PDA

View Full Version : Greedy 2500+


Ageeb
10-18-03, 08:01 PM
Apparently my 2500 doesn't like any combo in the 2.3Ghz+ range unless I pipe a line from excel to it and burry it somewhere in the arctic. it's fine right now at 215x10.5@1.85v, but if I go up at all with multi or fsb or any combo thereof (leading to a greater Mhz) I need like 2.x volts for stability... :( . Anyone have any experiences with this stepping they could share? Should be in sig.

c627627
10-18-03, 08:12 PM
Top notch mobo, cooling, top notch PSU, nop probs there LxOxSxI

Your chip does 2300 as is the expected average.

Yet another example that luck plays a role in above average OCs.
Another chip may do better but that's it for this one.

Ageeb
10-18-03, 08:21 PM
It's not what I wanted to hear, but I feal good anyhow... thanks all. Oh, memory is holding onto 5-2-2-2 so far (bout 30-45 minutes into testing)

c627627
10-18-03, 08:29 PM
PC2100 DDR 266 133 MHz
PC2400 DDR 300 150 MHz
PC2700 DDR 333 166 MHz
PC3000 DDR 366 183 MHz
PC3200 DDR 400 200 MHz

Maybe you can't do more FSB because you have RAM that's rated to do 183 FSB and can't be OC's too far into the 200s.

They say this one hits 220 almost every time:
http://www.memoryx.net/gn256pc3200.html

Ageeb
10-18-03, 08:43 PM
I was testing my cpu with real loose timings. 11-5-5-2.5 (won't boot with CAS3). So memory shouldn't have been the issue right?

EDIT:
typo

siavash_s_s
10-18-03, 08:54 PM
i have a 2200 tbred b and msi kt4 ultra and generic 333 ram
according to my indicators im failing at memory intialization when i try over 150 fsb even with loose ****ty timings.my conclusion: better ram is better

tom10167
10-18-03, 09:09 PM
Loose timings will not make up for ram speed, they're different things.

Ageeb
10-18-03, 10:57 PM
You're missing the pt. I used loose timings to find my cpu's limits. After that limit was reached (the first message in this thread) the timings were tightened up to 6-2-2-2 thus far. Not sure if sig is updated yet...

c627627
10-18-03, 11:08 PM
Can you do higher multi with lower FSB? That way you can maybe rule out RAM as a weak point...

All in all I'd say it's the specific chip you have maxing out, you're one of the few people that can actually say cooling and power supply are not keeping you down for sure, they're that good.

ManOfKnight
10-18-03, 11:29 PM
I agree with the message above me. It could still be your ram. Increase your multi and drop your FSB. This will tell you if you need new ram. If not, then 2.3Ghz is not a bad O/C...especially with the RAM you have

Ageeb
10-19-03, 02:00 AM
I've tried alot of combo's with FSB and multi's to =2.3Ghz or better and it's essentially all the same voltage req's. That's what I thought was odd. This chip REALLY hits a brick wall. It seems it doesn't care how you arrive at it, but again, if it's 2.3+ your callin excel :). Maybe I'll try upping the FSB more and dropping the multi to try and gain some performance around the 2250 mark...

EspElement
10-19-03, 06:16 PM
i cant seem to get my 2500+ to 2.4 ... id really like to hit that tho... ive tried lots of things... well see... i ordered a 1800 and if that dont do good then i will go backt o my barton with the new psu and the new fan i ordered... 92mm tornado..

it seems i cant go above the 2.0volts without it screwing up on tests... my motherboard setting is at 1.975 but it will read at 2.0v.. if i go to 2.0v on motherb oard setting it hits more and just never passes one test...

well see... i can boot at 2.4... thats it.. :-\

Ageeb
10-19-03, 06:24 PM
I can do lots 'o things but pass prime at 2.4 :(

EspElement
10-19-03, 06:26 PM
no ****.. :-\ that sucks what anout the new steping? i heard it was varly good... and it ran varly fast... like 2.4 or somethin

Gautam
10-19-03, 06:28 PM
You've got a great system and a great overclock thus far. Its fairly common for Bartons to need over 2v to sustain above 2.3ghz. Some people get the lucky steppings, some don't and that's life. KHX3000 have either excellent BH6 chips, or legendary BH5 ones. With 3v or so, you should be into the 230 domain with tight timings.

Since your power supply is so full of juice, you could try running some ultra high voltages(2.2v+) just for the fun of it and see where you end up. Due to the lack of information on how long these processors last with high voltages, I'd avoid over 1.8v in normal day usage, however.

