• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

E6600@3.0 crashes in games but stable in all tests??

P

Ph1l

Junior Member
Guys what to do?? It's 9x333 on P5B Deluxe, Geil Ultra 4-4-4-12 @833

Vcore = 1.3625
VDimm = 2.15V
FSB term = 1.45V
NB = 1.4V
SB = Auto
PCI = 33.33
PCIEx = 100

ORTHOS test is going over 1h and no errors. SuperPI 1.5 mod = 32M - no problem, but when I enter for example Splinter Cell it crashes after few minutes of playing with the critical error of dx9drv.dll (running on GF7800GTX not overclocked). Sometimes before it crashes I get few minutes slide-show for example in NFS also. What shall I change to make it better. 3mark 2003 also is running without any problems.
 
That is directx9.0 device drivers...

Sounds like video card driver errors..what card and what are your current drivers. I got that once when I installed ATI CAT drivers first and then DX drivers...

I would use drivercleaner and clear out all applicable video card drivers. reinstall DirectX9.0c drivers again...reboot...then install video card drivers again. Make sure no beta drivers.
 
The drivers are 93.71 - no modded, the system is totaly new. When I came back to 2.52GHz (9x280) all is OK, but it's not O/Cking
 
Try giving the SB a little extra voltage too. Worth a shot.

Also try higher Vcore just to see if anything changes at that point. Rule out everything.
 
I think I'll do soon, now I turned the "PEG Link mode" to Normal from Auto. Strange situaltion also happened when I changed the timings to 5-5-5-15 from 4-4-4-12 on the same rate : 833. On bigger timings the ORTHOS failed after 3 seconds, on lower timings it's going stable whats going on!! Changed the VCore to 1.375V
 
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
 
Remember to turn off spread spectrum, turn off memory remap, turn off C1E etc.

also turn off static read control for your memory.

All of these things hinder overclocking so turning them off helps.
 
since 9x333 =2.977 does NOT work in the game, and 9x280 = 2.52works. you think it is the memory. But have you tried anything to isolate this such as keeping the memory speed low or constant like at the stock speed, just keep raising only the cpu to that 2.977 GHz to make sure that is fine (should be), then set the cpu low to like the stock via the multiplier and keep upping the speed so the memory reaches it's highest. once you know the answer to that you can tweak far more accurately. That is my best experience when i go and overclock.
 
I think I got it. It's memory!! When I decresed memory voltage from 2.15 to 2.05V everything works fine. I'm after few hours of playing in Splinter Cell which was crashing quite often, and nothing happened... On 2.10V I got BSOD with typical for memory IRQ_NOT_LESS_OR_EQUAL, on 2.25V I got the same BSOD, but even before the windows loaded, so I started decresing voltages and it seems 2.05V is fully stable. I'll make some more tests and give you the replay.

PS : The problem with game crashing also came with 2.52GHz, the problem generaly started when I started to set the Vdimm manualy.
 
BTW, 1 hour of Orthos is no where near stable-I mean its up to you, but 8 hours is my personal minimum and normally I go for 12+ when I am trying to achieve max system stability. But, sounds like either memory or graphics errors all the same.
 
Maybe its a heat problem. Since orthos only stresses your CPU and/or memory, it wont crash. However, since games will ALSO stress your GPU, there may be too much heat in your case. That said, it can also be that your PSU craps out when your rig draws too much power (again, games would make this weakness evident).
 
Run your memory at those timings but do 1:1 for 666 instead of 5:4 for 833. Most likely it will be slightly faster and you won't be stressing your RAM so much. I'd set the voltage of the RAM at 2.0V. I've been running some Kingston RAM like this and it runs beautifully.
 
Originally posted by: Markfw900
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
Click to expand...

this is what i am thinking too....
 
Originally posted by: Ph1l
I think I got it. It's memory!! When I decresed memory voltage from 2.15 to 2.05V everything works fine. I'm after few hours of playing in Splinter Cell which was crashing quite often, and nothing happened... On 2.10V I got BSOD with typical for memory IRQ_NOT_LESS_OR_EQUAL, on 2.25V I got the same BSOD, but even before the windows loaded, so I started decresing voltages and it seems 2.05V is fully stable. I'll make some more tests and give you the replay.

PS : The problem with game crashing also came with 2.52GHz, the problem generaly started when I started to set the Vdimm manualy.
Click to expand...


Nice work - I'd re-run a few more tests in depth to check, this time for a few more hours.
 
Originally posted by: bob4432
Originally posted by: Markfw900
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
Click to expand...

this is what i am thinking too....
Click to expand...


With a single 7800GTX and a Conroe? I'd think even a 300-350w PSU could handle that easily.


