• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Whats goin on here?...Hard to OC 2500+

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

PoX Freak

Member
Joined
Jun 14, 2003
Location
North Carolina
Hi guys (and gals, if any);
Here's the problem.
I have an athlon xp 2500+, DKV4D, AQEXA 0330SPMW,
This puppy does'nt want to go past 2250Mhz on water without the temps going nuts.
@ 2500+ speed, 38ºC idle, 39.5ºC load.
@ 2800+ speed, 39.5ºC idle, 41ºC load.
@ 3000+ speed, 41ºC idle, 43.5ºC load.
@ 3200+ speed, 42.5ºC idle, 44ºC load.
I should note that in order to hit these speeds, the Vcore has to be pushed up progressively, in order, from 1.65 stock, to 1.75.
Now, my last cpu was a jiuhb t-bred 1700+, and on water that ran as fast, if not faster than this barton now is, and was totally stable.
Have i got a bad on-die sensor that reads 5º lower, or maybe a bad CPU, or what?
What also do you think of the temps, since after testing my block, which is pictured in the avatar, C/W is rated at 0.26, and the temps shouldn't be this high, i think.
BTW, my son cooked the t-bred for me when i was in the middle of a case swap.
The t-bred had been clocked to 2353Mhz, with a 1.975 Vcore.
 
Last edited:
What was the vcore on the XP1700 if it was 1.75 then remember the default may have only been 1.5 and if not that 1.6V so I would suggest upping the vcore on the Barton a bit more - dont worry it can take it!
 
Hrm that temp at stock speed seems a bit high. Not alarmingly high, and maybe its just not accurate, but I would expect lower on water. Sorry I am too lazy to calculate the actual speeds for the rest of the stuff. Most people typically talk in terms of real clock speed and not AMD's silly PR speeds ;)

More voltage would help I think, but if those temps are accurate things could get pretty hot with more juice...
 
i got same mobo and cpu.my core is set to 1.8 and oc'ed to
xp3200(11x200=2200mhz).my temp for 3200+ is about 44C with no a/c while folding.summer temps no a/c was about 50C while folding(at work).with my window a/c on folding temp was at 40C.
i'm aslo watercooling.
 
Good lord, you are going 2.2v on air? That thing is going to go poof after about 6 months of that. I wouldn't trust a mobo reading at that voltage either.
 
motherbaords are actually good at reading voltages.
I have never seen one (quality brand that is) be off in its readings by more than .01V. Dig out your multimetre and have a go at it. you will be supprised how accurate it is....

Now if they could only get temps right i would be ALOT happier
 
sorry, Deathknight. By putting down the AMD PR ratings, it made it easier to show current volts and temps.
BTW, on air, i coast along at 43ºc at full load, stock.
Maybe i'm not doing a barton right by just clocking it up a little.
probably because i remember the horrid smell a cpu makes when it cooks.
I've already burned 3 cpus, and i dont plan on roasting this one anytime soon.
I have yet to send in the second waterblock, pictured in the avatar, to test it for efficiency and C/W, but as stated, the estimated C/W is about .26 to.28, and my eheim 1048 pump is almost level with everything else in the cooling rig, including the rad.
 
2.2v! Holy crap!

about the 2500+, doesn't seem to be such a bad overclock. I calculated the average 2500+ overclock at 1.75v in the CPU database, and the average is 2270Mhz, so you are only 20Mhz lower.
 
I just received my 2500+ barton today and it is currently set at 1.675 (bios=asus probe showing 1.728 though), running 11 x 200 on my asus a7n8x deluxe (rev. 1.04) with corsair xms pc 3200 memory running async. I havn't tried going higher yet as things are running stable and smooth at the moment. My temps are 38c idle, 40c on internet and running programs, and 42c running 3d mark 03 with an oc 9500 red pcb at 350\295 for 5326 3d marks. This is with the stock fan and heatsink that came with the processor.:D
 
UPDATE:
2500+ is now running prime95 stable, 12 Hrs, @2.424 Ghz, 1.90 Vcore.
Waterblock (bong-block) has been finalized, and C/W tests will be made, as soon as i can get another built. Hard to get an estimate, since EPoX boards have this freaky problem reporting system temps to sandra, but MBM and USDM report fine.
 
same here something worng 2.5 with 1.98 volt's?

11.5x218 100% stable
 
Just looked at your first post, from the CPU frequency vs temperature numbers, it said that the CPU had not reached overclocking limit.

2800+ to 3000+, MHz increase of 84 MHz (from rating), load temperature increase 2.5 C
3000+ to 3200+, MHz increase of 84 MHz (from rating), load temperature increase 0.5 C

average MHz / C = 168 / 3 C = 56 MHz / C (from your measurement)

Since you had multiple temperature data points, even the temp probe was off by 5-10 C (in absolute number), but the relative difference could still be used to measure overclocking.

I estimated for Tbred B/Barton, the CPU reaches diminishing return of overclocking frequency at or below 10 MHz / C. It had not reached limit at that point (first post).

Now your last post says it goes much higher from 2.2 GHz to 2.4 GHz, the above analysis kind of confirming that. When a CPU reaches 10 MHz / C and below, overclocking will slow down more and more.

E.g. When Vcore is increased by 25 mV, get a temperature increase of 2 C and frequency increase of only 20 MHz, that is 10 MHz / C. I consider that as point of diminishing return, and overclocking will slow down. 3 MHz / C is the practical limit.

For details:
Some numbers to determine max CPU overclocking frequency - Vcore vs temperature,
When do the CPU's slow down?
(page 13)
Explanation (page 13)
 
Last edited:
Back