EspElement
10-19-03, 07:00 PM
i run a daily usage of... 212 x 11 = around 2.32 with a volt of 1.925 if i remember correctly...

right now... im just running at 2800 speeds for the fun of it... cause my other memory in the other cmoputer im gonna put this chip in... is 333 memory... and im not gonna overclock that memory

i run this stable... 2.09 at like 1.750 or somethin volts.. who the fu ck knows lol

the other system will haev 1 gb of ram... and its otehrs i used in the past... thats why its only 333 ... im probably gonna use it as a small game / recording machine

Ageeb
10-19-03, 09:01 PM
I'll try that.. I suppose my temps will start to overwhelm air cooling at that voltage though.. we'll see I guess. Hey, one more thing. I have a difference of around 4-5C between my probe on core and the mobo diode temps. I'd assume probe is more accurate due to it's contact, correct?

LxOxSxI

Twister1234
10-20-03, 07:57 AM
hm,how do you know,how high a cpu might go?
mine is @1,725 vcore,and i just want to know,what i can trie,when my new heatsink arrives---its prime95 stable,with these settings.
AQXEA

Ageeb
10-20-03, 05:37 PM
Like you've done, you raise your FSB(usually to the number where you memory maxes). Find the multi that equals your cpu's original Mhz. ex: 2500+ with FSB of 200 (where memory maxed out possibly)=multi of 9(roughly) and stability-test. Should pass.

then increase your multi 1 step and stability-test it. If it fails try again with a touch more voltage. If fail again, repeat voltage process.

Eventually, there will be a point where you can't seem to give it enough voltage. Kinda like the first message in this thread.

Someone please correct me or add to this if I am wrong vague. I'm just a greenhorn myself :)

LxOxSxI

EspElement
10-20-03, 05:59 PM
well i did it differant... kinda got myself too 200... with 10 multi... then went to 192 with 10.5 multi... up to 200 x 10.5 multi... 192 with 11 so on... and up the voltage as needed... but its the same thing... only it pushes more out of it... u can find ur total max

eventually... go to like 210 with 11 or somethin

Twister1234
10-20-03, 10:24 PM
hm,wouldnt it be easier to raise the multi,instead of the fsb...like 12x200?

EspElement
10-20-03, 10:38 PM
what u mean? well the higher fsb... the faster ur overall system goes... so u want to get the fastest fsb

OC Detective
10-20-03, 11:34 PM
Well if you wana find the limit of your ram or maybe your mobo best way is to lower the multi to something real low such as 6 and then up the FSB until it freaks out then you know the ram or mobo limits. From there you can do the reverse (low FSB and high multi) to find the cpu limits - and based on this info find a good balance of FSB/multi for your system.

ZaP37
10-21-03, 12:06 AM
My Barton behaves in the same way as yours, LxOxSxI. I run 2328MHz completely stable, but if I try 2400MHz I bluescreen as soon as Windows starts to load...even with a vcore of 2.2v...

What is your processor's stepping? See my stepping in my sig...

Twister1234
10-21-03, 04:58 AM
ok,i tried a little bit...
pc will start,prime stable,with fsb 215 and a 6/5 ratio(my mem causes problems running over 420 MHz),if i try 220 fsb,it looks,before posting...would it be better,to lower fsb,to get my ram working at maybe 205 fsb?

so,if i understand right,this seems to be my max fsb,and no i can try to rise the multi?

edit: running at 215x11,needed to raise vcore to 1,85,temps are stil ok(idle 39 C,load 47 C) 2369 MHz,think i will not try more,cause my southbridge isnt cooled yet,and my northbridge is standard cooled...

what can i exept of this cpu?will 2,4 MHz be possible,with air cooling?(slk 900 a,with an enermax UC-9FAB-B,cfm about 57,2600 rpm,92 mm)

new edit: :p i can write a guide about that...if i disable cpu interface,i will boot,with my mem@215 fsb...but it wont be stable,running 2,8 volt,and i dont wont,to get a reason to by new mem^^

so,i lowered fsb to 210,everything works fine,and my pc is stable@2317 MHz(hm,dont know why mbm5 sais that,should be 2310...),with mem@420 MHz and vcore 1,825...

i tried a 11,5 multi,but it causes bluescreens@1,875 vcore...

ogboot
10-21-03, 01:03 PM
my aqxda in my sig was stable @ 11.5*190 1.85vcore, but once i hit 200 fsb the voltage requirement has really gone up, don't remember whats in my sig right now 209*10.5 1.9vcore 'bout to put it up to 2.0 and see what it will give me

edit:meant 11.5 instead of 11

Ageeb
10-21-03, 07:57 PM
ZaP37, AQXEA 0330VPMW


LxOxSxI

EDIT:

ZaP37, just curious as to what you define stable as... I suppose you primed it, but for how long did it pass at that?

EspElement
10-21-03, 08:45 PM
at 3.65 i could prime it all night long... but that was the highest ever done... and i cant seem to get that anymore... dont know why :-\