orthos fail, safe?

V|per · 13 Mar 2007 at 23:22

Spec is: e6600 on a bfg 680I with zalman CNPS9500

Running at 400x9 with mem at 500mhz 5-5-5-15 2t (2x512 corsair xms2 pc6400)

Basically at this speed orthos fails after a couple of seconds, bet everything else that i have thrown at it (3dmarks/superpi/games/browsing/music/dvd) runs absolutely sweet as a nut with no crashes whatso ever.

Is this still safe to run (idles at 50c though..)

It's not the memory, as it does the same when its at stock.

This also occurs from about 3.2ghz upwards, at which i need to increase the voltage a fair bit (1.5v+) to make it orthos stable. I will eventually be putting this baby under water, so in the future im not too bothered about the voltage, i've seen people running these at 1.65v with temps still below 60c on orthos.

Would it be better to clock it down a bit and wait for the water cooling, or just leave it as it is, due to it not crashing at all?

Cheers,

Phil

andyh · 14 Mar 2007 at 06:28

What did you use to measure those temps i.e TAT, Coretemp or some mobo utility and what are your temps under load?
If its stable when doing everything you would normally do then I personally wouldnt be worried if orthos falls over after a few seconds.

kitfit1 · 14 Mar 2007 at 10:25

andyh said:
If its stable when doing everything you would normally do then I personally wouldnt be worried if orthos falls over after a few seconds.

I really don't know how many times this has to be said on this forum, or any others for that matter. If a rig is not Prime or Orthos stable, then it's not stable, period. If andyh is'nt worried about corrupting his install and data, that dos'nt mean the op should'nt as well. So back it down to a point where it is Orthos stable for 8hrs.

SidewinderINC · 14 Mar 2007 at 11:07

definately make it stable, you may not SEE it doing much damage at the present. but i corrupted my C&C3 demo install and that was a pain in the ass i can tell ya

no biggie, but what happens when you come to open a file one day and its corrupted itself? or windows decides that one bit isnt meant to be there and deletes it? it does happen

simonnance · 14 Mar 2007 at 11:12

have you tried tweaking the other voltages?

I am the proud owner of a DFI RD600, and i've worked out how vital the Vnb is (and all the other little voltages). You may not have the same range of options, and 400 isnt THAT high, but it's still worth checking out, save you frying your CPU with extreme volts.

overcl0cker · 14 Mar 2007 at 11:19

What CPU voltage are you running?

w3bbo · 14 Mar 2007 at 11:29

kitfit1 said:
I really don't know how many times this has to be said on this forum, or any others for that matter. If a rig is not Prime or Orthos stable, then it's not stable, period. If andyh is'nt worried about corrupting his install and data, that dos'nt mean the op should'nt as well. So back it down to a point where it is Orthos stable for 8hrs.

Q F T

spb251272 · 14 Mar 2007 at 12:31

Agreed - I tried to push mine to 3.7 at it's current vcore and when it failed Orthos after about 8 minutes, it was straight back down to 3.6 since I don't really want to push vcore any higher (set at 1.55, showing 1.51 actual).

I've done two seperate 12 hour tests with Orthos at it's current settings - just to make sure. One of the many things that I've learned on this forum is that an Orthos stable PC is a necessity

You may want to look at your cpu fsb, mcp and spp voltages as they may be what's holding you back. I have mine at fsb 1.3, spp 1.5, mcp 1.6.

andyh · 14 Mar 2007 at 16:31

kitfit1 said:
I really don't know how many times this has to be said on this forum, or any others for that matter. If a rig is not Prime or Orthos stable, then it's not stable, period.

That depends on your definition of stable, to me if it does the things I want it to do on a day to day basis without any issues, the fact that it fails orthos does not make it unstable.

If andyh is'nt worried about corrupting his install and data

Just because it fails orthos does not mean that my data will be corrupted.

that dos'nt mean the op should'nt as well

I never said he should, the op asked for opinion, I gave mine.

So back it down to a point where it is Orthos stable for 8hrs.

