i know, its an expensive PSU, but as it is a 510w unit from anyone would barely be coping

Quote from Jakg :

i know, its an expensive PSU, but as it is a 510w unit from anyone would barely be coping

theres no way that a pc with his speacs would consume more than 300-350w so unless he has a no name one like lc power he should be fine

Quote from srdsprinter :

I'd hate to recommend him a new $150 PSU and find out that it isn't the problem...

@ Devil, he has a 6400, which has been discontinued since the 6420's release. I would recommend he looks up the stepping he has to see if others have issues. He's got a good cpu cooler.

you dont have to explain me the CPU version.I have more then good overview.6420 has 4MB L2 memory - nothing new for me.

I am still failing to see what his real issue.I see only 2.
Too high tempreture and too much overclock.

Possibly bad stepping and also EIST might be enabled in BIOS.

Quote from Jakg :

i know, its an expensive PSU, but as it is a 510w unit from anyone would barely be coping

Jak dude, thats rubbish and you know it...

I've seen tests done with a (stock admitadly) Core 2 Quad and a GTX running on a 380W PSU, possibly 350W, can't remember now.

I'd be wary of saying the PSU is at fault.

PSU:
I can measure how many Watts are being used since I have a voltagemeter(?) attached to my power outlet:

Idle: 390-400W
TAT @ 100% Load: 442W
AquaMark3: 435W
LFS Vixen 01: 460W

Seems high? It's because it's also measuring power drawn by my 22" CRT monitor :P

LFS Vixen 01 w/o monitor: 360W

Those are the levels I'm peaking at.

Was looking for EIST in BIOS, but couldn't find it - I've looked before too - also looked for Speed Step, but they just can't be found anywhere, even though EIST is mentioned in the P5B Deluxe manual

I'm on latest BIOS v1101 - and checked the latest PDF manual.

I've noticed something before, but wanted to google it to see what is is: Vanderpool Technology - before I found out anything about it, I stumbled upon a thread with ppl listing their BIOS settings. I noticed one mentioning he's PCI-E rate would get reduced from 16x to 1x if he crossed a certain amount of FSB (37x-ish)

This is what mine looks like:
http://www.digital-ray.dk/LFS/pci-e_speed.gif


Back to the Vanderpool - according to this, I can just ignore it - should probably just turn it off, if only I could. It's ghosted.. hm.

I'll see what I can come up with regarding the reduced PCI-E rate, and then back to the the questions, comments etc. But it does give a better answer now, as mentioned earlier RAM and CPU speeds seems fine, but when gfx was involved, I would lack performance.

Thanks to my search for Vanderpool, I stumbled upon the PCI-E @ 1x rate issue.

Problem is now solved See the updated post #1.

But what I don't understand is..... why is my temperature readings lower now?

Hey! Glad you got it sorted out! Very interesting the pci-e speeding down.

My only haphazzard guess as to lower temps is that either the cpu was somehow trying to pick up slack for pci-e X1 gpu, it was somehow picking up extra voltage, or the gpu wasn't running its fan properly due to X1 pci-e speed. I doubt any of those are right.

@Devil - i figured you knew 6400 discontinued, but you had said the "latest batches" or something, which made it sound like you thought they were still making them.

Well you can still buy E6400 is some shops and as I said some batches of latest C2D generally are not so good overclockers as were the initial ones.

The E6420 doesnt bring too much with 4MB cache.Like 3-7percent but still something.

r4ptor,
I am gald you sorted it.Pretty interesting bug has this mobo.
I cant help but I heard this year a lot of weird issues with ASUS mobos.DFI is hard to spot now and same with ABIT.This is getting really confusing as the compnaies which had good reputation with their mobos having such a issues.

It is always hard to guess just from the screen in the other part of the world what the issue might be.Looking directly at the PC is always easier as you can check more things quicker.I think we would found that issue with PCI-E running just at 1x pretty quickly in normal condiditon.There are programs for checking these things.

Cheers

Quote from srdsprinter :

My only haphazzard guess as to lower temps is that either the cpu was somehow trying to pick up slack for pci-e X1 gpu, it was somehow picking up extra voltage, or the gpu wasn't running its fan properly due to X1 pci-e speed. I doubt any of those are right.

GPU voltage is the same, so there should be any difference because of that. But dunno..

Quote from DEVIL 007 :

I am gald you sorted it.Pretty interesting bug has this mobo.

...

