Help, good RAM can't even hit 205mhz

^yeah Bop, i got lucky... i haven't even pushed this thing yet :)


And Cyp, i thought instead 'painting't the line between the pins, you could use some glue and something else to connect em...

Yeah. You don't have to get your shorts in a knot.
 
pardon me for considering physical motherboard modifications for a few extra benchmark points stupid. at least on my DFI i can run a factory bios on a stock board and still do with same without jamming metal into my motherboard or trying to paint a super fine line while not letting it get into the wrong pin holes.
 
JackieChanIsGOD said:
^yeah Bop, i got lucky... i haven't even pushed this thing yet :)


And Cyp, i thought instead 'painting't the line between the pins, you could use some glue and something else to connect em...

Yeah. You don't have to get your shorts in a knot.
Click to expand...

Not at all.. just wondering what glue method you're referring to?? Thought perhaps you got confused with using glue to help unlock a Thoroughbred.
 
CzarDestructo said:
pardon me for considering physical motherboard modifications for a few extra benchmark points stupid. at least on my DFI i can run a factory bios on a stock board and still do with same without jamming metal into my motherboard or trying to paint a super fine line while not letting it get into the wrong pin holes.
Click to expand...

What pin holes?? you do it on the back of the mobo..away from the CPU ..

Never doubted your DFI... read what I wrote earlier this thread:


Yes with Dual Channel on nForce you should run the last timing at 11. This has proven many times to be the most stable and the best performance. There's been many posts in the past on this.

Also, make sure you do the L12 trick to get past 215 FSB.. some boards can get much higher after the trick and some will have a small gain. Mine, for example, was a small gain. I couldn't get past 215 but after the mod it went to 227 as the limit.

DFI LanParty Ultras can get over 240+ usually.
Click to expand...

You fellas need to read up before posting.
 
CypHill said:
Not at all.. just wondering what glue method you're referring to?? Thought perhaps you got confused with using glue to help unlock a Thoroughbred.
Click to expand...

Ooo... hmm, probably got stuck in my head... i was thinking that you could connect the pins on the top of the chip with conductive ink, then fill the gap with glue :) lol.
 
I think that 430w may not be enough for my CPU. I set my voltage to 1.875v int he BIOS and in the EQ it shows it at 1.82v while the system is idle and it goes down to 1.78v during Prime95.

My CPU is also reaching 54C and I expect it to hit 55C soon. I think I need to get a SLK900A now, or maybe even a SP97. Are those worth the extra installation time?
 
unmounting a motherboard and painting those on for 15mhz seriously isn't worth it. also,how is it easily reversable? would you cut the material after it drys, risking the traces below?

and yes i read your other post, i was implying that if you want FSB it would be better to get a product that can do it rather than have to modifiy one that can't stock.
 
CzarDestructo said:
unmounting a motherboard and painting those on for 15mhz seriously isn't worth it. also,how is it easily reversable? would you cut the material after it drys, risking the traces below?

and yes i read your other post, i was implying that if you want FSB it would be better to get a product that can do it rather than have to modifiy one that can't stock.
Click to expand...

Omg.. seriously stop arguing something you know nothing about. I hate to be rude but please.. you're sounding ridiculous.

This material is a copper based paint... when it's applied and cured it's like whiteout texture almost. You can easily scratch it off lightly with a needle. If you're complaining about taking out a mobo, what are you doing on [H]ardOCP in the first place? We're here to overclock what we can.
 
Dang, Prime lasted 10 minutes at 11.5x220. It lasted an hour at 10x220 until I stopped it so it looks like it is the CPU that is holding me back or my PSU, seeing as the vcore fluctuates between 1.78 and 1.82, and I set it to 1.875v in the BIOS.

Or maybe 55C is too much for the CPU to handle?
 
the material is whatever you can get at a local store, i've had to use a kit in the past and the one i got at auto zone certainly wasn't a copper based material that would easily come off. it was a thick silver paste that hardened to an almost solder like material. thats what i pictured so don't get bent out of shape over me questioning your ethics.

how am i sounding ridiculous?

unmounting a motherboard for a major cooling upgrade or computer case, sure! In my opinion unmouting a motherboard for 30mhz FSB is a stupid waste of time.