This is your opinion and thats fair enough

For me this was an issue of risk versus reward. I look at the risks, well, my system may be less stable than one that is Orthos stable for 8 hours but can that difference in stability be quantified? no it cannot.
If my system does fall over what may happen? it could be that it reboots in the middle of a game or worstcase scenario it corrupts some data.
I then look at the reward, does my machine work faster, play games with better framerates, when it runs at the higher clock speed.
I then decide whether to, accept that risk or not and based on the info provided by the OP then I would take the risk in his shoes. After all it probably is no more of a risk than overclocking, lapping and all the other things we do to get a faster machine.

Justintime · 14 Mar 2007 at 16:54

Well in my experience, a few seconds to crash orthos or prime95 or whatever form it takes usually means a fairly unstable PC. While p95 might not be the absolute it does give a good indication of instability since the pc can't calculate know values.

xsnv · 14 Mar 2007 at 17:29

Overclocking a cpu means making it run faster than stock. That implicitely assumes it works as well as it does at stock. If it doesn't then you havn't really overclocked it properly as you've lost some funcionality (in the case of prime the ability to accurately compute a calculation).

If you bought a cpu and it failed prime at stock speeds but "apparently" everything else yo uwanted it to would you return it or keep it?

kitfit1 · 14 Mar 2007 at 17:50

andyh said:
That depends on your definition of stable, to me if it does the things I want it to do on a day to day basis without any issues, the fact that it fails orthos does not make it unstable.

Just because it fails orthos does not mean that my data will be corrupted.

I never said he should, the op asked for opinion, I gave mine.

This is your opinion and thats fair enough

For me this was an issue of risk versus reward. I look at the risks, well, my system may be less stable than one that is Orthos stable for 8 hours but can that difference in stability be quantified? no it cannot.
If my system does fall over what may happen? it could be that it reboots in the middle of a game or worstcase scenario it corrupts some data.
I then look at the reward, does my machine work faster, play games with better framerates, when it runs at the higher clock speed.
I then decide whether to, accept that risk or not and based on the info provided by the OP then I would take the risk in his shoes. After all it probably is no more of a risk than overclocking, lapping and all the other things we do to get a faster machine.

This forum's function is to give people good and well trusted advice about overclocking. If you don't care about the stability of your clock and the risk that puts your data at, then that's fine. But to offer the same "advice" to other people is irresponsible. Your views are just that, "views", the op wanted sound advice which "I personally wouldnt be worried if orthos falls over after a few seconds" certainly is'nt.

overcl0cker · 14 Mar 2007 at 17:58

It's rather lol if you want to overclock your computer and not check the stability. The slightest things that put your CPU under 100% strain for a few seconds such as, lots of programs running and you open a browser, it struggles for a few seconds.. and.. your ...dead. And there goes all of your work documents!

What I've always done in the past is a minimum 30 minute Orthos whilst upping the FSB, and once I've reached what I want or what is my PC's potential I've ran 12 hours!

Overclocking is:

20% Actually doing it, getting it right
80% Checking it over. If somehtings wrong and it fails on you you've had it. So stability is the key to all overclocks, even the slightest ones.

ghgh · 14 Mar 2007 at 18:18

If its not 100% stable you will most likely get random corruption of data as I found out when I oc'd my pc too far it corrupted windows after a few days so I had to reinstall.....it might be more stable than that however and only corrupt a game install or a random programs install....I suspect that you will get reboots at random if its not orthos stable when under 100% load for periods of time but if this doesnt happen you are lucky guy,

Bizarre · 14 Mar 2007 at 18:37

I've had orthos fail my system after 6 and 7 hours both in the same day, it nearly had me in tears lol (ok, so I did cry a little bit

).

Any hewww... unless it passes 8 hours+ (12+/1 round to be certain) without any errors then quite simply I will not deem the system stable.

aztechnology · 14 Mar 2007 at 20:08

The good safe advice is to get it as stable as possible.

I had an a64 that I thought was OK at 45mins prime.

When it went it corrupted a hard drive and the Motherboard bios. The bios was a reflash, the harddrive needed a reformat before it was happy again.

Personally. I now leave a stability check at least overnight for 8 - 12 hrs.

People often quote "it's stable for 2 hours and I don't play xxx for much longer", but that's not the point. If it's not stable, it can go at any time, not at x minutes so I'm ok for x - 1 minutes. Testing for 8-12 hours just lets it check a lot of calculations.

