Highest Stable OC Challenge.

I began wondering about this a few minutes ago, and realized, I rarely if ever see what people actually run thier PC's at, day in and day out.
sure theres a ton of great benchmarks and screen shots here everday, but most of it's the top of the line unstable, passing by a hair, cross your fingers and pray to god kinda OC's.

I want to see what people can do 100% stable.
everyday, under every condition.
but how can you prove somethings 100% stable?
well, I can't really guess.
but I have this idea.
just a series of benchmarks.
the most grueling torturous benchmarks we all avoid.
some may be unecessary.
some may be boring.
but each will prove one thing, atleast as best as possible, who actually does have the Highest Stable OC.

heres, the challenges.

1. Complete 32M on Super Pi - Time irrelavant
2. Complete Hexus Pi-Fast- again time irrelavant
3. Complete 12 hours of Prime 95 atleast.
4. Complete FULL 3DMark2001SE test (not just game) and provide link.
5. Complete FULL 3DMark2003 Test (again all tests) and provide link.

Theres the Challenge.
I know I cringe when I think about trying some of it myself, cause I know the trolls falling down the MC-Escher thing in 3Dmark2003 is a pain in the ass to pass when heavily overclocked, and waiting for 32M to pass on Pi-fast can give you a heart attack if your sitting there watching the clock, hoping it won't freeze up on you.
But- thats the point of a challenge, isn't it?
to be tough.

theres other benchmarks I'd like to include or see, like rendering benchmarks, and sisoft ones, quake3 benchmarks and such, but in all honesty I know those are alittle obscure, so I tried to pick the most common benchmarks I could, but ones I know can be very tough on a unstable system.

This should be a nice challenge.

ill try it later this week when i get new hardware.

ill also post pic of the tests, like superpi and stuff to prove its stable

Kunaak you should put a note on what type of tests to run with P95, i.e. running Small FTTs for testing full CPU stability, or running Blend to test RAM stability.

Cheers, great idea! :toast:

`s

mines 2.2ghz :p:

sad isnt it...

I run at 2.3ghz now... Gota try testing at 2.53.

kunaak mate you want the highest MHZ stable clock ?

wouldnt that be unfair for us with 2.8 or lower cpus compared to users with 3.4ghz chips? :D

my 2.8 is prime stable @ 4.2ghz with 1.75V....

i think i am gonna push the vcore higher for this challenge...1.85V sounds right?

think were gonna see loads of cpus dead int he next few weeks lol

should we make more categories on this?

highes % stable overclock etc?:D

1.85V is to dangerous for my tastes :p:

And I think the point is stable 100% 24/7 and using the benchmarks for proof. Not just stable through the tests. Like gaming rig stable.

hmm, lots of good ideas here.
how about this.

Prime 95-

enter the password 9876 (perfectly legal, comes with the program in a little text file, incase you wanna push your CPU even farther)

click on "round off checking"

then "torture test"
thats always been my way of running Prime.
if someone has a better idea.
put it on here.

brwmogazos

I like your CPU % idea.
so perhaps that should be apart of it, but I am not sure how.
what about those guys with say 3.4C's and running 4.5 ghz everyday?

indeed a powerful machine, but not high in %....

this challenge is just for fun, nothing to freak out about.
if your PC can't runn say 3.6 ghz today, then don't try the benchmark at 3.6 tommorow.
this is just to show STABLE OC's.

how about seperating it to Highest CPU type?
P4 against P4.
Xp against Xp.
A64 against A64 and so on...

why not make it the highest percentage overclock that is stable? so that way it doesnt matter what cpu you have?

That sounds good to me. YES the cpu types need to be separated.

well, I guess I don't want this post to end up being a showcase for one persons machine.

CPU against CPU lets just about anyone post in here.
thats kinda what I personally would like to see, lots of OC's. Air, water, phase-whatever.

whats round off checking do?

mmm yup any cpu and than just calculate the overclock percentage:

OC % = (OC - Stock)/Stock *100

Quote:

---

Originally posted by Kunaak
hmm, lots of good ideas here.
how about this.

Prime 95-

enter the password 9876 (perfectly legal, comes with the program in a little text file, incase you wanna push your CPU even farther)

click on "round off checking"

then "torture test"
thats always been my way of running Prime.
if someone has a better idea.
put it on here.

brwmogazos

I like your CPU % idea.
so perhaps that should be apart of it, but I am not sure how.
what about those guys with say 3.4C's and running 4.5 ghz everyday?

indeed a powerful machine, but not high in %....

this challenge is just for fun, nothing to freak out about.
if your PC can't runn say 3.6 ghz today, then don't try the benchmark at 3.6 tommorow.
this is just to show STABLE OC's.

how about seperating it to Highest CPU type?
P4 against P4.
Xp against Xp.
A64 against A64 and so on...

