Is SuperPi 32M enough as stabilty test? [Archive]

mongoled

05-14-2005, 09:03 AM

Hi peeps, very simple poll, as the title says. Ive noticed recently many threads with peeps posting the settings that their CPU are run at. Alot of these are not backed with any real proof of stabilty. Some post just a SuperPi 1M and thinks this means its stable

:p:

Ive found tht only a few proggies show any real stabilty, for me they are prime95, OCCT and FarCry. These hammer the CPU, superPi 32M unfortunately is not one of them. Here is a screenie of wht the Venice I have can do

SuperPi 32M: 3003mhz 1.67V (333x9 Mem@273 5:6 Divider)

http://homepage.ntlworld.com/m.meliou/SuperPi_32M_3003mhz.gif

SuperPi 1M: 3037mhz 1.67V (337x9 Mem@276 5:6 Divider)

http://homepage.ntlworld.com/m.meliou/SuperPi_1M-_3037mhz.gif

this is no means stable, even though the test has hammered the CPU for over 37 minutes!! So wht do peeps think SuperPi 32M enuff or not??

i found nemo

05-14-2005, 09:14 AM

nah not really imo

bachus_anonym

05-14-2005, 09:16 AM

SuperPi 32M calculation cannot be treated as an ultimate stability test. It's still a benchmark, but one that requires much more stable system to complete, compared to SuperPi 1M, Pifast or 3D benchmarks. It's probably one of the most demanding benchmarks there is. After that, there are only torture-type apps.

Also, NO... SuperPi 32M stable doesn't mean same as Prime95 stable; it never did. I can give you perfect example... My system is 32M Pi stable @ 3034MHz/DDR550, but for Prime95/OCCT to NOT fail, I need to back down to 2900MHz/DDR530.

[XC] leviathan18

05-14-2005, 09:25 AM

stability is your conception maybe sp32mg is ok or even if you can play all your games for hours without any problem that is stability many chips wont pass prime and they can play games and everything

LoKi2k

05-14-2005, 09:31 AM

a few days ago i get my system "superpi 32m-stable". but after 10 minutes playing cs:source my computer ´get off.

so my answer is: NO!!!

ken

05-14-2005, 09:33 AM

nope :D

charlie

05-14-2005, 09:38 AM

a few days ago i get my system "superpi 32m-stable". but after 10 minutes playing cs:source my computer ´get off.

so my answer is: NO!!!

that's becasue the instability was in your GPU, not CPU/MEM subsystem.

C

dnottis

05-14-2005, 09:43 AM

Some post just a SuperPi 1M and thinks this means its stable

Yea, that always kills me 1M LOL. Even Super PI 32M doesnt mean anything. Everyone here posts their 1.39v @ 2.8 Ghz with a Super PI 32M - BFD show me Prime for 24 hours screen shot. You need to be Super PI 32M, 24 hours or Prime95 and gaming stable. I fould 3dm 2003 to lock if my CPU is unstable even when 3dm 2001 SE and 3dm 2005 complete fine, just one more thing I check for "true" stability.

Dean

trgonz

05-14-2005, 10:57 AM

I think both of these programs do a good job of telling you where you are. I guess the real question is... What is the best program for testing the entire system? And not just for a few nimutes on a overclocked system... But for 24 hours to give you the best look at where your sustem sits.

TG

EMC2

05-14-2005, 11:10 AM

TheMeatFrog

05-14-2005, 11:17 AM

I think I voted no but I can't remember :slap:

ben805

05-14-2005, 11:26 AM

if all you want is run SuperPI 32M all day long, then the answer is YES. But I voted for NO because I needed my rig to run more than just SuperPI 32M, for ultimate stability tests...you will need to run atleast several more tests, for gamers...if you can loop 3DMark overnight, it's almost good enough. For me...ultimate stability mean being able to perform multiple tasks simultaneously like running Prime95 blend test in the background while I rip and burn DVD, surf web, play mp3 in kernel streaming mode, and do a bunch of other stuff at the same time.

ben805

05-14-2005, 01:07 PM

My system is 32M Pi stable @ 3034MHz/DDR550, but for Prime95/OCCT to NOT fail, I need to back down to 2900MHz/DDR530.

even with the chilly1 phase change you still can't get it to prime stable at 3Ghz??

:shocked:

zA.Gosu

05-14-2005, 01:31 PM

ben805, you use foobar?

ImportantAwareness89

05-14-2005, 01:38 PM

I like seeing 1M times because I don't memorize 32M times. 1M times is a good way to show that your CPU is friggn fast.

ben805

05-14-2005, 01:46 PM

ben805, you use foobar?

yup, sometime I also use the winamp with the kernel streaming plugin :D :D

Nettwerk

05-14-2005, 01:59 PM

2 times 32sp is stable enuff for me

CrimeDog

05-14-2005, 02:24 PM

you can run superpi 32m at speeds that will crash other applications and benchmarks, so of course it's not enough. not referring to video card or other instabilities either.