V|per · 15 Mar 2007 at 00:36

it was at 1.6v at one point

D) just to get it at 3600mhz, but i deemed it way too unsafe, i dont really like to see the temps of this exceed 60c, although it's probably good for 80c.

With regards to toher voltages, i have played around with them, and most have actually been upped to keep it stable (non-orthos that is)

Managed to back the voltage down to 1.45 and is 24h orthos stable @ 3300mhz.. This is fine for me, i got me a bunch of screenies anyways to say that 'this is what i got

'

How does the hdd corruption work then? thought most of these were linked to pci speed locks etc? Athough i see where most of you are coming from with the 'needs to be orthos stable for x time' etc, find it odd that orthos running both cores maxed leaps out at 3 seconds, but supreme commander running both cores maxed out runs indefinately (yes i tried this for about 10 hours.. i know, way too much time on my hands...

)

Cheers for your imput guys

easyrider · 15 Mar 2007 at 00:46

andyh said:
That depends on your definition of stable, to me if it does the things I want it to do on a day to day basis without any issues, the fact that it fails orthos does not make it unstable..

Rubbish.

If your cpu cannot calculate data correctly and it fails in orthos.Then the cpu is not doing its job.

Just because it fails orthos does not mean that my data will be corrupted..

The cpu cannot calculate prime numbers if it fails orthos so data corruption is already happening.

I never said he should, the op asked for opinion, I gave mine..

Thats fine.But you are kidding yourself.A failed clock is errors in orthos especially after a few seconds.

For me this was an issue of risk versus reward. I look at the risks, well, my system may be less stable than one that is Orthos stable for 8 hours but can that difference in stability be quantified? no it cannot..

yes it can

If my system does fall over what may happen? it could be that it reboots in the middle of a game or worstcase scenario it corrupts some data.
I then look at the reward, does my machine work faster, play games with better framerates, when it runs at the higher clock speed..

running at higher clocks speeds should not mean instabilty.If you have overclocked and you are getting instabilty you have clocked
incorrectly.Simple as.Its a failed clock as the pc is not doing its job correctly.

I then decide whether to, accept that risk or not and based on the info provided by the OP then I would take the risk in his shoes. After all it probably is no more of a risk than overclocking, lapping and all the other things we do to get a faster machine.

Overclocking should not be a risk.If you cannot get a clock stable then lower the clock until it is or incres vcore etc...to compensate.

Having a clock that is not orthos stable is pointless,bad practice and pur failure.

Overclocking is done to increase perfromance while still maintaining a perfectly stable data integrity pc.

No ifs,no buts.

Thats how it is

Corasik · 15 Mar 2007 at 01:48

Sorry, but if its not orthos stable, even if your just use the computer for gaming and non serious applications, it can still be a pain in the ....

Imagine someone with a non orthos stable rig running seti work units (or folding or anything else like that). Their work units could contain errors, so they are worthless.

Or perhaps you've just installed a new game, but mid install the cpu miscalculates and instead of writing your game to sector 546663 of your hard drive, it writes it to sector 146663, and overwrites your registry. Oops windows corrupted.

Your CPU is the heart of the computer, and even a single calculation in error could corrupt your windows install, crash your game, or make your data worthless.

People complain about windows being badly programmed and crashing a lot, but I can count the number of crashes my PC's made in the last 9 months on 1 hand. And I use my PC heavily every day.

Of course, if the risk of application/windows corruption doesnt bother you, and you dont participate in any distributed computing groups, and it doesnt crash often enough to bother you, then sure go ahead, after all, the very worst that can happen is, a) you'll lose some time in a game if your PC crashes. b) if your windows install gets corrupted you may need to reinstall windows. Afterall, no matter what, even if your hard drive gets totally corrupted which would be very unlucky, there will be no physical damage, the cpu wont get a 'memory' for its mistake and become permanantly damaged. At least not if your sensible with the voltages and temperatures.

malccy · 15 Mar 2007 at 13:30

Orthos stable 12 hours and then crashed in cod2; increased voltage from 1.52 to 1.53 and is now fully stable.