---

sounds nice

seperating them P4 vs P4 etc sounds nice

if we are not looking into the highest % then users with a 3.4 and phase change getting 4.4ghz stable are winners straight away ;) thats why i suggested highest percentage or something...

Yep all the test should be completed cos we want stable oCs not max post :D

oh forgot...

and before anyone jumps in with an uber 2.4A prescott overclocked, we need to specify that Prescotts will be compared with prescotts, northy Cs with northy Cs etc :p

bugger, sorry brwmogazos didnt see that you had already suggested it. Will have to seperate the AMD's and the Intels?

what about also separating the cooling categories such as air cooling, water cooling, phase change, etc.

Quote:

---

Originally posted by Confused_User
what about also separating the cooling categories such as air cooling, water cooling, phase change, etc.

---

That I would think would be automatic, you can't honestly compare Water cooling with Phase change lol

`s

Can we have 2 entries? like Max clock, and Max FSB? Cause I run 2812x 255 but I can also run 2933 @ 225 both are pretty good feats I think ;)

Quote:

---

Originally posted by WildKard
Can we have 2 entries? like Max clock, and Max FSB? Cause I run 2812x 255 but I can also run 2933 @ 225 both are pretty good feats I think ;)

---

i would believe this challenge is about whatever max u run your system at most of the time and meets all the requirements mentioned in first post, can't u run 255x11.5 instead 255x11 and have same CPU clock that u get with 225x13?

like, u see, my max benchable is 244x11 but i run it 24/7 @ 239x11...

ok, gotta work on those tests now... ;)

Why bother seperating everything? It won't matter as long as everyone posts their cpu, mobo, videocard and cooling. People here are smart enough to judge what is a good overclock and what isn't.

this is one i did a 2 weeks ago

2.8C SL6Z5 week 34, 2003 with March 18, 2004 pack date
Abit AI7 16beta1 bios
2x 256MB KHX3200K2/512
5:4
NB Strap 800FSB
GAT = SREAED

Gigacube 9800PRO Extreme @ stock 398/702
Corsair Hydrocool 200

http://www.fileshosts.com/pentium4/D.../cpuz_idle.jpg

http://www.fileshosts.com/pentium4/D....65/pifast.jpg

http://www.fileshosts.com/pentium4/D...hrs21min_2.jpg

http://www.fileshosts.com/pentium4/D...8-351_6433.jpg

didn't run 32M though only SuperPI 16M

Will have ago at rerunning some of the other tests :)

http://168.144.93.22/cpu-zscore.jpg

24/7 gaming settings. air cooled no vmods.

3.0c - 1.55v ( cpu-z doesnt show it for some reason )
Ai7
2x 256 mushkin 2-2-2 (no CPC)
9800pro

well if you seperate everything into all of the catagories that have been mentioned its gonna be a pain in the ass to keep straight. much too complicated. i think it should be p4 vs p4
axp vs axp
a64 vs a64

You have to separate it according to cooling (air/water/phase-change) otherwise there's no contest, and its just the person with the most money who wins :(

Quote:

---

Originally posted by bachus_anonym
i would believe this challenge is about whatever max u run your system at most of the time and meets all the requirements mentioned in first post, can't u run 255x11.5 instead 255x11 and have same CPU clock that u get with 225x13?

like, u see, my max benchable is 244x11 but i run it 24/7 @ 239x11...

ok, gotta work on those tests now... ;)

---

I wish, but certain multipliers and FSB combinations wont boot into windows for me

I can boot 12x250 for 3ghz but thats just benching speed and far from 24/7 but I cant boot 11.5 x 255 otherwise that would be my 24/7 setup...its very strange this system :( I cant even run 12.5x230 which is 2875 :(

Quote:

---

Originally posted by eva2000
this is one i did a 2 weeks ago

2.8C SL6Z5 week 34, 2003 with March 18, 2004 pack date
Abit AI7 16beta1 bios
2x 256MB KHX3200K2/512
5:4
NB Strap 800FSB
GAT = SREAED

Gigacube 9800PRO Extreme @ stock 398/702
Corsair Hydrocool 200

http://www.fileshosts.com/pentium4/D.../cpuz_idle.jpg

http://www.fileshosts.com/pentium4/D....65/pifast.jpg

http://www.fileshosts.com/pentium4/D...hrs21min_2.jpg

http://www.fileshosts.com/pentium4/D...8-351_6433.jpg

didn't run 32M though only SuperPI 16M

Will have ago at rerunning some of the other tests :)

