sigh... very... very... unhappy

[grayg1]

ok

No one was saying you are stupid. You need to stop snapping though, or people are going to lose interest.
It was humour aimed at myself

Drivers, including the latest chipset drivers from the intel website, the latest graphics drivers from the nvidia website, the latest lan drivers from gigabyte.
Tried

Clean the old drivers with driver cleaner before you install the latest ones.
Tried

Try disabling the jmicron controller if you dont have an ide hard drive/dvdr.
Tried

Take off the NB heatsink and check it is making contact properly.
Unlikely because the same problem exists over 2 different motherboards? / would prefer not to

Manually set the correct voltage and timings for your ram.
My RAM is value selct so it uses a lower voltage, all timings/voltage auto in correctly. I have tried running PC4200 and PC5300 with them though and have also tried bumping the voltage up by 0.1v

Try the ram in different slots.
Tried

Try the ram one stick at a time in all of the slots.
Tried

Run memtest on each stick of ram individually.
Could this make much of a difference if both pass memtest together 9 times in a row?

If you have a sound card, take it out, if you are using on board sound, disable it in the bios.
I am using onboard and have already tried disabling in BIOS and also not having it installed.

EDIT: looking back there was no smile in my orginal comment and well this being just words rather than real speach there's no real way to tell whether it's a joke or not

[Clunk]

Yes, set them to their rated speed and voltage in the bios and run each stick for about 5-6 hours or evernight on test 5.

[Jay]

can he not just take it in and see for yourself? (or are you too far away?)

you could then be happy looking at the lovely girl on the front desk while they sort it for you!

lol

[grayg1]

I'll do the tests tomorrow, hopefully Okami surprisngly turns up through the post, if not it's going to be a boring day tomorrow... Don't like leaving the PC on overnight. Besides it sounds like jet engine and it's in the same room I sleep in

hhmm does jay have a slight crush on someone at Scan perhaps?

[Jay]

hhmm does jay have a slight crush on someone at Scan perhaps?

me?

Never!

[grayg1]

hhmm now I haven't tested them seperate yet but I will

I just did test 5 with both RAM sticks it passed the first test and then got 108 errors as soon as test 2 began basically. EDIT just realised it reports the errors when the pass has finished.
Bare in mind I have tested the RAM seperately already when playing games and get freezes with either stick.
Now I read up on test 5 and this tests RAM and the memory controller? 2 different motherboards with a faulty memory controller seems unlikely really? I have been on Corsair website and checked and they say the Corsair Value Select PC5300 is compatable with the Gigabyte 965P-DS3P.

I have been thinking for a while now that the CPU was the cause as it is the only thing I can't really test by changing. What do you think?
Or maybe Corsair are wrong and haven't really tested well enough?

I will go test both sticks seperate now.

[grayg1]

Test 5 seems to take about 1 minute per GB of RAM. I only did 5 tests per stick each but there was no errors but together they get 108 errors in the first test and it is consistant too because I reran it with both sticks and it got 108 errors again on the first pass.

[mike306dt]

TAT (which records temperatures much higher than any other temperature tool) says CPU is 65C 100% load both cores.

TAT reports your true cpu temperature from the diodes though doesn't it. So this is far more accurate than the bios temperature display.

[grayg1]

Yes that's what I mean though it reports much higher temps than other temp software. Also there is no test you can do that makes the temperature go as high as the TAT 100% load does.
Anyway back to the RAM testing...

I just did 101 passes with 1 of the RAM sticks no errors and before I used memtest on it I played C&C 3 and it froze within 30 seconds.

I'm thinking memtest won't find errors on the sticks seperately but will only when together for some reason. Though obviously having both sticks running together is not the cause of freezing.

Phillip you have anything to suggest?

[Clunk]

Are you getting errors on just one of the sticks, in all of the slots, or one of the sticks in one of the slots?

I still think this is PSU related.

[grayg1]

It seems I get errors only when the sticks are together. Like I said though 101 passes no errors for 1 stick on it's own and it still freezes though.

I tried another PSU previously a Tagan 420watt with the 8pin eps and I used it's own power lead. I can take off my case fans, floppy, DVD and CDRW drives and it makes no difference either (just trying to reduce power load).

If you're perhaps unsure of my Coolergiant (like what it is capable of) I think it is a Noisetaker on the inside, just a more fancy look.

[Clunk]

I'm not doubting the PSU's power abilities, but from what you have described, everything points to a faulty/flakey PSU and/or bad ram.

I would also have checked the motherboard, as its not uncommon to have faulty memory slots, but as you have tried 2, its not that.

Have you actually tried ringing scan to see what they can and cant test for you?

If scan wont do it, take the whole lot to a local PC shop and get them to test the stuff.

I would certainly start by borrowing some new ram, or at least trying the stick that errors, in another pc to see if it still does it.

[grayg1]

I wouldn't like to suspect the motherboard though because over 2 different motherboards same problem? Both having dodgey slots?

Scan took basically my whole computer last time and they said the graphics card was causing the freezes after their testing. They replaced the graphics card and said the problem is solved. The problem is I told them I know it isn't the graphics card that causes it because I have already tested using a different graphics card and it didn't fix the problem. The manager chose to overule me by saying they had done long testing and they are confident the problem is fixed.

Now I get it all back and the problem isn't fixed, I mean what exactly am I supposed to be requesting they do? Take it all back and test it again? If I knew what was wrong exactly I wouldn't be still in the same situation. I did tell them I haven't been able to swap the CPU with anything else and I haven't been able to swap the RAM with a different model.
Perhaps I should request exactly what they did during the testing if possible? Like did they follow my notes and swap the CPU out or the RAM?

I didn't deny that the graphics card maybe faulty to them you know? I said it may well actually be faulty but it isn't the cause of the problem.

[IAmATeaf]

Have you checked windows event viewer to see if there any entries and what voltage do you have the RAM set too.

[Clunk]

Well, for what its worth...

I still reckon that your PSU *could* be faulty, the fact that you have a faulty stick of ram and have had a faulty graphics card just reinforces this.

If you bought all the stuff from scan, return the lot to them and Im sure they will test it for you.

Make sure they do exactly what you do, when it freezes, and then they can use their own troubleshooting methods from there in.

[grayg1]

No entries related to crashes etc in event viewer. RAM voltage is 1.8v (stock) have tried 1.9v.