My Core Duo scored 555 in 3DMark06, whats wrong? (CPU-Z pics)

[13thmonkey]

I read ages ago about the death from static discharge being slow and not instant, perhaps this is what it looks like.

I find it very very strange as any 'errors' created in a cpu I would have expected to crash the system very quickly, and not build up into more and more frequent and crashes.

You said that your temps were ok, what were you measuring it with (TAT is best) (I assume you are aware of all the problems with measuring temps when C2D's first hit the market)

Yes it does look like your CPU is a problem, BUT and this is a big BUT, the cpu is in all but a very few tests going to be reliant on external components to function properly. Your memory looks low, this may be dependent on the CPU to a small degree.

HDD's look good and virus is probably ruled out, and generic HDD failure is also ruled out as the format should have either spotted the bad sectors and warned you, or it should have accounted for them, so you should have started from scratch.

Its getting late over here so thats all from me, and I'm out of ideas other than over heating, but you seem to know a bit of what you are talking about so if you used something decent to measure load temps....

Another thought, have you looked around the CPU socket on the mobo to see if any of the capacitors or other components are discoloured or showing signs of damage?

 

[monsterrocks]

CPU would affect the memory I think. Doesn't the CPU have some sort of RAM controller or something? I think if your CPU has problem it affects the RAM. He ran Memtest, so if it passed, his RAM if fine. I used Memtest to find a faulty stick of RAM in my board. Very slight fault, but errors came up all over...No, if his RAM had anything at all to do with it, Memtest would have gotten an error somewhere...Did memtest have any errors at all? You should put the results of that up so that we can know whether or not it is your RAM. I think that it is conclusive that it is your CPU or your motherboard. Go into the bios and turn off speedstep and manually set the speed to it's stock value. Then run the test again. I think it is 266x9, right?

 

[skittle]

CPU would affect the memory I think. Doesn't the CPU have some sort of RAM controller or something?

Only AMD cpus have an integrated memory controller. Intel based motherboards have the controller in the chipset... thats what makes Intel cpus less dependant on memory speeds.

 

[Zorg]

I didn't see where you posted the results of your memtest. For grins run the CPU bench again and have CPUZ open to the CPU tab, you should see the multiplier go to 9 and speed to ~2.4. Download and run Core Temp while you are running the bench and post temps.

 

[Grimmy]

Kano523 mentions its been working fine for 7 months.... I'd suspect a power spike/surge happened, but the system survived it. I'd rule out ESD as a cause. (thats just me)

I noticed he didn't mention what kind of HS he was using, stock?

Since Kano523 says his temps are fine, I assume he has read the "Core 2 Quad and Duo Temperature Guide".

Perhaps it would be a good time (excuse *cough*) to build a quad system. :lol:

 

[Mondoman]

mr - you may want to consider doing more research before posting so as to improve the correctness of your posts. BTW, Orthos is *not* the exact same as PRIME95; rather, it is a shell that allows one to automatically run two copies of PRIME95, one per CPU core. The GUI improvements are secondary.
I'd be interested to hear the reasoning behind "That error is probably his RAM or CPU running at too low or high of a voltage," as neither cause would fit his symptoms. Grimmy's question about GPU/CPU temps at least fits the symptoms, where the drastic slowdown comes during gameplay, while non-game benchmarks seem to be within 2x of expected values.

Edit: OK, now I read the last page of posts. Slow deterioration and increasing instability of the system suggests the following possibilities to me:
1) Cooling for one or more key components (e.g. CPU, north bridge, PWMs) is failing, for example due to dust accumulation or improper heat sink mounting (very common on LGA775 builds). What are your CPU, system, and PWM temps under idle in the BIOS, under idle in Windows, and as you run Orthos?
2) Key system files are becoming more and more corrupt, perhaps due to memory errors. Have you run memtest86+ (not to be confused with "memtest") through at least a couple of complete passes without any errors?

 

[Kano523]

Alright I have another screen shot that should hopefully prove useful in diagnosing this problem, I ran the SiSoftware Sandra Lite CPU arithmetic test with CPU-Z and Core Temp opened. As you can see this screen shot was taken during the test and the multiplier does not increase to x9, however I should mention CPU-Z did get grayed out for the beginning of the test like it was having trouble refreshing so it might have been trying to display the change but wasn't able to. So you all be the judge if my CPU is performing well and at proper temperatures during the benchmark.