---

my 2.8 M0 has the same fpo matey

costa rica ;)

does ~4.2Ghz prime stable with phase change and 1.75V

cant u push a bit higher? :)

sounds awesome

and it is feasible to have all those things open on your screen for a screenie if you just leave them open after theyre done running

Quote:

---

Originally posted by blinky
sounds awesome

and it is feasible to have all those things open on your screen for a screenie if you just leave them open after theyre done running

---

Or do them seperately, save the screenshots and make a collage in paint or photoshop. ;)

I think that is much easier if you are doing multiple benchmarks considering you'll be rebooting before each (most likely)

4250MHz (Prime95 torture)

Was Prime stable at 4.3GHz with ajusted droop. Dont have picture thou :(

Quote:

---

Originally posted by brwmogazos
my 2.8 M0 has the same fpo matey

costa rica ;)

does ~4.2Ghz prime stable with phase change and 1.75V

cant u push a bit higher? :)

---

look at my cpu temps and compare to yours ;)

i can bench at 2.8C @280fsb but not prime stable cpu load temps don't go much over 50-54C

i still have 2 exact same fpo/batch 2.8C SL6Z5 i have yet to test so who knows :)

unfortunately not prime96 stable but still

Still with Abit AI7 and 2x 256MB KHX3200K2/512 5:4 ratio
NB Strap: 800FSB
GAT = SREADD

http://www.fileshosts.com/pentium4/D...uperpi-32m.jpg

http://www.fileshosts.com/pentium4/D...PAT/pifast.jpg

http://www.fileshosts.com/pentium4/D...enderbench.jpg

This sounds like fun. I will see what I can do

32M superPi

http://members.home.nl/youpy/superpi.jpg

Still havent found a "bullet proof" stable setup in that i get prime errors and such, but the clocks I am running are 24/7 benchable gameable and etc.

I think its because i'm running 3x256 sticks of ram..? Oh well, will find out soon enough.

so whats gonna determine which entry is faster between two? 3dmark01 score? or is this just for CPU OCs

the first factor is whether or not they can actually complete all the benchmarks.

theres a few benchmarks posted so far, but no ones done the big 5 yet.

then from the people that can complete the big 5.
it's pretty easy to figure out who's got the best OC of them.

Here goes test 5:

3d mark 2003 -- 488 points with crappy Xabre 200 :D
http://service.futuremark.com/compare?2k3=2380321

Dunno if I'm going to pass test 3 (prime crashes for no reason every time)

My computer can do 24 hours in Prime and I can play for hours. I pass this ;)

2500+ @ 2.6 Ghz @ 1.85V FSB200-225
R9800 @ 480/385
NF7-S with L12 mod
all watercooled

Now this is what I've been waiting for.....finally a reason beyond 3DMark to top out my OC and have it 100% stable. Sad thing is I'm stuck with a bunk AI7 that artifacts over 230FSB.....has anyone actually found a sure-fire fix for this yet???

EDIT: If i can't find a fix for this artifacting then I'll probly just wait 'till I get my MAX3 to actually "submit" my results here, but kudos to you Kunaak on puttin' this up. :cool:

With my 30 cap SL6WK 3.0 i reached 3.8Ghz(253Mhz FSB) on my max3...

Is there anyway i can get it higher??? All specs in my sig...

Tomorrow or next day i get a fresh new pair of pc3500 mushkin lvl2 ;o)

Do i need to Vmod them to have low timing(2-2-2-5/6/7)@~255Mhz???

Quote:

---

Originally posted by Mini
Do i need to Vmod them to have low timing(2-2-2-5/6/7)@~255Mhz???

---

From what I've seen there arent any actual volt mods needed, you just need to GIVE them lots of voltage. You are talking about PC3200 Black lvl2 right? Most people that take it THAT high are looking at around 3v to have those timings at 1:1 up to and above 240FSB or so. Have you ever tried bumping the dram voltage that high on your MAX3? Some MAX3's have issues with the dram votlage in which it'll start to overvolt the memory beyond about 2.8-2.9v.....and it isn't very healthy for you memory at all.

No... Its Mushkin pc3500 Level 2... http://www.danskdatasalg.dk/visvare.asp?id=533

Im givin my ram 3.2V as it says in sig...

Well i heard from others in here that they modd to give 3.3 and over...