It is always hard to guess just from the screen in the other part of the world what the issue might be.Looking directly at the PC is always easier as you can check more things quicker.I think we would found that issue with PCI-E running just at 1x pretty quickly in normal condiditon.There are programs for checking these things.

Yeah, sure.. but I never suspected the PCI-E to slow down like this I had CPU-z and could have had seen it all along.. but never looked for it.

I would now like to turn this thread to discuss temperatures. I read quit a bit on the issue last night, and although I got a good clue to what to do and how to read temperatures, I'm still just uncertain about it and especially afraid that I might read them wrong.

Apparantly, I can use CrystalCPUID to get an exact DTS reading on my core temp.

I can then use this reading to verify core and cpu temps displayed in various monitor programs - SpeedFan can do offsetting, so I'll be using that - guide to how to do this.

The core temp. displayed in SpeedFan matches, so no offsetting needed - however my CPU temp should be ~15 degrees lower according to this, but it's not - it's only 5-6 degrees lower.

After offsetting my CPU temp with -9, I end up @ 39-40 degrees with 54/52@ cores as I'm writing this right now.

I did a thermal test last night. Core went up to 79-80, while the cpu temp (before offsetting) went up to 67.

Nothing happened - no throttling, no crashes, no nothing.

With the -9 offsetting it would give more sense to why nothing happened, as that would return a cpu temperature @ 58-59 degrees.

Also, how buggy is SpeedFan? I started an Orthos test when I was about to head for bed, but SpeedFan later displayed that my CPU was @ 89 degrees, while the hottest core was @ 55. After restarting SpeedFan, it still showed same values, and only after a computer restart I would get proper readings.

Perhaps an instability glitch caused by the Orthos test? It didn't return any errors though.

Also, how do you guys measure ambient temp - and does it matter at all to know it?

Ambient - I use a clock on my desk - it projects the time, tells me the date, and that my room is 23.4°C

Ambient = not that useful (unless you've got temp issues), 21°C and expect my CPU to be at 42°C - 26° Ambient and i'm looking at 50°C CPU temp, but i have a Zalman Reserator crappy water cooling, and it's cooled purely by the air around it - which explains my massive temp gaps

Yeah, I suspected it doesn't really matter - just another variable that in best case further explains a thing or two.

I don't know what ambient is today, but it's extremely hot all of the sudden :| But I can't see this being reflected as higher CPU temp.

it really does depend - on my old cooler it didn't matter (because i had some "bitching" airflow on that thing for testing) but on this it really does

Doing another Orthos test and just noticed that my GPU is @ 74 degrees - top of my computer case is pretty hot, actually. Is a high GPU temo normal during CPU stress tests? Seems odd to me.

R4ptor, take a look at this thread. That guy found his 6600 was build like a satelite dish, with a massive dip in the center. He ended up lapping it, and now sees temps of 30 degrees under load with a Tuniq Tower, really quite amazing.

If your top of the case is pretty hot then the CPU must be REALLY hot.
I said it from start that the temps are really bad.

What kind of CPU cooler you have.I must have missed that in the previous posts.
This will sound kidna silly but I have good experience when you put your fingers on the bundled coller and cant keep them on the cooler for more then 3 sec the overheating is the real issue.

If you will continue like that with higher overclock,raised vcore the CPU will die pretty soon.

Quote from pb32000 :

R4ptor, take a look at this thread. That guy found his 6600 was build like a satelite dish, with a massive dip in the center. He ended up lapping it, and now sees temps of 30 degrees under load with a Tuniq Tower, really quite amazing.

Thats insane! I will definitely have a look at how flat my E6400 as well as heat sink are - but wont be able to do it for another 2 weeks or so.

Quote from DEVIL 007 :

If your top of the case is pretty hot then the CPU must be REALLY hot.

I said it from start that the temps are really bad.

I've been running Orthos for almost exactly 4 hours now at priority 4 - I've also just finished watching a movie .. on the computer :P

Core Temp and SpeedFan are telling me that both cores are at 65-66 (TAT @ 64-65/62-63.. I wont be using TAT anymore. I've read several places that it's unreliant, which I can see now).

GPU temp is currently @ 78

Ambient temp is still high (it's a really hot day/evening today.. but I've experienced worse).

I don't want to be stubborn, ignorant or anything else, but I really don't think my temps are too high.

Except for the GPU temperature - I never monitored it before, so I don't have much reference to go by, other than placing my hand on the case. I did that last night too, but it surely wasn't this hot - it feels like as if I've been using the GPU for hours.

