OC my 2600XP ???

[Soulhunter]

Ok, Ive replaced my old cpu-fan with a new one...

The temperature dropped from 64°C to 49°C now !!!


So, I thought its time for some overclocking... :D :D :D

Someone knows a good (safe) initial FSB speed to start with ???

My processor is a AMD Athlon XP 2600 (Barton) / Core @ 1916.5 MHz / FSB @ 166.7 MHz


Tia n' Bye

[TotalChaos]

Wow soulhunter 64 degrees celsius!!!! You are pushing the limits there if you asked me....(guess you didn't ask me though hehehe) I havn't looked up the AthlonXP's max temp, but I know that most P4's max is around 70 celsius.

I believe you know what your doing as I've seen you all over this forum. You've been at this awhile. Just thought I'd boost my post counter by 1 and put in my two cents.

[Soulhunter]

Yeah, 64°C was with the boxed cooler and no case fans... ;)

But it was running perfectly without any heat related crashes !!!

So, with the new cooler now, I wanna raise the FSB to push it back to 60°C... :D


Bye

[jimmy basushi]

my old athlon runs completely stably at 70 degrees (not kidding), ive done 24+ hours of straight encoding with no problems at all.

but onto your subject soulhunter, why is your computer running so hot? my brother-in-law has a 2500xp oc'd to 3000 completely stable and it runs at about 53-58 degrees at stress with the stock cooler. its not like australia here is a cold country either...

i cant specify the exacts of his oc'ing but, i hope someone else can shed light on that one

[Soulhunter]

I assume it was running so hot because I had no case fans !?!

No matter, now I run it with a FSB of 190MHz @ 51°C... :eek:

Wow, only 51°C while running @ 3000XP speeds !!!


Maybe I could even run it @ 3200XP speeds...

Nah, first I gonna make a 8h UT'04 stability test to be sure !!!


Bye

[Mnl]

Or you could use prime95 (http://www.mersenne.org/freesoft.htm) to test stability. That way you can use your PC for other things while stressing it, plus Prime95 is better at making the CPU develop heat than UT'04, I think :)

[Soulhunter]

I use UnrealTournament 2004 to test the overall stability (CPU, RAM, GFX) !!!

To test the CPU my usually XviD encoding should be enough... :D

But thanks, I will add prime95 to my tests as well !!!


Bye

[theReal]

A good FSB to start overclocking would be 167 ;)

But hey, no kidding, it's the safest way to check out what you can do (OK, maybe don't start with 167 but with 175 or something like that). One more MHz and then check with Prime torture test for at least a few hours.

btw. temp displays can vary a lot from board to board (and from diode to diode). Up to 15°C difference measured from two different diodes could still be the same actual temp...

I have watercooling with my Athlon XP2700+ (TBred Core). I'm running it at default voltage and a minimalistic OC of 268 FSB - I get a maximum temp of 42°C after hours of Prime/avisynth (room temp at max. ~26°C - it hasn't been any warmer since I got the watercooling)
In my experience there are occasional instabilities with avisynth/Prime as soon as the CPU temp hits 50-52°C or above.

As I don't want to risk any instabilities whatsoever, I got an even better waterblock for my CPU now, I also want to add the Northbridge and GPU to the watercooling system. Better cooling never hurts ;)

[Soulhunter]

Ok, after 4 hours of UT'04 it hits 55°C @ FSB 190MHz !!!

And after 4 hours of XviD encoding it hits even 56°C !!!

But no crashes or errors so far... :)


Bye

[guada]

Hello "Soul"

Say me, how much have you won in frequency in relation to your initial value.
For my part, I limited myself to 700Mghz (small info: I kept the fan of origin and I didn't add anything on my configuration of basis)

My P4 2.4ghz overclocked to 3.1ghz (no problem until now)
Be kind with your new "baby" :D :D :D

[Soulhunter]

Before it was @ 1916 MHz (2600XP), and now its @ 2185 MHz (3000XP) !!!


Bye

[Asmodian]

@Soulhunter

I have noticed that UT2004 and 2003 are both very forgiving of an overclocked system - I have overclocked two computers recently (both multiplier locked XP2500s on nforce2 motherboards) that would never crash UT but had errors in prime95 after only a few hours (or minutes depending on particular settings). These might have been chipset/ram problems because I was >200Mhz on the FSB but it made me less trusting of stability tests using UT.

[Soulhunter]

I should mention that Ive also oc'ed my GFX... :D

So, I ran UT'04 and 3D Mark to test its stability as well !!!

But I also made tests with XviD and MemTest to test my CPU/RAM stability... ;)


Bye

[guada]

Yes, i think is a good idea ;)

[theReal]

I found that the Sisoft Sandra Burn-In test for cache speed and memory speed are really hard on your hardware.