CypHill said:
Omg.. seriously stop arguing something you know nothing about. I hate to be rude but please.. you're sounding ridiculous.

This material is a copper based paint... when it's applied and cured it's like whiteout texture almost. You can easily scratch it off lightly with a needle. If you're complaining about taking out a mobo, what are you doing on [H]ardOCP in the first place? We're here to overclock what we can.
Click to expand...
 
Ok, I'm confused now, I lowered the multiplier to 11x220 and Prime lasted 3 minutes compared to 10 with 11.5x220.
 
how quickly it creates errors doesn't tell you how stable it is. all that you need to know is its still unstable and will fail eventually, may be 10 minutes, may be 6 hours.

Bop said:
Ok, I'm confused now, I lowered the multiplier to 11x220 and Prime lasted 3 minutes compared to 10 with 11.5x220.
Click to expand...
 
CzarDestructo said:
the material is whatever you can get at a local store, i've had to use a kit in the past and the one i got at auto zone certainly wasn't a copper based material that would easily come off. it was a thick silver paste that hardened to an almost solder like material. thats what i pictured so don't get bent out of shape over me questioning your ethics.

how am i sounding ridiculous?

unmounting a motherboard for a major cooling upgrade or computer case, sure! In my opinion unmouting a motherboard for 30mhz FSB is a stupid waste of time.
Click to expand...

So is volt-modding a mobo or a vid-card to get an extra 15-30mhz ... but [H]ardOCP'ers do these things to OVERCLOCK to their satisfaction. If you don't understand that then fine.. it might not be worth YOUR time.. but its different for everyone.
 
CypHill said:
So is volt-modding a mobo or a vid-card to get an extra 15-30mhz ... but [H]ardOCP'ers do these things to OVERCLOCK to their satisfaction. If you don't understand that then fine.. it might not be worth YOUR time.. but its different for everyone.
Click to expand...

i'm sure you modded it with higher expectations. i'm also fairly sure our ram uses the same chips. i'm sure you already tried this but jack your voltage to 3 and increase the timings to cas 3 4-4-7. it should run 500mhz no problem.
 
CzarDestructo said:
i'm sure you modded it with higher expectations. i'm also fairly sure our ram uses the same chips. i'm sure you already tried this but jack your voltage to 3 and increase the timings to cas 3 4-4-7. it should run 500mhz no problem.
Click to expand...

Everyone mods with high expectations. Whatever gain I get is considered a success. I'm not worried about what my memory did.. The fact was I tried everything there was but it's that the NF7-S was limiting me. That's okay. I know I should have purchased the DFI and been better off. However I sold the system. I'm getting an A64 rig soon. :)
 
This is very confusing, I don't know what is causing the instability.

It can run 12x200 (2400mhz) and not have problems in Prime, but at 11x220(2400mhz) it lasts 3 minutes. 10x220 shows no problems so far. I don't know whether to blame the RAM or the CPU.

Its at 50C load with 10x220 and 32C case temp, maybe its just a CPU overheat issue and the SLK900A and Zalaman 92mm 66CFM fan I purchased will fix it but they don't arrive until the 3rd.

I'm running 10.5x220 with Prime for an hour and if that doesn't work I'm running at 10x220 until the cooler gets in.
 
LOL!

if it doesn't fail with stock ram speed,but does with OCed ram speed,then there is your problem!
 
The memory is running the same speed at 10x220 and 11x220, thats where the confusion comes from.

The multiplier only affects the CPU, or at least thats what I think.
 
your only testing it for a few minutes at a time. thats NOT stable.

for it to be considered stable you have to run prime for 24 hours at least,with no errors. just lower your multi, make it really,really low. it will save you much time. find a safe ram speed that seems safe and test it for stability overnight.

then do the same for the cpu with the ram at stock speeds.

then mix the results for the best combo of high ram speeds and cpu speeds.

its REALLY not that hard...
 
