fx57 2d runs with ln2

[k|ngp|n]

Took a little break from running 3d to get some quick pi numbers. Here are some quick runs off of my 3d benching drives for 1m, 8m, and pifast (my 3 favs). With more tweaking and the OS setup right, these times could get pretty sick with these cpu clocks:

DFI SLI-d
1 gig mushkin lv2 bh-5
Dragon pot ln2

1m @ 3950mhz

validate:
http://valid.x86-secret.com/show_oc?id=62550

8m @ 3928mhz

validate:
http://valid.x86-secret.com/show_oc?id=62548

Hexus pifast @ 3962mhz

validate:
http://valid.x86-secret.com/show_oc?id=62546

I am not the best PI tweaker around. I usually just set the clocks and let it rip, so any suggestions on how to get my times lower using OS or registry tweaks are always appreciated

[lawrywild]

those are WR's right???

Congratz!!!

[giorgioprimo]

nice .....really nice, is that a 530 isn't it?

[k|ngp|n]

pifast is AMD WR. I am going to have to battle team japan for the sp records. I have the clocks for it, just need a little more tweaking.

[piotres]

I think it must be 0530apaw

Shoot 4.2 ghz hehe

[lawrywild]

i think, can't remember exactly, he said it was CABNE 0530 apmw in the trifecta thread

[[XC] moddolicous]

Yea, I'm almost positive that it was a CABNE 0530 apmw in the trifecta page. Yea, you definetly have the clocks to get team japan, although it wont be easy lol. I think you'll get it though.

[NiCKE^]

You've got a great pair of memory and got very nice clocks, just the tweaking part and you will have the WR

[STEvil]

limit it to 127mb of memory or so.. dont remember the exact number. Run 16k twice before running 1m (OPB says its good?), and.........?

[afireinside]

max mem 104

Nice KP, was 3950 max 1m? Odd that 3d05 was higher...

[k|ngp|n]

I almost had a 1m pass at 3980 but it errored out loop18. mem controller seemed to error out alot after that point, so I just went for a safe run before shutting down so I could at least get a decent 1m score.
I suspect pi doesn't put quite the load on the processor that 3d does and the cold bug trips quicker giving rounding and converg errors

[POGE]

Man thats crazy! So close to 4.0... I bet you can still hit it!

[dinos22]

pifast is AMD WR. I am going to have to battle team japan for the sp records. I have the clocks for it, just need a little more tweaking.

+ OS: Windows Server 2003

[NiCKE^]

Maybe if you run prime or similar a little while before starting the 1M round?

[T_M]

Run 16k twice before running 1m (OPB says its good?)

Interesting, thats a new one to add to my list of tweaks

Vince, have you done the usual basic reg tweaks such as SecondLevelDataCache etc?

Ive got a small basic tweak guide i use for WinXP 3D benching which is also mostly applicable to 2D (+maxmem etc), not sure if all is applicable for server 2003 (presuming you are using that yes?)
http://i4memory.com/showthread.php?t=833
But youve probably done all these already...

[zakelwe]

Wonder what it wil ltake to get under 30s for pifast ? Strangely I much prefer pifast to superpi, I like the 2 stages, the countdown then the calculation at the end, you always think you will be faster than you are.

Nice work Vince.

Regards

Andy

[Athens[2004]]

Xtreme 57@runs

[Captin Insano]

Very Nice Kingpin.. But if its 0530... I thought most if not all the 0530 has cold bug issues. Maybe im wrong.. But everyone ive seen and had have had it. Anyhow very nice OC

[MaRtIe]

Whoa... nice, fastest stable FX clocks ive seen my self.

[bazx]

Very Nice Kingpin.. But if its 0530... I thought most if not all the 0530 has cold bug issues. Maybe im wrong.. But everyone ive seen and had have had it. Anyhow very nice OC

hi captin its the 0531 that bug out 30 is ok

@KP how do you keep a constant temp with the ln2?

[dinos22]

hi captin its the 0531 that bug out 30 is ok

@KP how do you keep a constant temp with the ln2?

CABNE 0530APMW Opties based on that stepping have the bug too......different temps...seen a few from past posts

[VictorWang]

FX57 under LN2 , soooo good to avoid coldbug

[k|ngp|n]

Wonder what it wil ltake to get under 30s for pifast ? Strangely I much prefer pifast to superpi, I like the 2 stages, the countdown then the calculation at the end, you always think you will be faster than you are.

Nice work Vince.

Regards

Andy

Not much prolly andy. Run I posted was at 31.03 on a very dirty OS and not 100% fully tweaked, so I am thinking 30.** for final AMD pifast score. That would be awsome. I love pifast too..My previous best was 31.72 with 273x14=3838mhz on week 16 chip.

[k|ngp|n]

Interesting, thats a new one to add to my list of tweaks

Vince, have you done the usual basic reg tweaks such as SecondLevelDataCache etc?

Ive got a small basic tweak guide i use for WinXP 3D benching which is also mostly applicable to 2D (+maxmem etc), not sure if all is applicable for server 2003 (presuming you are using that yes?)
http://i4memory.com/showthread.php?t=833
But youve probably done all these already...

Thank you very much for the info. I didnt use alot of those tweaks. I will when I bench a dedicated pi session.

[Cornelious0_0]

Very nice k|ngp|n, I knew you had some magic left...now get tweakin' and show us some more WR's.