 

[13thmonkey]

Thats too hot. Much too hot.


Probably a bad heat sink seating, very very common.

 

[The_OGS]

LoL - Kano523 your CPU appears to be throttling as one core reaches 85c.
At least in the good old days, the thing would just catch fire and burn up, eliminating all the endless speculation... heheh.
There are many ways to measure temps - BIOS, tools, mobo utilities etc.
I thought you assured us (last page) all your temps were 'fine'...?
Run the thing on its side, with the case open. Is your CPU cooler hot?
If it is not getting hot, it is improperly mounted.
Note: OEM cooler compound is a 1-time thing. If you remove the cooler (or even move it, breaking the seal) you must clean it all up and reassemble using new compound.
Supposedly there is even better stuff but Arctic Silver works well for me.
Back in the early 2000's we joked about all the new thermal throttling technologies - how CPU coolers were going to become completely trashed over time, fans not even spinning, and calls would start to come in:
"Something's wrong with my computer. It seems kinda... slow."
LoL! The truth it seems is nowhere near as humorous though.
Check out your cooler and get back to us,
Regards

 

[Kari]

yep, it is throttling it down to avoid complete silicon failure.. 82-85C is BAD

reseat the heatsink with new thermal compound...
make sure you removed the plastic from the bottom of the heatsink, that doesnt conduct heat very well

edit: the_ogs was little faster than me..

 

[monsterrocks]

Mondoman-

"mr - you may want to consider doing more research before posting so as to improve the correctness of your posts. BTW, Orthos is *not* the exact same as PRIME95; rather, it is a shell that allows one to automatically run two copies of PRIME95, one per CPU core. The GUI improvements are secondary.
I'd be interested to hear the reasoning behind "That error is probably his RAM or CPU running at too low or high of a voltage," as neither cause would fit his symptoms."

Wow, Orthos says on overclockers.net that someone took the EXACT engine from prime and updated the UI. There is a multicore Prime95 to. So while orthos made that change first, they originally used the EXACT Prime engine. Do some research....
As to the other thing...lol. That was not to this guy. Some guy said "I'd be interested to know what orthos error this is" and provided a link. The error was as follows: "Error, rounding was 0.5, expected less than 0.4." So that was what that voltage part was too. Do us all a favor and read what people write in context, not just what they write...

 

[Mondoman]

...So while orthos made that change first, they originally used the EXACT Prime engine.

My point was just that *using* the exact Prime95 engine is not the same as *being* the exact same as Prime95. Reread my post for clarification.

...Some guy said "I'd be interested to know what orthos error this is" and provided a link. The error was as follows: "Error, rounding was 0.5, expected less than 0.4." So that was what that voltage part was too. ...

Thanks for the background info; I'd still be interested to hear the reasoning behind "That error is probably his RAM or CPU running at too low or high of a voltage."

PS - the "envelope with left arrow pointing to it" icon at the bottom right of a post allows you to quote the post in your reply, with formatting to clarify what is quoted and what is your reply.

 

[Mondoman]

Kano, check out steps 2.5 onward in this post for details of how to make sure your LGA775 CPU heat sink is fully seated (there are a few tricks): http://www.tomshardware.com/forum/194385-31-part-assembly#t1184072

 

[monsterrocks]

My point was just that *using* the exact Prime95 engine is not the same as *being* the exact same as Prime95. Reread my post for clarification.


Thanks for the background info; I'd still be interested to hear the reasoning behind "That error is probably his RAM or CPU running at too low or high of a voltage."

PS - the "envelope with left arrow pointing to it" icon at the bottom right of a post allows you to quote the post in your reply, with formatting to clarify what is quoted and what is your reply.

About the prime thing...point taken. I guess I really didn't stop to think about the fact that I could have figured that out the whole time. Wow, now I feel like an idiot. I was disproving myself in my response by the very fact that it uses the same ENGINE...
As far as the voltage thing goes, I get that error message when I overvolt my RAM or undervolt my CPU. When I change the voltages and do tests, certain voltages do better and better until I get the right one. Then the error goes away, no matter how long I run Prime95. That is why I said it. Although it can also mean you have too high of a CPU clock; or so others have said on various forums at least. The voltage thing was from my personal experience; I don't know what else the message might mean.