I had Prime Torture Test run for 12 hours without any errors, then started the Sandra Memory Burn-In and got a BSOD in no time (sometimes immediately, sometimes after 15 minutes, depending on RAM timings)

Finally I decided my system was stable enough when, after a few hours of Prime, I could run the Sandra memory tests once without anything bad happening.

So far I have not encountered any reallife application that was more demanding than Prime Torture test (avisynth 2.5 together with Realproducer commandline is about the same demanding as Prime). So I guess it should be enough to do a test with Prime (for at least a few hours - I like 12 hours or more to be safe)

[Tuesday]

Just for temp comparison, I have a P4 3.0C, while not overclocked at all I'm using a zalman 7000BCu on silent mode in a case with no fans 'cept the psu and my processor regularly gets to 70-75 C during heavy use and has yet to crash even throughout 48hr Xvid encodes running in the background while I play Half life 2 :)

While hot this is still less than the 80 C maximum reccomended temperature, and it keep my system all nice and quiet :)

[Soulhunter]

Gah, casefans... >.<

I had to install some after the last GFX upgrade !!!

The old GFX blowed the hot air out @ the backside, the new one not... :(

Well, good that I have a 7.1 speaker-system to drown the noise my pc makes !!!


Bye

[Nematocyst]

There is no particular "safe" FSB to start experimenting with since each processor has its own limits. Anyway, assuming your temps are accurate, I wouldn't recommend increasing voltage. That's what will kill your temps. Just find the fastest speed you can get stable using the default voltage.

And yes, prime95 is a very good tool for testing stability. The only problem I've ever had with it is that sometimes it tries to use too much memory, causing paging which makes the test worthless. But if you go 24 hours w/out paging in prime95, then you are pretty stable.

[Sharktooth]

I have one of those "older" bartons 2600+ with unlocked multipliers and i pushed it to 2300Mhz (11.5x multiplier) with 400Mhz (200x2) FSB -> 1.65 Vcore and 2.85 Vdimm.
It's safely running 24/7 at 43 °C (actual temp) with a coler master Aero7 cooler.

EDIT: Yes, i'm a lucky bastard :D

[GrofLuigi]

Originally posted by Nematocyst
And yes, prime95 is a very good tool for testing stability. The only problem I've ever had with it is that sometimes it tries to use too much memory, causing paging which makes the test worthless. But if you go 24 hours w/out paging in prime95, then you are pretty stable.

AFAIK, it's a good thing (it's allocating/testing all of your memory) and yes, it's a pain when you do something else - but you're not supposed to do anything else. :)

I've also noticed that it doesn't stress the CPU to the max (in all three tests) - I usually run it together with another "cpu burn-in/number crunchin" program and reach 10 degrees more!

But overall, out of several similar programs for CPU testing I've tried, I haven't found better program for that purpose.

GL

[Nematocyst]

AFAIK, it's a good thing (it's allocating/testing all of your memory) and yes, it's a pain when you do something else - but you're not supposed to do anything else.

The point I was making is that it can use more than there is, causing paging. This makes the test completely worthless because the CPU is waiting for the HD, and is thus, not being stressed at all.

[Shinigami-Sama]

hmm
I'd love an AMD procesor again I used to have an old amd k6 @ 266mhz I could never hurt it once and now with htis damed p4 I;ve run into more troubles go figure eh?
anyways I;d say you go jsut a little bit higher soul a few more mhz on you FSB bus kinda scary how low it is compared to mine but if that litle one does hte same work as my bigger one..
hehe I guess it true eh?
size doesn't matter ;0

[JoeShrubbery]

Well just to add my 2 cents (and CPUs) in the mix here. My Barton XP2500+ is currently running just idling fine in the low/mid 50s maxing in the mid 60s at 2200mhz (11x200) using the stock heatsink, but I had to up the CPU core voltage a little higher than I'd like (1.825) to keep it from freezing and crashing such in short order. Meanwhile, my Sempron 2500+ runs fine at 2100mhz (10.5x200) using it's stock core voltage of 1.60 and using just the stock heatsink, and temperature readings in the mid 50s while idling as well. I was actually running that Sempron at 2310 for a month or two (10.5x220, 1.80 core voltage) but I noticed it was maxing out at 71 degrees under full load and felt that was just a few degrees more than I was comfortable running it at (the problem of moving from a cold winter to a warmer spring, by summer I might have to clock the XP down a bit more). I should look into trying to unlock the multipliers on both of those chips at some point in the future: I'd like to push my FSB & RAM a bit further than either CPU (when using the locked in multipliers) can handle reasonably.