Before I moved my memory slots from 1&3 to 2&3 it only lasted 3 hours on 11.5x205, but after switching slots I expect my memory to be much more stable.

I'm thinking of just getting the Mushkin PC3500 and have a guaranteed 216mhz FSB(unless the motherboard has reached its limit) and then adjust the multiplier from there.

I always seem to get around a 1.76-1.78 vcore when I set my voltage to 1.825-1.875, my PSU must suck or something...

At 1.825v setting in the BIOS my vcore goes from 1.74 to 1.79.
 
you won't ever notice a difference between 400fsb and 433fsb unless the latencies are hugely different.

and switching the ram around only made it operate in single channel mode, that shouldn't effect it, much.

most socket a motherboards have a tendency to under-volt a bit, perfectly normal.

your motherboard could very well be the problem of your FSB woes, but it might not be. i'd just setting with what you got and stop dwelling on it.
 
I am still in dual channel mode, if 433 and 400mhz aren't much of a difference then I guess I could settle for lower FSB and a higher multiplier. I always thought that FSB gave the biggest increase in peformance though.

UPDATE: Apparently I'm not the only one with FSB OCing issues with the Mobile Barton. Over at the ABIT forums I read this thread: http://forum.abit-usa.com/showthread.php?s=&threadid=58307

Someone had a Tbred that could do 220x10.5, he took that out and put the mobile in the same computer and he got a error in Prime after minutes. Some people are getting 208mhz FSB max with the mobile. A few tried the L12 mod with the mobile and got no increase in FSB.
 
FSB in most cases used to be associated with raising the CPUs core frequency due to locked mutlipliers.with a mobile barton you can set the mutli and core speed to whatever you want.

roughtly, depending on your mobo, ram,etc, i estimate 100 extra mhz of FSB is only like adding 200mhz to the core IF you keep the same latencies. it gives the core's cache more bandwidth to play with. when your raising the ram latencies to accomedate the higher frequency your largely canceling out your gains.
 
Well at least I know I'm not the only one with FSB OCing issues with a mobile. I'm running it at 11x215 with Prime for an hour and then I'm going to try 11.5x or 12x. I'm letting it run overnight after that.(EDIT: err I'm hitting 50C load at 11x215, I'm stopping here until I get my new cooler)

I guess as long as I hit 2.4ghz I'll be happy, or do what KillaBee did and put it at 200mhz and ramp it way up.

People have been saying that if you give the CPu a little too much voltage with the speed it will fail but I think that is BS, you can read about it on that ABIT forums thread I posted.
 
I ran memtest overnight(12 hours) at 215mhz, 2-3-2-11 timings and I got zero errors. I set it to do ALL test too. Maybe I do have BH-5 and I can tighten my timings ever more? I'll try it again tonight with 2-2-2-11.

Looks like its my CPU/motherboard/PSU. I'm gettining my SLK900A soon so if that doesn't work than I have reached the maximum OC my CPU/motherboard can handle.

I noticed that my vcore voltage rail dips way low when under load. It hovers around 1.8-1.82 when idle but dips to 1.76-1.776 when under load. Could a dip like that cause a game to crash at higher speeds? I was playing UT and ran into a BSOD after a few minutes at 11.5x210(2.4ghz). Some article reccomeneded at least 1.775v for that speed and mine dips to 1.76.
 
Bop said:
I think that 430w may not be enough for my CPU. I set my voltage to 1.875v int he BIOS and in the EQ it shows it at 1.82v while the system is idle and it goes down to 1.78v during Prime95.

My CPU is also reaching 54C and I expect it to hit 55C soon. I think I need to get a SLK900A now, or maybe even a SP97. Are those worth the extra installation time?
Click to expand...
Bop said:
This is very confusing, I don't know what is causing the instability.

It can run 12x200 (2400mhz) and not have problems in Prime, but at 11x220(2400mhz) it lasts 3 minutes. 10x220 shows no problems so far. I don't know whether to blame the RAM or the CPU.

