Can anyone prove an 8hr Orthos Screenie of 4.5ghz on air with a retail E8400?

[Forsaken1]

Well give us another run with the window shut then lol.

Plenty of peps doing that(window shut).Im intrested in max stable prime, acouple of hours regardless of cooling method.Any one doing 4.5(prime) on air is either open window or very deep into temp max.

[Zucker2k]

Plenty of peps doing that(window shut).Im intrested in max stable prime, acouple of hours regardless of cooling method.Any one doing 4.5(prime) on air is either open window or very deep into temp max.

Besides, "Air" is very subjective because not all "air" temps are the same. Plenty people do "air" with PC in front of their AC units.

Edit: Here's a typical example of sub-zero "air":

http://www.xtremesystems.org/forums/...18#post2744218

Notice those voltages for those clocks? I only needed 1.536v (1.550v set in bios) to prime @ 4.6Ghz. Besides, doesn't it take some prepping to bench at sub-zero temps? Gah!

[Forsaken1]

True man.I rember a trip to the ukraine.The gov
(powers that be) did not turn on the heat until the pipes froze .Thats some chilly air .

[godsfist]

Mine topped out at 4.2 GHz but I use a silent 120mm fan spinning at 800 RPM

[Sniper]

Besides, "Air" is very subjective because not all "air" temps are the same. Plenty people do "air" with PC in front of their AC units.

Edit: Here's a typical example of sub-zero "air":

http://www.xtremesystems.org/forums/...18#post2744218

Notice those voltages for those clocks? I only needed 1.536v (1.550v set in bios) to prime @ 4.6Ghz. Besides, doesn't it take some prepping to bench at sub-zero temps? Gah!

Its nice to see all kinds of oc, be it air, water and other. But when someone shows of their cpu, its always nice to know what kind of cooling they used, or else it is kind of pointless. Claiming normal room temp when rad gets cold air from window or other sources is not nice, even though the temps where you are sitting is warm enough.
Try to put less ego into it, its just a cpu.

[Speederlander]

Plenty people do "air" with PC in front of their AC units.

Air conditioned rooms are pretty standard.

People just need to list their ambient air temps. Solves all of the problems.

[Don_Dan]

Air conditioned rooms are pretty standard.

Only in the U.S. and maybe Australia, but not in Europe or on most other places on earth.

[little_scrapper]

so far i passed Dual Super PI 32MB @ 4.5ghz with 1.45v, 3dmark06 and 05, but everytime i run orthos or prime95 it will crash instantly, i am still trying to find out what causing the problem.

I know what the problem is!! Your CPU is unstable at 4.5GHz and 1.45V. 3DMark scores really mean "SQUAT!!!!" in terms of stability. Because even if your CPU is making massively huge calculation errors on 3DMark it may may only manifest itself in ways such as rendering a pixel in the wrong color. Yet the program will finish no problem. You will never notice that every 5th pixel is being rendered in the wrong shade of green.

I put much more faith in programs like prime and such that actually make long string calculations that are checked at the end for accuracy. But at the end of the day what really matters is if your data integrety lasts for months. You can pass every benchmark and prime number but if you comp is crashing periodically (every 3 - 6 months) or you are finding that you are losing pictures of your kids that you have saved in "my pictures" then that is the true test of stability.

Not trying to get all weepy or even Existential but I had a comp that would pass ANY benchmark you tossed it yet every 3-4 months or soo it would crash unexplicably and I was finding that files were coming up corrupt. I first thought virus but to this day I have never been able to find any correlation between ant virus and system crashings. I resolved myself the the fact that it was probably the defrag while OC'ed was causing errors that no other programs caught. Could still be wrong theough. I almost lost my kids entire set of 1st birthday pics due to a system crash. Luckily for me I have learned from the past an I make several backups... and I do it often!!! phhew... the wife was not happy for a few days until I produced proff positive that I in fact did have all the pics restored....

[dinos22]

i was testing my new X38 board yesterday and did some quick 10 minute prime jumps to see where my chip is at

basically 4.4GHz 1.44-1.45v on water at NORMAL 25C ambient
load temps around 61C

i am anticipating further 100MHz jumps with 0.05v added to vcore so it can be done but very stupid IMO as this vcore is getting too high for 24/7

[Cronos]

I am so eager to see ANYONE pass Linpack64 benchmark on new 45nm CPUs. Both for benchmark results, to compare efficiency of new 45nm Core2 with previous 65nm one, and to see how far it will go under REALLY stressing test.