I noticed my GPU temp last night when I enabled monitor for it, and it was idling @ somewhere between 58-66. I didn't pay too much attention to it though.

If I place my hand on the side of the case next to the GPU, then I feel a little less heat here, but placing by hand where the CPU is, the temperature drops allot.

Quote from pb32000 :

What kind of CPU cooler you have.I must have missed that in the previous posts.

I'm using a Scythe Infinity, but have replaced the stock fan with Enermax Warp UC-12AEBS running @ high 1100 RPM (lowest RPM).

It's mounted on one side of the heatsink, sucking air into/through it. On the other side of the heat sink there as 4cm gap, and then another 120mm fan mounted at backside of the case. This one is blowing air out/away from the case/cpu.

Quote from pb32000 :

This will sound kidna silly but I have good experience when you put your fingers on the bundled coller and cant keep them on the cooler for more then 3 sec the overheating is the real issue.

The cooler is midly warm at the top, and a bit warmer at the base. I could hold my fingers on it for an eternity.. (yeh, I know - doesn't mean it's a good sign.. especially since I'm suspecting that my CPU is uneven - but.. monitor doesn't show THAT hot temp.. or is my perception at fault here?).

Quote from pb32000 :

If you will continue like that with higher overclock,raised vcore the CPU will die pretty soon.

Yeah, I know But I won't be continuing like this. I'm actually running @ 3GHz right now because that issue with the PCI-E seems to be random. I fix it.. it works.. I reboot and then need to fix it again. happened 3 times so far after I learned a bit about what I could do about it, and I don't want to keep doing it.. I'll just wait for the BIOS fix.

Thought about RMA'ng the board and saying about the PCI-e glitch? worth a try if you can live without the board for a week or so

Quote from Jakg :

Thought about RMA'ng the board and saying about the PCI-e glitch? worth a try if you can live without the board for a week or so

I can't For the same reason why I won't be looking closer at the CPU. I work home and it would be bad timing to do it now as I'm in the middle of a project.

Update on the GPU temerature:
Downloaded and installed RivaTuner - manually set the fan to 100%, and temperature is now 62 degrees after only few minutes. Top of case is already much much cooler.

I'm wondering what happened? I played with some fan speeds in uhm.. SpeedFan, but I returned all values back to 100% I just verfided and they are indeed @ 100% - I didn't even notice the GPU fan being affected when I tried to see what fans I could control from there

you cant control the GPU temp or fans in SpeedFan

Quote from Jakg :

Thought about RMA'ng the board and saying about the PCI-e glitch? worth a try if you can live without the board for a week or so

Actually, I'm just just thinking I wouldn't be able to do so at all. They typically ask how to reproduce the error, and all I've seen so far is that it only occurs when attempting to overclock. Don't think they would accept that as a RMA reason

Quote from Jakg :

you cant control the GPU temp or fans in SpeedFan

Yeah, I know.. just saying that that was all I did - I'm puzzled about why the GPU fan would get lower.

Ps. GPU @ 60 degrees

Yer sadly projects/work mean messing about with my main rig has to be put on hold, I just cannot run the risk of failure atm.

GPU temps seem fine and dandy, my 8800 gets almost too hot to touch the heatsink under load, and it heats up the case a lot around where it is and where its bolted to the chassis.

I very briefly took my 6600 to 3ghz on the same mobo as you and everything was fine, as you say it isnt a 100% certainty that that problem will occur, I hadn't heard anything about it before you found out about it.

Anyway otherwise everything seems to be good. Ambient temps have such a big effect on my CPU temps. At the moment its pretty warm in my room, been a warm day and it's idling at 35 degrees, however normally in a cool room i'll see about 28 degrees idle, and if I leave my window open at night with the cold air directed into the pc case lol, I see about 24 degrees. So yer, there's a rambling useless piece of imformation for you

Quote from r4ptor :

Actually, I'm just just thinking I wouldn't be able to do so at all. They typically ask how to reproduce the error, and all I've seen so far is that it only occurs when attempting to overclock. Don't think they would accept that as a RMA reason

Why not? explain that on the box it brags about the overclocking prowess of the board and yet on trying it, it screws everything up.

Worth an email or two at least.

pb32000:

I'm quiet satisfied with it for now - when I give it another look, I think I'll also have replace the paste on the various smaller heatsinks.

jakg: Year, doesn't cost anything to ask.. so I'll just do that