Its at 50C load with 10x220 and 32C case temp, maybe its just a CPU overheat issue and the SLK900A and Zalaman 92mm 66CFM fan I purchased will fix it but they don't arrive until the 3rd.

I'm running 10.5x220 with Prime for an hour and if that doesn't work I'm running at 10x220 until the cooler gets in.
Click to expand...
AGGGGGGGGGGGGGGGGGGGGGGGGGGGGGH! I see you've finally mem tested. But you're not sure what is causing the instability because you rely solely on prime to discover 'stability.' Prime95 is only good for total system stability. If it fails it doesn't tell you anything. It should be the last stability tester you use. Memtest86 should be the FIRST at higher FSBs than stock. Memtest overnight and see what happens, you can play with timings and/or speeds (bump them up) and try again. This is REQUIRED. After you have found your memory's sweetspot you can also rely on your motherboard being capable of running those speeds.

THEN you find the chip's sweetspot. I would recommend doing this at an underclocked/stock ratio to your memory. This will give you a lot less headache-action. It sounds tedious, and it IS, but this method will save you time, as i and CzarDestructo have been repeatedly telling you. I'm not trying to patronize you or anything, but you could make things a lot easier on yourself...

As far as your PSU, you have NO CLUE what the voltage drop is... The hardware monitor isn't even accurate for 'ballpark' estimates, it could be TOTALLY off. You need to find a voltmod guide so you can find a measuring point for vcore and then use a multimeter to measure it. (Black to ground [probably in a molex plug] and then red to measuring point). If it's an Antec True430 (i'm assuming it is, since i haven't seen (m)any other power supplies at a 430 rating... it's probably a pretty good PSU. If you measure it with a multimeter under load (which you haven't specified if you've done yet) i'm sure you'd be pretty surprised how stable it is under load. My board will report 11.6-8 by motherboard sensor (MBM/Abit EQ reads off the same sensor [as long as you set it up correctly]), but my multimeter will never drop below 11.95 (and fluctuation is VERY little, as opposed to the .2v flucts of the MBM sensor).

For reals ;P
 
What I just don't get is that 12x200(2400mhz) runs fine but something like 10.5x220(2300mhz) crashes in like 2 seconds. And my memory is fine so that is taken out of the equation. And I guess I can trust my PSU is up to the task, so its either the motheboard or the CPU. Over at the ABIT forums alot of people seem to have the same problem with the mobile XPs. Does a higher FSB create more heat than a higher multiplier or what?

EDIT: I was wondering, would increasing the chipset voltage help with the issue? Maybe it can't handle a higher FSB with a high multiplier, but I don't know if 1.7v would fry the NB or not, but I do have the ABIT HSF on it.
 
Bop said:
What I just don't get is that 12x200(2400mhz) runs fine but something like 10.5x220(2300mhz) crashes in like 2 seconds. And my memory is fine so that is taken out of the equation. And I guess I can trust my PSU is up to the task, so its either the motheboard or the CPU. Over at the ABIT forums alot of people seem to have the same problem with the mobile XPs. Does a higher FSB create more heat than a higher multiplier or what?

EDIT: I was wondering, would increasing the chipset voltage help with the issue? Maybe it can't handle a higher FSB with a high multiplier, but I don't know if 1.7v would fry the NB or not, but I do have the ABIT HSF on it.
Click to expand...
A higher FSB would stress the board more, considering it's running out-of-spec. It's not like a board will be unable to run it, it's just that it's not 'officially (read guaranteed)' supported. It is pretty odd that you can't run a measly 220fsb though. And now that you mention your low VCC voltage, i'd bump that up. The Abit HSF will be able to handle that pretty easily, just keep an eye on the temps (or touch the HSF with your finger... if it burns, you probably want to upgrade the cooling). Chipset voltage is pretty big in stabilizing a high(er) overclock. 1.7 isn't too much for a NB by any means.

Sometimes on a motherboard, the VAGP is closely linked to the VCC, so if that doesn't work, you might bump your vAGP up .05v to see if that helps. It won't give you any gains on your video card, and will pretty much guaranteed not do it any harm, since modern voltage regulation is handled on the PCB of the card.
 