Quote:

---

Originally posted by Mini
No... Its Mushkin pc3500 Level 2... http://www.danskdatasalg.dk/visvare.asp?id=533

Im givin my ram 3.2V as it says in sig...

Well i heard from others in here that they modd to give 3.3 and over...

---

That's surely some nice ram you've got there.

You're referring to voltmodding the mobo then right? I'm glad to hear that 3.2v is workin' out for ya.....cus some people do get stuck with boards that are quite unstable with anything 2.9v+.

Yes i am... Im asking if i need to Vmodd to reach around 260 FSB@2-2-2-5/6??

Quote:

---

Originally posted by Mini
Yes i am... Im asking if i need to Vmodd to reach around 260 FSB@2-2-2-5/6??

---

I can't say for sure, but I do see a lot of people requiring 3.4v or so to get above 250FSB so it IS a very real possibility.

I wouldnt say it's like MC escher.

think im gonna have to do this at some point cuz obviously when not hardcore benching i want stability. at the mo i'm running 2.4ghz perfectly stable but i dont have the time to prove it.

...saving spot for future benching

& hi Cornelious0_0, hows it going

Quote:

---

Originally posted by fr33ze
think im gonna have to do this at some point cuz obviously when not hardcore benching i want stability. at the mo i'm running 2.4ghz perfectly stable but i dont have the time to prove it.

...saving spot for future benching

& hi Cornelious0_0, hows it going

---

Hey, not bad.....man, it's nice seeing a lot of you guys over hear too.....just so I don't feel COMPLETELY on my own here. ;) :)

yeah i know, been a bit weird without the [H]ard crew but i'm starting to like it hear, taking my knowledge to a new level that [H] couldn't really provide, phase change is starting to sound awful tempting.

Quote:

---

Originally posted by fr33ze
yeah i know, been a bit weird without the [H]ard crew but i'm starting to like it hear, taking my knowledge to a new level that [H] couldn't really provide, phase change is starting to sound awful tempting.

---

Wow, sounds like you and me are on basically the same page here. Good to hear.

I think I have it. for now..

Waiting for grafx to "take" 3Dmark:D

M1, M8 M32 Pifast, etc. etc. you name it. All pas 3300 Mhz fastest mem timings

Inc "funny" German bench:
http://www.xtremesystems.org/forums/...7&goto=newpost



M1 (3360)
M8 (3320)
M32 (3300)
Pifast (3391)
CPUMark99 (3343)
3Dmark (3320-3360) not officiel results yeat. ;)

105 kg of solid Cascade -120 C / -180 F Max load 350 W
Just finish after almost 1 year.

More later
:D

:D

Quote:

---

Originally posted by Tom Holck
105 kg of solid Cascade -120 C / -180 F Max load 350 W
Just finish after almost 1 year.

More later
:D

---

You go dude.....truely awsome. :slobber:

Tom Holck, wtf, thats like the biggest cascade i've ever seen...then again 105kg to go with it is hardly gonna be small, nice work.

oh and 3.3ghz a64, dam thats a nice oc, like to see it paired with a x800 xt

Quote:

---

Originally posted by fr33ze
Tom Holck, wtf, thats like the biggest cascade i've ever seen...then again 105kg to go with it is hardly gonna be small, nice work.

oh and 3.3ghz a64, dam thats a nice oc, like to see it paired with a x800 xt

---

For sure dude.....you've got me salivating, that's for sure. It might not take a lot to GET me like this with hardware.....but it must still count for something right??? ;)





:slobber:

nice tom you rock.

Here's what I did with my rig for the toaster sig below...

The only thing i have not been doing for over 5months now is 3dm2k3.. :rolleyes:

I combinated the lower two benches with the upper fullscreen... :)

Good god man!!!!! That's some tight :banana: :banana: :banana: :banana: right there. ;)

wait what do you have to do to get a toaster in your sig? lol
and who givesd out these toasters?

Quote:

---

Originally posted by blinky
wait what do you have to do to get a toaster in your sig? lol
and who givesd out these toasters?

---

Angry_Games/king of n0thing over at the DFIstreet forums... you must have a DFI board though... And you have to pass the tests at either 250fsb or 2.5GHz+... dual channel.

Tests: 8h+ prime95, stop it, then do 3dmark01 and open two instances of cpuz (mem ->dual, clocks) after that. you have to let them all on your desktop and take a screenshot. No photoshopping... must all be run in series.

Further info there.

Sounds good. Except I don't have a DFI board. :(+

SAE...cooling??

Quote:

---

Originally posted by SwimmerBoy
SAE...cooling??

