Jump to content

Core 0 is the black sheep of the family.


Recommended Posts

I know, I should be happy with my X2 4400+, but I'm not. It just maxes out at 2400 Mhz, everything else isn't stable and for 2400 Mhz I already need 1.45v. And it's just core 0 that is the culprit. 2500, 2600... all fine with core 1. 2450 and core 0 goes mental in Prime95, even with 1.56v.

 

I changed my Zalman Cu-7700 for my shiny new Koolance Exos-2 today (see sig), hoping that I may get that cpu stable at 2600. But no, it wasn't the temperature at all, the Koolance lowers my temps about 10C with no noise at all, so I'm happy with that, but my overclocking results suck.

 

Yeah, I've got the equivalent of a 4800+, maybe a bit more with memory running at 1:1 @ 240 Mhz, but... oh well, I just have to force myself to be satisfied with that. Hard for someone who started overclocking in the early 90s by soldering 50 Mhz clock generators (instead of 40 Mhz ones) on Amiga 1200 Turbocards :)

 

/rant over.

Share this post


Link to post
Share on other sites

  • Replies 38
  • Created
  • Last Reply

Top Posters In This Topic

some may say taht the x2 is the hardest to tame. especially both the cores aren't the same u you would have to tune both cores in an ordily fashion... i suggest u read around and ask for help. ocforums.net would be a nice spot to visit as for here i haven't seen much of users with x2... a lot with fx's tho

Share this post


Link to post
Share on other sites

Thanks for that page, I'll have a look. I knew that this could happen before I bought it and my system is still very smooth and fast and so much better than my 3500+ @ 2.5 Ghz. I'm just greedy :nod: :angel:

 

Edit: ocforums.net doesn't exist. it's .com :P

 

Ok, not even 1.6v can convince Core 0 to run at 2450 Mhz. Nothing I can do. But 1.44 is enough for 2400 Mhz. :rolleyes:

Share this post


Link to post
Share on other sites

My black sheep is core 1. I just got her last night so still playing around with it. Currently at 2539MHZ @ 1.36V She ran all the way up to 2475 with 1.3V the core 1 would fail 20 minutes in prime. Currently at 30 minute with the new voltage and speed :nod: . All I want is 2.6 and I will . my pants. But is has to be 12+ hrs prime stable and 3D stable. Ultimately though If I can get 2.5 stable I should be happy and I am sure I can squeezt that much out of her.

 

Did you try running a divider to try and rule out the Ram or timings ;)

Share this post


Link to post
Share on other sites

That is pretty disappointing for X2 overclock yeah. Mine goes to 2400 np on stock voltage and settings, 2650 with 1.4v isn't a problem with memory divider, however, the chipset temp seems to get pretty high at that rate, and the chipset fan on max is *loud*.

 

Turned it back down to 2400 for now because I couldn't stand the noise 8)

Share this post


Link to post
Share on other sites

I tried mem dividers and slower speeds, of course, but that shouldn't be a problem with my G.Skill LA 4800 anyway. One thing I will try on the weekend is to remove 2 of the 4 sticks and see if it's the core itself or it's memory controller... my old Winnie liked it up to 10x250 with quite fast settings.

Share this post


Link to post
Share on other sites

however, the chipset temp seems to get pretty high at that rate, and the chipset fan on max is *loud*.

 

Turned it back down to 2400 for now because I couldn't stand the noise 8)

 

Did you try to change the BIOS settings to "Full Speed only >60?" :)

Works wonders.

Share this post


Link to post
Share on other sites

Well... there was one last, brutal, suicidal option, wasn't there? I mean... risking the life of the CPU by doing that:

 

decap.jpg

 

It worked fine for Pentium 4s, so I thought, what the heck, it's just £440... :sweat: and I can't afford a new one and my old CPU is sold already, so I was a BIT scared, I admit.

 

But after going round the IHS with that razorblade, it came of easily. And that grey stuff between the CPU and the IHS is hard and dry and crumbles easily and I thought, well, if the CPU did survive that procedure, maybe it'll be cooler and needs less voltage. The core is pretty large compared to e.g. an Athlon XP, so contact between the waterblock and the core should be good.

 

Koolance adds 2 screws for the waterblock, one for CPUs with IHS, one (a little bit longer) for CPUs without. Lucky me, no modifying, just use the other screw.

 

Machine boots, I check the bios and the CPU at 2.4 Ghz with 1.5v is at 38C. Hooray :) Then I go into Windows... CPU-Z, Prime95-1, Prime95-2, 2.4 Ghz, 1.5v... maxes out at 48C.

 

The big moment: ClockGen. Change FSB from 240 to 245. Before, this instantly made core 0 crash in Prime95, within milliseconds. Now... nothing, calculates away. SUCCESS! FSB to 250 Mhz -> 2.5 Ghz. No error and still running fine. In fact, I am writing this now while both cores are still in Prime95 torture test! 51C core temp.

 

Woohoo! It's not core 0! It was the damned heatspreader stopping me! I am happy with 2.5 Ghz, but who knows what's possible now. Maybe less voltage, maybe 2.6 Ghz? To be continued...

 

Disclaimer: I'm not responsible for anything YOU want to do or will do because of my post. DON'T DO THIS AT HOME! Better even, DON'T DO IT AT ALL.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×
×
  • Create New...