Couple of SuperPi Questions

Virdi · 26 Sep 2006 at 15:36

Hi all,

Sorry - I'm a newbie to SuperPi...

Could somebody just advise what calculation everyone does? i.e. 1M or 32Mb etc..? and do you need to run two instances for dual core rigs..?

Thanks

wizardmaxx · 26 Sep 2006 at 15:38

Most of the 'top 10' or 'top 20' use the 1meg calc. Best i got was 47sec though

. Dont know about the dual cores though

Jobe · 26 Sep 2006 at 17:27

1m is the norm.
i get about 14 secs ( i think) on my 6600 @ 3.2
and only one run at a time is the norm.

Jobe

w3bbo · 26 Sep 2006 at 19:51

To fully test your cpu then you should run two versions (Super PI should only really be used to test for VERY basic stability). It appears however thay most only do 1 core/1 meg run. Imo that is a benchmark and by no means a test of stability.

jellybeard999 · 26 Sep 2006 at 19:54

w3bbo said:
To fully test your cpu then you should run two versions (Super PI should only really be used to test for VERY basic stability). It appears however thay most only do 1 core/1 meg run. Imo that is a benchmark and by no means a test of stability.

the test is supposed to be a benchmark :rolleyes:

1m = time it takes to calculate pi to 1 million decimal places, 32m = 32 million etc

Raikiri · 26 Sep 2006 at 21:03

Yep, as said 1m on 1 core is the standard done for comparisons. Damn you C2D people

33s is my best (805)

wizardmaxx · 26 Sep 2006 at 21:10

Raikiri said:
Yep, as said 1m on 1 core is the standard done for comparisons. Damn you C2D people 33s is my best (805)

And you feel stupid !!!

Raikiri · 26 Sep 2006 at 21:14

Yes... I feel stupid for building a near silent, heavily OCed, dual core system with 2gb of ram for around £350 months before C2D came out when I wanted/needed it.

Hamoodii · 26 Sep 2006 at 21:14

w3bbo said:
To fully test your cpu then you should run two versions (Super PI should only really be used to test for VERY basic stability). It appears however thay most only do 1 core/1 meg run. Imo that is a benchmark and by no means a test of stability.

Hmmm, well when my CPU was not stable I ran the SUPERPI 32m test and when my CPU was not stable it would come up: "NOT CONVERGENT IN SQ" or something like that!

vocch · 26 Sep 2006 at 21:16

wizardmaxx said:
Best i got was 47sec though .

And you even more.

wizardmaxx · 26 Sep 2006 at 21:20

vocch said:
And you even more.

Thanks feel more stupid now.

All u conroe dudes watch out when i get my 6300 in december and clock it 8ghz prime stable.

WJA96 · 27 Sep 2006 at 05:56

SuperPi is not a good benchmark as it's very memory speed dependent. My E6600 at 3.1GHz does 16s with the RAM at 667 5-5-5-15 or 15.5s with the same RAM at 667 4-4-4-12. If I turn the RAM up to 800 at 5-5-5-15 it does it in 14.6s.

stigggeh · 27 Sep 2006 at 11:31

Hamoodii said:
Hmmm, well when my CPU was not stable I ran the SUPERPI 32m test and when my CPU was not stable it would come up: "NOT CONVERGENT IN SQ" or something like that!

yeah your overclock has to be very flakey to do that.... usually though you can pass 32mb of pi and fail prime after an hour or two, hence super pi should be only used as a method of pimping your pc

i had sub 25s OVER a year ago too