Pifast Benchmark

[Comp-Freak]

Here I come with my sad score My p4 won't get any higher, maybe soon with phasechange cooling.

- P4 northwood 3,0
- Gigabyte ga-8ik1100 rev 2
- zalman cnps7000 Cu
- 2x Corsair XMS3200C2PT rev 4.2, 256MB, 1:1, 2-3-3-6

[SynGroW]

Celeron ''Northwood'' 2.4GHz @ 3815MHz



Intel Celeron 2400 @ 3815MHz
Asus P4P800SE
Corsair TWINX512-3200LLPT v1.1 (BH-5) @ DDR398 2.9v
Custom PhCh - Single Stage (Epsilon's Promkiller)

No validation link because the system rebooted when I tried to save that file

[Pedro Rocha]

Here is my FX-57, cooled by an MachII ..DFI mem volt jumper still at 3.2v so only 257mhz on the Redlines



Anyaway not a bad start

[GoriLLakoS]

Here is mine at 3520mhz....memory at 3.70V with maximizer ona a dfi sli

[M.Beier]

Uhm, missed this thread..

33.19 sec, reached at netbyte using LN2 on a FX55 clawhammer

Aint home, so cant post pictures, but quite some time ago.

[fredyama]

By New Beetle

Return match of Pifast with his M780.
Many thanks to Hipro5 for your mods

Time: 28.83s
Clock: 4098.2MHz(FSB256.1MHz x 16)
M/B：ASUS P4P800-SE #2 mod + CT-479 mod / 55MM339*** / BIOS：P4P8SE1011.ROM
M/B Bed：「V-TEC ARC Bed Rev.3.3」
CPU：Pentium M 780 / 7506B***
CPU Cooling：「KuriPro２」+ LN2
Memory：SanMax PC3200 512MB x2 + DICE / CL2.0-2-2-5-4-A-A-A
Vcore：1.85V（Actual）/ Vddr：3.40V / Vcca：2.40V / Vio：3.65V
VGA：ATI XPERT 98 PCI 8MB
HDD：IBM-DTLA-307030 /U-DMA 5 (ATA100)
PowerSuplly：Zippy-460WS #2 （3.3Vmod）

[zakelwe]

k|ngP|n on 18.5s

http://pic20.picturetrail.com/VOL256.../176350862.jpg

Regards

Andy

[Boble76]

FX57 with DI, one-timer pifast. 32.88s

[sierra_bound]

X6800 + Vapochill LS

[Garrett]

I have 36.61... see my sig... pretty pathetic NOW huh?

[demiurg]

24.45 - E6700 @ 3980MHz (398*10) 1.6V
http://www.hwbot.org/compare.do?resultId=558614

[jvargaszabo]

Hi, i'm new, but i'm doing a research project about cpu temps when calculating pi. If anybody could post some times temps n cooling methods they used to achieve them, i'd appreciate it. Any information helps.

thx

[flclisgreat]

i feel dumb for asking this but how many number's are we supposed to calculate?
how many is normal?

[Viss]

E6600 and P5B dlx:

[sysfailur]

i feel dumb for asking this but how many number's are we supposed to calculate?
how many is normal?

hexus.pifast.net download the package, run the bat, not the exe

Also.... (air)

[sparkY1000]

Sorry to sound like a complete newb but how do you run this benchmark? It gaves me so many options and im just not sure which ones I want? sysfailur says something about running a bat file but there wasnt one in the download just an exe and a txt

[PMM]

I Stopped maintaining Pifast because of all the cheating as it is cheatable, I did have a word with the program maker but he was un-interested in sorting a fix.

However its as simple as... downloading .... http://pifast.hexus.net/hexus_pifast.zip

and then run the file hexus_pifast.bat

[spaglia]

I don't think QuickPi is cheatable. How are the cheats done?

[PMM]

A person would have to go to extra lengths to effect QuickPi but I have just confirmed it can be messed with... see pic below

A liitle time and effort and easily made less noticable.

[spaglia]

A person would have to go to extra lengths to effect QuickPi but I have just confirmed it can be messed with... see pic below

A liitle time and effort and easily made less noticable.

I see, of course -- changing the system clock skews the times. I was thinking more along the lines of tampering with the output. Ah well, where there is a will, there usually is a way...

[spaglia]

A person would have to go to extra lengths to effect QuickPi but I have just confirmed it can be messed with... see pic below

A liitle time and effort and easily made less noticable.

I think I can fix this, anyone interested?

[PMM]

After running this again on my home machine a few times I like it esp as its multicore friendly it kind of brings in a new type of tinkering to get max speed the run above I did at work.

I think if you can certainly put into place....
1) An anti cheat
2) A Validation at the end to confirm the calculation is all is OK
3) Suggest an autorun batch file to say kick off a 10M run as a fixed benchmark standard.

I think you have a winner :nod:

[spaglia]

After running this again on my home machine a few times I like it esp as its multicore friendly it kind of brings in a new type of tinkering to get max speed the run above I did at work.

I think if you can certainly put into place....
1) An anti cheat
2) A Validation at the end to confirm the calculation is all is OK
3) Suggest an autorun batch file to say kick off a 10M run as a fixed benchmark standard.

I think you have a winner :nod:

Version 4.5 is is available here:

http://home.istar.ca/~lyster/pi.html

Let me know if you can easily cheat it. The app can compute a checksum over all the digits and you can verify the correctness that way (by comparing it to a known good checksum). Use the -check option. Other options are the -bin and -stats options. By submitting the resulting binary file, it can then be independently varified.

[PMM]

Passed my test but I admit I'm not clued up in the ways of tripping progs up.

I have PM'd Kunnak here in the view if he is willing to put his superior know how to give it a thumbs up.

Excellent work by the way