However -> Leadtek GF7800GTX Extreme 490/1250 [no o/c] [8800GTX in January]

Not without a new PSU first Mister! 26A on the 12V rail isn't going to cut it with an 8800GTX.
 
Originally posted by: Golgatha
Originally posted by: bob4432
Originally posted by: Markfw900
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
Click to expand...

this is what i am thinking too....
Click to expand...


With a single 7800GTX and a Conroe? I'd think even a 300-350w PSU could handle that easily.


However -> Leadtek GF7800GTX Extreme 490/1250 [no o/c] [8800GTX in January]

Not without a new PSU first Mister! 26A on the 12V rail isn't going to cut it with an 8800GTX.
Click to expand...

just sounds like a power issue from the information given. i am not familiar with spire but there are many times i have read about people maxing out the cpu with orthos or 2xprime95 only to have it fail in a game because the gpu takes so much power. really puts a serious strain on it. otherwise it would fail under orthos if it were memory or cpu related. now i do think he should test it for longer, usually i do my setups 12+hrs before i call them stable, whether it be a o/c or a new build.

gaming really stresses a system because of the extra draw the gpu uses. recently i started f@h and used the gpu client along with a cpu client. well the gpu client needs a cpu core too, so with 1 cpu client and 1 gpu client i have both cores and the gpu maxed out. machine puts off a ton of heat, but stable (stopping using the gpu client due to heat output with my x1800xt - too much)

the game company of heroes is one that really stresses stuff pretty hard. you can get a demo and mess around with it. good game and a good stress test. i have heard of more than 1 person saying there machine was dual prime stable for 12+hrs and it would not be able to handle coh. not sure how splinter cell is.

i guess it could be a heat issue if the case has no airflow and the fans are clogged...maybe check that out too op.

definately let us know what your conclusion is. remember, not all psus are the same and manf specs don't tell the whole story. sometimes you are better off getting a lower rated psu from a better rep than a higher rated one from a not-so-well know rep. again, i don't know about spires rep, for all i know they could be good and you got a bunk psu.

to see about heat, maybe install speedfan, something that will log temps, so you can see how much hotter your cpu and gpu get during the game. this is another avenue.
 
Originally posted by: Markfw900
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
Click to expand...
Spires are about the same quality/output as Rosewills. I'd guess that the psu is the problem is this case, also.
 
Originally posted by: myocardia
Originally posted by: Markfw900
that spire 600w, I have never heard of. could be a weak PSU. If that 600 is really like=400 watt seasonic, thats your problem.
Click to expand...
Spires are about the same quality/output as Rosewills. I'd guess that the psu is the problem is this case, also.
Click to expand...

I don't think so. I read some about this PSU before I bought it, no problems, voltages stable. Nothing happened from the last time, so it was memory. I'm also sure before switching to another new graphic card I'll also change the PSU, just in case of...

look at the tests, before you are saying something you don't know what it is.

http://www.3dgameman.com/vr/spire/600w_psu/written_review_03.html

I had the option not for North America (without active-PFC), but European option with active-PFC
 
Originally posted by: Ph1l
I don't think so. I read some about this PSU before I bought it, no problems, voltages stable. Nothing happened from the last time, so it was memory. I'm also sure before switching to another new graphic card I'll also change the PSU, just in case of...

look at the tests, before you are saying something you don't know what it is.

http://www.3dgameman.com/vr/spire/600w_psu/written_review_03.html

I had the option not for North America (without active-PFC), but European option with active-PFC
Click to expand...
Um, I went directly to Spire's website to check the specs on that psu before I posted, and noticed that it's an ATX ver. 1.3 psu, which were built to power the cpu from the 5v rail. And since it's a well known fact that the majority of hardware sites are paid to give good reviews of products (this is a very informative read, BTW), just know that any review that's linked to directly from the manufacturer isn't going to be a very honest review. That's why we trust known trustworthy sites, like Anandtech, Silent PC Review, JohnnyGuru, and only a very few others.

 
Um, I went directly to Spire's website to check the specs on that psu before I posted, and noticed that it's an ATX ver. 1.3 psu, which were built to power the cpu from the 5v rail. And since it's a well known fact that the majority of hardware sites are paid to give good reviews of products (this is a very informative read, BTW), just know that any review that's linked to directly from the manufacturer isn't going to be a very honest review. That's why we trust known trustworthy sites, like Anandtech, Silent PC Review, JohnnyGuru, and only a very few others.
Click to expand...

I agree here and my instict is the PSU also. Keep in mind that "IF" this problem is the PSU that you can in fact damage your system by using it. When a weak or cheap PSU gets hot and heavily loaded, the ripple current goes up. Ripple is actually AC current bleeding thru into your computer which can kill components over time. GPUs and HDs are the most susceptible to this.

 
You must log in or register to reply here.

TRENDING THREADS

Back
Top