---

NVentiv (R.I.P.) Mach2 GT... watercooled NB, MCW50-T vid card... sinks everywhere :lol:

Quote:

---

Originally posted by SAE
NVentiv (R.I.P.) Mach2 GT... watercooled NB, MCW50-T vid card... sinks everywhere :lol:

---

Yeah, pretty much one of the highest fsb/clock speed combos ever seen on AMD XP. Makes me and my pathetic 2.5ghz on air feel like this :owned:

Quote:

---

Originally posted by mrlobber
Yeah, pretty much one of the highest fsb/clock speed combos ever seen on AMD XP. Makes me and my pathetic 2.5ghz on air feel like this :owned:

---

:lol:

I was surprised of that myself ;)

Quote:

---

Originally posted by Kunaak
1. Complete 32M on Super Pi - Time irrelavant
2. Complete Hexus Pi-Fast- again time irrelavant
3. Complete 12 hours of Prime 95 atleast.
4. Complete FULL 3DMark2001SE test (not just game) and provide link.
5. Complete FULL 3DMark2003 Test (again all tests) and provide link.

---

Heya Kunaak,

What happens if I don't have if I don't have a valid 3DMark03 key??

The other tests are no problems but the last test is...

How's about posting a pic of the results??

Someone's probably asked before me but I just thought I'd ask again anyways...

Cheers,
Minnyboy

lool SAE I always thought that toaster was some kind of casemod project you had going..

and then its a toaster that´ll toast and toast and toast and toast and.........



btw. very nice clocking as always :)

Quote:

---

Originally posted by bias_hjorth
lool SAE I always thought that toaster was some kind of casemod project you had going..

and then its a toaster that´ll toast and toast and toast and toast and.........



btw. very nice clocking as always :)

---

Nononono... no project - just a DFI thingy you have to qualify for.

Guess what!? I have the highest clocked toaster ;)

SAE..allways worst:D..damn 9h in P95@3ghz+.....i think ya GT owns my r404-not-tuned MachI:D.But i think i have the 2cnd highest:D

Quote:

---

Originally posted by gabbax
SAE..allways worst:D..damn 9h in P95@3ghz+.....i think ya GT owns my r404-not-tuned MachI:D.But i think i have the 2cnd highest:D

---

:lol:

gabbax, it's pure fun to compete with you ;)

Quote:

---

Originally posted by SAE
:lol:

gabbax, it's pure fun to compete with you ;)

---

And against u 2 SAE...but the Q, damn..8h LN2...na dont think so:D

Quote:

---

Originally posted by Minnyboy
Heya Kunaak,

What happens if I don't have if I don't have a valid 3DMark03 key??

The other tests are no problems but the last test is...

How's about posting a pic of the results??

Someone's probably asked before me but I just thought I'd ask again anyways...

Cheers,
Minnyboy

---

a pic with the full test results should be fine.

but so far, nones actually completed the entire challenge, just the easy ones....

Quote:

---

Originally posted by Kunaak
a pic with the full test results should be fine.

but so far, nones actually completed the entire challenge, just the easy ones....

---

Hmm. What do you mean with just the easy ones :confused: ?

Is 3dm03 so hard??? I don't think so... I later do a run and post a screenie. Then my challenge should be complete, shouldn't it??? :D

1. Complete 32M on Super Pi - Time irrelavant
2. Complete Hexus Pi-Fast- again time irrelavant
3. Complete 12 hours of Prime 95 atleast.
4. Complete FULL 3DMark2001SE test (not just game) and provide link.
5. Complete FULL 3DMark2003 Test (again all tests) and provide link.

a few people have done 2-3 of the challenges, but not the Prime95 one.
stability is the point of this challenge.
not who can just get the highest/fast scores on certain benchmarks.

I will have a go kunnak, but prime95 will be hard because I do not have my computer on all the time, guess I will have to get up early:)

Edit: Kunnak are you a photographer?

Quote:

---

Originally posted by JWB


Edit: Kunnak are you a photographer?

---

sure am.
thats how I make the majority of my money.

Quote:

---

Originally posted by Kunaak
sure am.
thats how I make the majority of my money.

---

These are great pics :)

You seem to be not only a great overclocker :D :thumbsup:

hey Kunaak, my 9800pro broke yesterday and all i have is an mx440. this means that nature in 3d01 and most tests in 3d03 will not be able to run because of them being dx9. so this means only people with compatible cards will be able to do the challenge.

do whatever parts you are able to actually do.
even with your MX440.

it's not who can get the highest super score.
it's stability that counts on this challenge.

Oh sorry kunaak I forgot all about this I will try this weekend.