 

[Zorg]

I was afraid that your temps weren't what you thought they were, that's why I asked you to post the pic. Mystery solved, fix your HS. I'm not going to go back and look, but if you are using the stock HS throw the POS in the trash. Get a nice HS with a backplate and install it. Here is a review that has some good ones in it. Pick one, install it and move on.

AnandTech: Thermalright Ultra 120 Extreme: Is More Better?

Oh, in case you didn't already know this,

DO NOT BOOT THE PC UNTIL YOU FIX THE TEMP PROBLEM YOUR CPU IS ON FIRE. [:zorg:2] [:zorg:2]

 

[firetatoo]

your e6600 is a 1.6ghz?
i thought that was 2.6ghz chip.

 

[monsterrocks]

80c is really bad. I think toms did a test to see how hot a CPUs silicon has to get before it melts and it was something like 130c, but I am surprised you haven't burnt that chip into the ground. Even if you get a new coller, that chip might not last to long. With temps like 80c, you can do some pretty serious damage (which is permanent) to those chips...If you CPU hits in the 80s, it probably throttles down to speeds you wouldn't believe.

 

[Kano523]

Yup, the HS was loose....I now must begin the long and painful battle with my own incompetence. Though however moronic I may indeed be the problem is fixed and performance is vastly improved. I ran SiSoftware Sandra Lite and my CPU scored right where it should be, with the CPU temp being 50C during the test. I also ran 3DMark06 again and got these scores:

3DMark Score 9207 3DMarks
SM 2.0 Score 4306 Marks
SM 3.0 Score 4103 Marks
CPU Score 2162 Marks

I really appreciate everyones help in solving this problem, you were all helpful and informative. I'm glad the problem is fixed, but boy do I feel like an ass. Nevertheless I have some gaming to do, thanks again!

 

[Grimmy]

:lol: . o O (82-85C? yep.. dat do it)

Glad we all could help ya.

 

[Mondoman]

Although not immediately a problem, 50C still seems somewhat high for your CPU. Did you remove the old thermal compound and apply fresh thermal compound when you secured the hsf? Did you make sure all 4 pins were turned all the way in the direction *opposite* the arrow, and that they went all the way through the motherboard?

 

[Bik3r]

Er... have you run DXdiag?

Just run it quickly and make sure direct draw and direct 3d acceleration are enabled... your results are consistent with D3D running in software. Fingers xxed ;-)

 

[gwolfman]

your e6600 is a 1.6ghz?
i thought that was 2.6ghz chip.

2.4GHz

 

[spongebob]

50C still seems somewhat high for your CPU.

50C under load... might be high, might not. Ambient room temp. can have a significant impact.

 

[Grimmy]

Yup, the HS was loose....I now must begin the long and painful battle with my own incompetence. Though however moronic I may indeed be the problem is fixed and performance is vastly improved. I ran SiSoftware Sandra Lite and my CPU scored right where it should be, with the CPU temp being 50C during the test. I also ran 3DMark06 again and got these scores:

3DMark Score 9207 3DMarks
SM 2.0 Score 4306 Marks
SM 3.0 Score 4103 Marks
CPU Score 2162 Marks

I really appreciate everyones help in solving this problem, you were all helpful and informative. I'm glad the problem is fixed, but boy do I feel like an ass. Nevertheless I have some gaming to do, thanks again!

He's already taken care of it. From that last screen cap, his load temps on CoreTemp were 82-85C. So the throttling was the cause of his low scores.

 

[cpburns]

in the rare case that his cpu's cache is dying, i could understand this kind of performance hit. it does look like a cpu problem. i would try RMAing the chip, see if that fixes things. if it doesn't, i would say move onto the motherboard, and then ram, and then power supply, and then video card.

i know you don't want to waste time RMAing things forever. i wouldn't either. but it might be what you're forced to do at some point.

i would test the cpu by turning off speedstep, setting a 9x multiplier, and slightly overvolting the processor. see if it turns stable. maybe it won't. is your cpu in good physical condition?

again, i'd RMA the cpu and see how the new one performs. it will only cost you shipping and some time, and it will rule out or confirm the cpu being at fault.

good luck!

Edit:
hehe, i wasn't going to suggest the heatsink. figured you'd have checked that O

good to hear you're back up and running though.