[rge]

I downloaded linpack64 a couple days ago. That program is very inefficient at finding unstable overclocks. You have to run it once for long enough to get an error (but you have no idea length run needed). Then run it again and compare results to see if there was an error. You get no clue as to ram or cpu, or how far off you are. And since it does not notify right when error occurs, you could run it for hours needlessly, when orthos or prime would notify you right away.

IMO, it is piss poor at efficiency and thus worthless for overclock testing.

If you want to argue that after you find your stable overclock with orthos, you could find out if it is even more stable by running linpack64, then first you would have to demonstrate a need, and second prove that its better.

If my system is orthos stable on both small ffts 8-64 and on large ffts or blend, both for 15 hrs, I have never had a system crash or instability problem. Therefore there is no need for a better program than orthos or prime for my desktop. And second, you haven't even proved that it does a better job of finding instability, even though its a worthless endeavor even if it did.

If there is more to linpack64, and an efficient way to help with stability testing, please educate me, I am all ears. Until you have shown me that, I will go on thinking you are selling something or smoking something.

[Bail_w]

I know what the problem is!! Your CPU is unstable at 4.5GHz and 1.45V. 3DMark scores really mean "SQUAT!!!!" in terms of stability. Because even if your CPU is making massively huge calculation errors on 3DMark it may may only manifest itself in ways such as rendering a pixel in the wrong color. Yet the program will finish no problem. You will never notice that every 5th pixel is being rendered in the wrong shade of green.

I put much more faith in programs like prime and such that actually make long string calculations that are checked at the end for accuracy. But at the end of the day what really matters is if your data integrety lasts for months. You can pass every benchmark and prime number but if you comp is crashing periodically (every 3 - 6 months) or you are finding that you are losing pictures of your kids that you have saved in "my pictures" then that is the true test of stability.

Not trying to get all weepy or even Existential but I had a comp that would pass ANY benchmark you tossed it yet every 3-4 months or soo it would crash unexplicably and I was finding that files were coming up corrupt. I first thought virus but to this day I have never been able to find any correlation between ant virus and system crashings. I resolved myself the the fact that it was probably the defrag while OC'ed was causing errors that no other programs caught. Could still be wrong theough. I almost lost my kids entire set of 1st birthday pics due to a system crash. Luckily for me I have learned from the past an I make several backups... and I do it often!!! phhew... the wife was not happy for a few days until I produced proff positive that I in fact did have all the pics restored....

well, my main rig is for me to overclock for fun so i dont have anything to lose

@rge - have you tired OCCT 30 mins or 1 hour test yet?

[fornowagain]

Anyway, not an 8400, but it is stable. Orthosing at 4.7ghz now.

Yeah on blend, gotta wonder if you can do that on smalls.

[Movieman]

Gentlemen:
Let's stop the name calling and accept this for what it is.
Screen shots shown at a specific speed with cold air hitting the system.
That's all this is and the poster made no attempt to hide that fact.
He even shows us a pic with the system sitting in an open window.
There is NO and I repeat NO reason for people to be arguing here.
There was no attempt at deception so I ask all of you nicely to take a deep breath and relax.
Thanks for reading!

[syne_24]

I was looking at the numbers and the overclock scaling seems to be the same on 45nm and 65nm. For instance, on E6600, the average orthos stable is 3.6 (that's 1.2ghz OC) Now on E8400, most seem to get stable around 4.3 which turns out to be 1.3ghz OC. So actually 4.3 orthos stable is quite good and to expect 4.5 is a little bit unrealistic IMO; especially on air/water.

In fact when Conroe first came out, not many people were able to hit 3.6 in those first few batches. And the ones that did hit those on air/water needed high amount of vcore. Not until the "B" stepping came and eventually GO that it became the standard orthos stable clock. So for Wolfdale first few batches to slightly beat Conroe in overclock is pretty damn impressive IMO. I think down the road 4.5 orthos stable would be possible with better revision. But for now I'm pretty happy with Wolfdale and it can only get better.