Gautam

05-14-2005, 03:13 PM

No way.

jikdoc

05-14-2005, 03:52 PM

I didn't vote, because it depends ;)

Stable enough for what?? To pop a wheelie? To drag-race? To run Daytona? To road-race? To do the 24-hours at Lemans? To run a taxi service in NY 24/7?

Stability is simply this - the ability to do what the user wants it to do at a particular moment in time for a specified period of time :D

perfectly states my point of view.

i treat spi32m as a bench, see how fast i can go. for 24/7 use, i run OCCT. there's NO appreciable difference in what i use my rig for (games, word processing, powerpoint) between 3.25Ghz and 3.0Ghz which is the difference in 32m and occt.

mongoled

05-14-2005, 05:19 PM

I didn't vote, because it depends ;)

Stable enough for what?? To pop a wheelie? To drag-race? To run Daytona? To road-race? To do the 24-hours at Lemans? To run a taxi service in NY 24/7?

Stability is simply this - the ability to do what the user wants it to do at a particular moment in time for a specified period of time :D

I like your answer

:)

Your spot on with wht u say, everything is RELATIVE, lets say the poll is rephrased to 'Is SuperPi 32M enough as stabilty test relative to your needs?' tht would be more appropriate, and then u could answer my question

:p:

mad mikee

05-14-2005, 05:21 PM

SuperPi is good 1st test but because of the way it works I speak of the above.
Thats Why I like OCCT, you can set it to use as much memory as prime but it hammers it ALOT more (loads and unloads every couple of minutes, and no it is NOT diskbound doing it!) :D

Napoleonic

05-14-2005, 05:25 PM

what about running ALL super pi benchmarks?

agenda2005

05-14-2005, 07:43 PM

okay here we go, the pics speak for themselves

Chip is from Monarch

ADA3700DAA5BN
CABGE 0515 SPMW

Set up
MB DFI Ultra-D with 310 Official Bios
Memory :GSkill LE PC 4400 at 246MHz 2.5-3-3-7
Cooling: Thermalright XP-120 with 120mm Panaflo at 110cfm
GFX: eVGA 6600GT at stock speed
Room Temperature : 29.4C
PSU : PC Power & Cooling 510 deluxe

ben805

05-14-2005, 07:53 PM

okay here we go, the pics speak for themselves

Chip is from Monarch

ADA3700DAA5BN
CABGE 0515 SPMW

Hey Agenda, can you please post a bigger picture? can't read any of them and I would like to see some of your results :D :D

By the way, I'm going to switch over to OCCT and ditch Prime95....it's very neat,
check it out (http://www.xtremesystems.org/forums/showthread.php?t=49487)

agenda2005

05-14-2005, 07:56 PM

Hey Agenda, can you please post a bigger picture? can't read any of them and I would like to see some of your results :D :D

By the way, I'm going to switch over to OCCT and ditch Prime95....it's very neat,
check it out (http://www.xtremesystems.org/forums/showthread.php?t=49487)

First open the attachments and then enlarge it by clicking on the picture once. It should be clear from there. If you still have problem viewing it, please let me know.

Gray Mole

05-15-2005, 01:30 AM

While I totally agree with it depending entirely on your particular stability needs, I voted yes, because even 1-2 meg means I'm stable enough to start hitting the 3d benchies with vengence...

If I need more than that, it's done on one of the other rigs in the house...

Heck if it boots, it's stable enough for me :rehab:

Gray

Pjotr

05-18-2005, 02:22 AM

Prime95 FFT test is good test for testing your *CPU* stability, SuperPI is not as good. However, this doesn't test your total system stability. For RAM stability, use memtest86 24 hours, especially test 5 and 6 will fail first. For graphics, the 3DMark series is a common way to test your basic graphics stability, but not conclusive. For HD stress testing, I'm not sure. For ethernet you could find some ethernet benchmarking programs. For USB/Firewire, don't know. Sound? Don't know.

As for *system* stability, use your computer the way you intend to use it. If you ever experience a crash or problem and you think it's not buggy software/drivers, you have to improve your system (lower clocks, better PSU, more cooling etc). In general, without overclocking and with a required PSU, you should never be unstable from hardware, only software/drivers.

Soulburner

05-18-2005, 07:08 AM

For me, it seems the speed I can run Super Pi 1M is the same speed I can game at...its wierd, but i've been able to game for hours and still crash 1M...

Pjotr

05-18-2005, 08:03 AM

So your game is leaner on the CPU than SuperPI, this is common.

Fr3ak

05-18-2005, 02:34 PM

I also vote for no.
But I stopped trying to get every stability test programm out there run successfuly, because all that matters for me is to play games without crashing. Memtest even fails at far below stock speeds, so I dont take everything too seriously :)

[XC] moddolicous

05-18-2005, 03:12 PM

I voted no becuase I've had it crash, even though it was superpi 32m stable. I find D2OL to be the best, right up there with prime. It just uses your comp @ 100% for a while.