Has anyone done close to 500FSB on X38 boards?

[zfooz]

My board ABIT IX38 (X38 Chipset) and E8400 just doesn't like anything over the 450ish FSB range.. I even tried like 480 x 6 or 500 x 6 and it runs but errors on prime. :/

If anyone has gotten this board or any other X38 board to run close to 500FSB (with a 45nm cpu) please tell me what voltages you set things to. I tried so many combinations my head is spinning.. I don't care if I get 500 x 7 (cuz this cpu won't do 4ghz without 1.4v or higher which I'm not doing), but i'd like to get the ram at 1:1 divider at 1000mhz if possible.

Thanks!

 

[runawayprisoner]

45nm dual-core:

http://i235.photobucket.com/al...ongvietpham/3DMark.jpg

45nm quad-core:

http://valid.x86-secret.com/show_oc.php?id=334952

Board: ASUS P5E-VM HDMI
Chipset: G35

For 45nm dual-core: all stock volts!
For 45nm quad-core: all max volts!

Yeah... quads require more than duals. But I gotta say... having an mATX board beating some X38 boards make me feel lucky.

But anyway, here's a piece of advice: try upping your northbride and FSB termination (VTT) voltages. They help in a lot of cases with FSB limitations.

 

[zfooz]

Well I can get it to boot at 520FSB and show a CPU-Z.. that doesn't interest me.. I want it to be able to run dual prime95 instances for 24 hours or more with 500 FSB.. If it's not possible then it's not possible and I'll do 450 or something but I just want to know.
I've tried northbridge voltages up to 1.58 and cup VTT up to 1.4 I think and I saw no difference at all at 500FSB. The only voltage that made any difference at all was the CPU voltage. But even up to 1.37v it still won't run at 4ghz or 500 FSB even at 3ghz.
Either I have to use 1.6+v on the northbridge (which I don't want to fry it) or 1.4+v on the cpu which I don't wanna fry it either, or higher than 1.4 on VTT which again.. to me is a lot..
Do you think it's safe to use 1.6V or higher MCH(NB) and 1.4 or higher CPU VTT?
I think I also have a DDR VTT which might help?? I haven't tried touching that yet and I also haven't tried touching the GTLREF

 

[bryanW1995]

I have an ip35 pro, I think that the adjustments are similar to yours. I have my mch at 1.52 and vtt at 1.31 in bios with my mobo completely stable up to to 455x8 on an x3350. 450-460 seems to be the "safe" range for most 45nm quads though I have heard of a few that will creep up closer to 500.

aigomorla has more quads than many small countries. He realizes that the numbers don't matter without orthos (or maybe occt) stability.

 

[runawayprisoner]

Try some higher vPLL then? And you know, sometimes, BIOS settings don't match the real voltage. There might be some tolerances. The only way to know the real volt going through your motherboard components is to measure them directly from the motherboard.

 

[zfooz]

Yea.. That's my feeling on the matter. I could probably run it at 520 and 4.2ghz or something but it would not be stable. Stable is orthos/prime blend test using 2gb ram for more than 12 hours at least, preferably 24h. Nothing else matters if you can't use the darn thing. I expect my computers to have high uptime. I reboot or power off maybe once a month and if they can't stay running that long it's not worth it.
I don't see why E8400 wouldn't do 500.. I see a lot of people doing 500*8 which would be perfect, but maybe theirs isn't stable. I wish we could set a standard, like don't even post your overclock unless you have 24 hour prime95. P95 blend test is the best, or custom settings to make sure it uses most of your RAM.
Right now I'm trying 451mhz * 8.5 with ram at 1:1.20 and I still need almost 1.29v in the bios for this.. crazy.. It will do 3.6ghz with 1.15v and needs another .15 just to get to 3.8something? Or maybe it needs a lot more VTT or MCH..
What is vPLL? Is that something to do with the GTLREF?
I haven't messed with GTLREF or DDR VTT. GTLREF on my board says 67%/67% and the DDR VTT is like +10mv +20mv etc.

 

[runawayprisoner]

Try DDR VTT at +30mv or +40mv then. Now I know one thing for sure: your RAM is what's failing. You need about at least 2.10v going through your RAM sticks to get them to run stable at 1000MHz.

 

[Xpoc]

I've had this board prime stable at 8 x 500mhz with my E8400, and I probably could have went higher.

 

[zfooz]

If it's my ram, why does it help increasing the CPU voltage?
I tried setting the ram to 2.275 volts even and it's rated at 2.1
At 500FSB I tried

MCH 1.55
CPU VTT 1.40
CPU 1.38
DDR 2.275

The GTLREF, DDR VTT I left alone and I bumped up the ICH stuff by .05
It won't even run at 500FSB * 6.5 or 7.. The ram works fine at 1080MHZ at 451 FSB with 1:1.20 divider, so why wouldn't it work at 500FSB ? Something isn't adding up here.. Either the board just can't do 500 FSB or the CPU just doesn't like it regardless of the MhZ
My cpu really doesn't like being near 4ghz , it takes at least 1.3V just to get to 3850mhz :/

Upping the CPU voltage at 500fsb was the only thing that got me a longer prime run. I went up to 1.375 and it still made no difference and I don't want to go any higher, ever

I'm thinking maybe the DDR VTT and the GTLREF might need some changing or maybe more CPU VTT but 1.40 is already .30 over the stock setting which is a lot.

Can you tell me your settings for 500 FSB with E8X00 cpu?

 

[Tempered81]

Originally posted by: zfooz
just to get to 3850mhz :/

Sorry man, that's unfortunate

 

[zfooz]

LOL
I'm dissappointed, I wanted a 1ghz overclock
It's running stable at 451x8.5 I'd post a screenshot if I could but it looks like I have to put the file on a server somewhere instead of uploading it here.
Almost 120mhz FSB overclock so I'm happy. Don't have to have 500 as long as the thing works and does what I want.. It's just one of those things I guess you want to try and do just to see if it can do it........ but i guess it can't!!