I think I got it stable at 11.5x215mhz. With 2-2-2-11 timings it fails in a minute but at 2-3-2-11 its still going on with an hour and 30 mins.

I might give it a rest here if it is stable, I'm probably only going to get 1-2fps increase in games if I go from 11.5x215 to 12x215. When I get my SLK900A in I'll give it a go but if it doesn't go too well I won't lose much performance.

I'll give the 10.5x220 one more go after I bump the chipset voltage and the CPU voltage, but if that doesn't work I'll have to settle for less and face the fact that my motherboard/cpu just can't handle it for some odd reason.

UPDATE: I tried 10.5x220 and bumped the CPU voltage to 1.9v in the BIOS menu and the chipset voltage to 1.7v. Things got HOT, the case temp was 37C and the CPU was starting to get past 55C. It lasted more than a minute though... it got to 3 mins and then I stopped it as it was getting too hot.

Does a CPU with a lower multiplier and higher FSB require more CPU voltage than a CPU with a higher multiplier and lower FSB? It seems like the case with my PC.

UPDATE: I have it at 11.5x215, 1.775 vcore and 1.6v chipset. I think I'm going to keep it like this, any more mhz won't give me alot of framerate it games. I get a case temp of 34C and a CPU temp of 51C.

I just have to hope it passes an overnight Prime session, if it doesn't I gotta bump the CPU voltage, and then the thing gets too hot...
 
Bop said:
I think I got it stable at 11.5x215mhz. With 2-2-2-11 timings it fails in a minute but at 2-3-2-11 its still going on with an hour and 30 mins.

I might give it a rest here if it is stable, I'm probably only going to get 1-2fps increase in games if I go from 11.5x215 to 12x215. When I get my SLK900A in I'll give it a go but if it doesn't go too well I won't lose much performance.

I'll give the 10.5x220 one more go after I bump the chipset voltage and the CPU voltage, but if that doesn't work I'll have to settle for less and face the fact that my motherboard/cpu just can't handle it for some odd reason.

UPDATE: I tried 10.5x220 and bumped the CPU voltage to 1.9v in the BIOS menu and the chipset voltage to 1.7v. Things got HOT, the case temp was 37C and the CPU was starting to get past 55C. It lasted more than a minute though... it got to 3 mins and then I stopped it as it was getting too hot.

Does a CPU with a lower multiplier and higher FSB require more CPU voltage than a CPU with a higher multiplier and lower FSB? It seems like the case with my PC.
Click to expand...
I have no definitive evidence or anything, but in my experience, at the same frequencies you will not need a voltage bump (for the processor at least). The motherboard might need a good bump for the vCC. Leave the VCC at 1.7, it's not too dangerous, as the poster above my last response stated, he ran it without the fan on the cooler safely for an extended period of time. I'd try running the same voltage but keep your northbridge bumped.
 
Ok, but dang my games are giving me more trouble than Prime!!!

BF Nam crashes to the desktop after 5-10 mins, Prime lasts longer!!! My CPU temp was only 47C and my case temp was 31C at the time of the crash.

I'm wondering if it is a software or hardware problem.

UPDATE: Ugh Prime crashed after 15 minutes on 11.5x217mhz(2497mhz). I had the chipset voltage at 1.7 and the CPU voltage at 1.8v set in the BIOS. I bumped the voltage up another .05 on the CPU, maybe that will take care of it.

UPDATE 2: Ok the CPU has been runnign Prime for over 30 minutes so far after I bumped the voltage to 1.85, so my limiting factor is voltage here. If it fails an overnight Prime I'm taking it down to 2.3-2.4ghz while I wait for my SLK900A then I'm bumping it to 2.5 again when I can run 1.9v at < 55C.

UPDATE 3: Damn! I got a BSOD during Prime that says PFN_LIST_CORRUPT. I'm taking the memory down to 215mhz. I'm also thinking about lowering my memory voltage, 2.9v seems too high.
 
You must log in or register to reply here.
Back
Top