[rge]

well, my main rig is for me to overclock for fun so i dont have anything to lose

@rge - have you tired OCCT 30 mins or 1 hour test yet?

Yep, I like OCCT and use it for quick runs as well. But thanks, that is a helpful program.

I was having problems couple days ago pinpointing a problem, til I ran memtest and my ram is bad...errors no matter settings...have some crucial pc8500 on the way, and will rma mine then for backup use.

I had downloaded linpack64 after hearing the same person tell how great it was...what a waste of time.

[fireice2]

Wish I could earn as fast and as much as you guys.

$220 take me a least a month or two to earn.

[Cronos]

If there is more to linpack64, and an efficient way to help with stability testing, please educate me, I am all ears.

See the thread on proper Linpack testing, i posted the detailed instruction on how to use it.
There is no way with any test to tell how much time it will take to find an error.

[sofarfrome]

I put much more faith in programs like prime and such ...

I had a system built around an E6600 L629B a while back that was Prime95 stable while looping 3DMark03 in the back ground. 12hours + at 3.71GHz using 1.42vcore. However, when I loaded up UT2004 and started playing it would crash within 10 minutes. Hence I do not put a lot of faith in P95 or orthos, especially with the latest CPUs.

[sofarfrome]

I was looking at the numbers and the overclock scaling seems to be the same on 45nm and 65nm. For instance, on E6600, the average orthos stable is 3.6 (that's 1.2ghz OC) Now on E8400, most seem to get stable around 4.3 which turns out to be 1.3ghz OC. So actually 4.3 orthos stable is quite good and to expect 4.5 is a little bit unrealistic IMO; especially on air/water.

In fact when Conroe first came out, not many people were able to hit 3.6 in those first few batches. And the ones that did hit those on air/water needed high amount of vcore. Not until the "B" stepping came and eventually GO that it became the standard orthos stable clock. So for Wolfdale first few batches to slightly beat Conroe in overclock is pretty damn impressive IMO. I think down the road 4.5 orthos stable would be possible with better revision. But for now I'm pretty happy with Wolfdale and it can only get better.

Not to mention that clock for clock the wolfdales blow the conroe away

[aoch88]

I had a system built around an E6600 L629B a while back that was Prime95 stable while looping 3DMark03 in the back ground. 12hours + at 3.71GHz using 1.42vcore. However, when I loaded up UT2004 and started playing it would crash within 10 minutes. Hence I do not put a lot of faith in P95 or orthos, especially with the latest CPUs.

Maybe it was something else - heat?
Orthos is good for multi-cores so use that instead of Prime95.

[rge]

See the thread on proper Linpack testing, i posted the detailed instruction on how to use it.
There is no way with any test to tell how much time it will take to find an error.

I read it. I followed it. You are missing the point. I run orthos, and it crashes in 8 minutes. So within 8 minutes, I know I am unstable. I dont have to guess how big of a run to make or spend two or three hours running linpack once, then again and comparing the two.

Plus, if I crash in 8 minutes on orthos, I know to go up a couple notches, because I crashed quickly. Linpack, you would still be running the second run to compare the two, and have no idea about anything other than some errors occured.

The program is worthless for finding a stable OC.

Again, the only potential use would be trying to find any unstability left over after a successful 12+ orthos run, and not only is that worthless even if possible, but you have absolutely no evidence it would be any better at finding an error.

[adamsleath]

yeah definitly belongs to intel section, anyway it is quite a task 4.5 ghz air on 8400...4ghz is normal, 4.1 is good, 4.2 is very good, 4.3 VERY VERY good....everything above is just exceptional and a very good chip.

enuf said.

Not to mention that clock for clock the wolfdales blow the conroe away

hardly/barely,
but any improvement is good

[Cronos]

I read it. I followed it. You are missing the point. I run orthos, and it crashes in 8 minutes. So within 8 minutes, I know I am unstable.
***
The program is worthless for finding a stable OC.
***

Your argumentation is so weird i can't even find an answer.
I think I am done with you. Suit yourself.

[sofarfrome]

Maybe it was something else - heat?
Orthos is good for multi-cores so use that instead of Prime95.

Not heat..custom WC system that would keep a hot running L629B E6600 at 55C under load at 1.45vcore. I just needed more vcore to keep the game stable.