.
 I= 6 L=      10        Time=       1.157 Sec.
 I= 7 L=      21        Time=       1.154 Sec.
 I= 8 L=      43        Time=       1.155 Sec.
 I= 9 L=      87        Time=       1.158 Sec.
 I=10 L=     174        Time=       1.154 Sec.
 I=11 L=     349        Time=       1.168 Sec.
 I=12 L=     698        Time=       1.157 Sec.
 I=13 L=    1396        Time=       1.155 Sec.
 I=14 L=    2794        Time=       1.153 Sec.
 I=15 L=    5588        Time=       1.151 Sec.
 I=16 L=   11176        Time=       1.144 Sec.
 I=17 L=   22353        Time=       1.128 Sec.
 I=18 L=   44707        Time=       1.099 Sec.
 I=19 L=   89415        Time=       1.024 Sec.
 End of main loop
 End of calculation.    Time=      22.791 Sec.
 End of data output.    Time=       0.145 Sec.
 Total calculation(I/O) time=      22.936(       0.613) Sec.
 ------ Ended super_pi run : Wed Aug 6 00:54:24 BST 2008

notageek · #9 6th August 2008, 04:00 PM

Okay I got 42 seconds on AMD Athlon64 2800+

Code:

 ./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 : Wed Aug 6 20:28:42 IST 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.600 Sec.
 I= 1 L=       0        Time=       1.864 Sec.
 I= 2 L=       0        Time=       2.132 Sec.
 I= 3 L=       1        Time=       2.128 Sec.
 I= 4 L=       2        Time=       2.124 Sec.
 I= 5 L=       5        Time=       2.128 Sec.
 I= 6 L=      10        Time=       2.120 Sec.
 I= 7 L=      21        Time=       2.128 Sec.
 I= 8 L=      43        Time=       2.124 Sec.
 I= 9 L=      87        Time=       2.120 Sec.
 I=10 L=     174        Time=       2.124 Sec.
 I=11 L=     349        Time=       2.120 Sec.
 I=12 L=     698        Time=       2.124 Sec.
 I=13 L=    1396        Time=       2.116 Sec.
 I=14 L=    2794        Time=       2.124 Sec.
 I=15 L=    5588        Time=       2.112 Sec.
 I=16 L=   11176        Time=       2.104 Sec.
 I=17 L=   22353        Time=       2.080 Sec.
 I=18 L=   44707        Time=       2.028 Sec.
 I=19 L=   89415        Time=       1.892 Sec.
 End of main loop
 End of calculation.    Time=      41.975 Sec.
 End of data output.    Time=       0.204 Sec.
 Total calculation(I/O) time=      42.179(       0.900) Sec.
 ------ Ended super_pi run : Wed Aug 6 20:29:24 IST 2008

YeOK · #10 6th August 2008, 08:11 PM

Code:

./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 : Wed Aug 6 20:10:56 BST 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.902 Sec.
 I= 1 L=       0        Time=       1.287 Sec.
 I= 2 L=       0        Time=       1.350 Sec.
 I= 3 L=       1        Time=       1.347 Sec.
 I= 4 L=       2        Time=       1.341 Sec.
 I= 5 L=       5        Time=       1.336 Sec.
 I= 6 L=      10        Time=       1.337 Sec.
 I= 7 L=      21        Time=       1.339 Sec.
 I= 8 L=      43        Time=       1.338 Sec.
 I= 9 L=      87        Time=       1.336 Sec.
 I=10 L=     174        Time=       1.347 Sec.
 I=11 L=     349        Time=       1.342 Sec.
 I=12 L=     698        Time=       1.336 Sec.
 I=13 L=    1396        Time=       1.336 Sec.
 I=14 L=    2794        Time=       1.334 Sec.
 I=15 L=    5588        Time=       1.334 Sec.
 I=16 L=   11176        Time=       1.327 Sec.
 I=17 L=   22353        Time=       1.308 Sec.
 I=18 L=   44707        Time=       1.274 Sec.
 I=19 L=   89415        Time=       1.188 Sec.
 End of main loop
 End of calculation.    Time=      27.092 Sec.
 End of data output.    Time=       0.129 Sec.
 Total calculation(I/O) time=      27.221(       0.816) Sec.
 ------ Ended super_pi run : Wed Aug 6 20:11:23 BST 2008

AMD X2 4800+ @ 2500.

notageek · #11 8th August 2008, 03:27 PM

On AMD Athlon X2 4800+

Code:

./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 Aug 8 19:56:40 IST 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.430 Sec.
 I= 1 L=       0        Time=       1.340 Sec.
 I= 2 L=       0        Time=       1.523 Sec.
 I= 3 L=       1        Time=       1.521 Sec.
 I= 4 L=       2        Time=       1.526 Sec.
 I= 5 L=       5        Time=       1.521 Sec.
 I= 6 L=      10        Time=       1.521 Sec.
 I= 7 L=      21        Time=       1.523 Sec.
 I= 8 L=      43        Time=       1.522 Sec.
 I= 9 L=      87        Time=       1.521 Sec.
 I=10 L=     174        Time=       1.523 Sec.
 I=11 L=     349        Time=       1.522 Sec.
 I=12 L=     698        Time=       1.520 Sec.
 I=13 L=    1396        Time=       1.521 Sec.
 I=14 L=    2794        Time=       1.521 Sec.
 I=15 L=    5588        Time=       1.516 Sec.
 I=16 L=   11176        Time=       1.509 Sec.
 I=17 L=   22353        Time=       1.492 Sec.
 I=18 L=   44707        Time=       1.452 Sec.
 I=19 L=   89415        Time=       1.353 Sec.
 End of main loop
 End of calculation.    Time=      30.079 Sec.
 End of data output.    Time=       0.139 Sec.
 Total calculation(I/O) time=      30.218(       0.848) Sec.
 ------ Ended super_pi run : Fri Aug 8 19:57:10 IST 2008

I'm dissapointed

JN4OldSchool · #12 8th August 2008, 06:55 PM

Code:

./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 Aug 8 13:50:26 EDT 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.349 Sec.
 I= 1 L=       0        Time=       1.070 Sec.
 I= 2 L=       0        Time=       1.222 Sec.
 I= 3 L=       1        Time=       1.223 Sec.
 I= 4 L=       2        Time=       1.223 Sec.
 I= 5 L=       5        Time=       1.222 Sec.
 I= 6 L=      10        Time=       1.222 Sec.
 I= 7 L=      21        Time=       1.222 Sec.
 I= 8 L=      43        Time=       1.222 Sec.
 I= 9 L=      87        Time=       1.222 Sec.
 I=10 L=     174        Time=       1.222 Sec.
 I=11 L=     349        Time=       1.225 Sec.
 I=12 L=     698        Time=       1.219 Sec.
 I=13 L=    1396        Time=       1.220 Sec.
 I=14 L=    2794        Time=       1.218 Sec.
 I=15 L=    5588        Time=       1.216 Sec.
 I=16 L=   11176        Time=       1.211 Sec.
 I=17 L=   22353        Time=       1.191 Sec.
 I=18 L=   44707        Time=       1.157 Sec.
 I=19 L=   89415        Time=       1.080 Sec.
 End of main loop
 End of calculation.    Time=      24.106 Sec.
 End of data output.    Time=       0.147 Sec.
 Total calculation(I/O) time=      24.253(       0.966) Sec.
 ------ Ended super_pi run : Fri Aug 8 13:50:51 EDT 2008

AMD 9850BE 4GB DDR2 1066

I must say I am a bit disappointed also. Too bad it isnt multi-threaded.

pete_1967 · #13 8th August 2008, 08:01 PM

Quote:

Originally Posted by JN4OldSchool

AMD 9850BE 4GB DDR2 1066

I must say I am a bit disappointed also. Too bad it isnt multi-threaded.

Makes an interesting comparison to my 9950 especially considering that we got the systems at the same time.

JN4OldSchool · #14 8th August 2008, 08:12 PM

Quote:

Originally Posted by pete_1967

Makes an interesting comparison to my 9950 especially considering that we got the systems at the same time.

If the test were more definitive I would overclock it briefly to see if that helped. As it is I turned OFF compiz-fusion and got WORSE results! I cant figure that one out. But they are all within tenths of a second difference, so it is close.

RupertPupkin · #15 9th August 2008, 01:18 AM

Interesting that I consistently get a slightly faster time when using F9 x86_64:

Code:

$ ./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 Aug 8 15:58:51 EDT 2008
 Start of PI calculation up to 1048576 decimal digits
 End of initialization. Time=       0.458 Sec.
 I= 1 L=       0        Time=       1.510 Sec.
 I= 2 L=       0        Time=       1.740 Sec.
 I= 3 L=       1        Time=       1.747 Sec.
 I= 4 L=       2        Time=       1.741 Sec.
 I= 5 L=       5        Time=       1.740 Sec.
 I= 6 L=      10        Time=       1.747 Sec.
 I= 7 L=      21        Time=       1.738 Sec.
 I= 8 L=      43        Time=       1.737 Sec.
 I= 9 L=      87        Time=       1.747 Sec.
 I=10 L=     174        Time=       1.743 Sec.
 I=11 L=     349        Time=       1.734 Sec.
 I=12 L=     698        Time=       1.746 Sec.
 I=13 L=    1396        Time=       1.737 Sec.
 I=14 L=    2794        Time=       1.733 Sec.
 I=15 L=    5588        Time=       1.736 Sec.
 I=16 L=   11176        Time=       1.726 Sec.
 I=17 L=   22353        Time=       1.704 Sec.
 I=18 L=   44707        Time=       1.671 Sec.
 I=19 L=   89415        Time=       1.563 Sec.
 End of main loop
 End of calculation.    Time=      34.377 Sec.
 End of data output.    Time=       0.167 Sec.
 Total calculation(I/O) time=      34.544(       1.145) Sec.
 ------ Ended super_pi run : Fri Aug 8 15:59:26 EDT 2008

This is on the same machine (AMD64 3700+ @2.1GHz) as the one I used before (I got about 38.4 sec in 32-bit F7).