Overclocking Q6600

Hi

I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I
noticed the PSU got very hot, so I decided to just run it at stock. It was
just a cheap 550w PSU.

Recently I bought a branded 700w PSU, it doesn't hardly get warm. Around
the same time I replaced the northbridge heatsink with a different branded
heatsink for overclocking, as the stock one was getting very hot at 400 fsb,
this one hardly gets warm. I also used some differant thermal paste on the
CPU.

I've found out by looking at CPU temps that the thermal paste is rubbish,
its hotter now than it was before, it was very old and gloopy and I probably
put way too much on.

I also put 4gb (1gb sticks) of memory in, I saved them from an old computer.

Anyway running prime 95 it would lock up windows at stock CPU settings, I
also ran memtest and it crashed after 11mins, no errors detected, just
locked up. So I thought it must be the memory, I took out 2 sticks and
tried Prime95, it got to the second test, so I took out those 2 and tested
the other 2 and it passed too. Then I tried some OCZ (other brand is
kingston value) it gets to the second test too. So I thought maybe it just
doesnt like 4gb, I guess I can live with 2gb.

Then I try it at 400fsb and 3.2ghz and it freezes up again soon after
starting the test, then I just came back down to stock again ran the test
again, and after it gets to test 3 it freezes up.

I use Core Temp 0.99.3 to measure the temp, at stock it goes up to about 53c
and overclocked to 61c, quite high but it is with this rubbish thermal
paste. Could this cause the computer to freeze?

I have no temp warnings set on the motherboard, so I doubt this is a
motherboard feature that stops the CPU from overheating. I'm now using the
memory I have been using since I made this computer (OCZ), when Prime95 was
working fine, so maybe its not the memory, but I'll give it a memtest
anyway.

I've played games and encoded a 2 pass xvid for over an hour (at stock) and
it didn't lock up but even the encoding doesnt get close to 100% load on 4
CPU's

Sorry if its a bit confusing!

'Random' wrote, in part:
I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I
noticed the PSU got very hot, so I decided to just run it at stock. It
was just a cheap 550w PSU.
_____

Your reported temperatures for the CPU are quite good; very low, in fact,
for air cooling with a Q6600 @ 1.4 volts and 3.2 GHz. The 1.4 volts seems a
bit high, especially for a mere 3.2 GHz.

"Quite hot" is not a very useful term - it really isn't a precise enough
basis on which to make decisions. Certainly not enough to warrant changing
components. Check through prior posts in this newsgroup for threads on
overclocking a Q6600, especially those by 'Ed Medlin'.

You haven't mentioned WHAT thermal paste you used. But never mind, there is
VERY little difference in the performance of one compared to another, even
butter does pretty well for the short run. There probably was no reason to
replace your NorthBridge heatsink. A good check to now: Overclock the FSB
to 400 MHz, but set the multiplier for the CPU to lower than normal; this
should give a differential diagnosis as to whether the problem is with the
new NorthBridge heatsink (or its installation.) Or you could reinstall the
original NorthBridge heatsink. It could be that the new heatsink is not
particularly flat and smooth. Or you could have incorrectly applied thermal
paste, or incorrectly installed the heatsink.

You haven't posted your motherboard either, a very important bit of
information for diagnosis. Nor your ambient room temperature, also
important. Nor your highest successful overclock speed test in your new
configuration. Nor what core voltages you have tried. Nor, in fact, why
you violated the first rule; if it ain't broken, don't fix it. So, among
other things, you might just try restoring everything to your original
successful overclocking setup.

Phil Weldon

"Random" wrote in message
...
Hi

I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I
noticed the PSU got very hot, so I decided to just run it at stock. It
was just a cheap 550w PSU.

Recently I bought a branded 700w PSU, it doesn't hardly get warm. Around
the same time I replaced the northbridge heatsink with a different branded
heatsink for overclocking, as the stock one was getting very hot at 400
fsb, this one hardly gets warm. I also used some differant thermal paste
on the CPU.

I've found out by looking at CPU temps that the thermal paste is rubbish,
its hotter now than it was before, it was very old and gloopy and I
probably put way too much on.

I also put 4gb (1gb sticks) of memory in, I saved them from an old
computer.

Anyway running prime 95 it would lock up windows at stock CPU settings, I
also ran memtest and it crashed after 11mins, no errors detected, just
locked up. So I thought it must be the memory, I took out 2 sticks and
tried Prime95, it got to the second test, so I took out those 2 and tested
the other 2 and it passed too. Then I tried some OCZ (other brand is
kingston value) it gets to the second test too. So I thought maybe it
just doesnt like 4gb, I guess I can live with 2gb.

Then I try it at 400fsb and 3.2ghz and it freezes up again soon after
starting the test, then I just came back down to stock again ran the test
again, and after it gets to test 3 it freezes up.

I use Core Temp 0.99.3 to measure the temp, at stock it goes up to about
53c and overclocked to 61c, quite high but it is with this rubbish thermal
paste. Could this cause the computer to freeze?

I have no temp warnings set on the motherboard, so I doubt this is a
motherboard feature that stops the CPU from overheating. I'm now using
the memory I have been using since I made this computer (OCZ), when
Prime95 was working fine, so maybe its not the memory, but I'll give it a
memtest anyway.

I've played games and encoded a 2 pass xvid for over an hour (at stock)
and it didn't lock up but even the encoding doesnt get close to 100% load
on 4 CPU's

Sorry if its a bit confusing!

Random wrote:

I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I noticed the PSU got very hot, so I decided to
just run it at stock. It was just a cheap 550w PSU.

My Q6600 G0 didn't like over 333 MHz in two different motherboards that
run my E6400 at 8 x 440. There's a lot of current flowing at a low voltage,
and a substantial voltage drop under load is very common. Frequently, four
sticks of RAM often won't run as fast as two. Also, I have had two sticks of
RAM go bad on me with this setup. Checking your memory with a lowered
multiplier, are we, hmm?

I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I
noticed the PSU got very hot, so I decided to just run it at stock. It
was just a cheap 550w PSU.

My Q6600 G0 didn't like over 333 MHz in two different motherboards that
run my E6400 at 8 x 440. There's a lot of current flowing at a low
voltage,
and a substantial voltage drop under load is very common. Frequently,
four
sticks of RAM often won't run as fast as two. Also, I have had two sticks
of
RAM go bad on me with this setup. Checking your memory with a lowered
multiplier, are we, hmm?

I did a check the other night of the memory I'm using now, OCZ 2GB DDR2
PC2-6400 Platinum Revision 2 Dual Channel, it got 3 full passes in memtest
86 so I think its fine. This is with the processor at stock which is a 9x
multiplier and 266mhz bus. It still locked up Prime 95 and it seems to do
it at exactly the same time. I've ordered some new thermal grease. I might
refit my old PSU and heatsink and see if it works again but I wouldn't be
suprised if I've slightly damaged my CPU or Motherboard so that it doesn't
overclock anymore. I was also trying to overclock my graphics card at one
time and it crashed the graphics driver a few times. Bit frustrating as
with stock CPU and graphics I get 12,500 3D 06 marks and overclocked (was
stable) CPU @ 3.2ghz and graphics 15,200 3D marks, at the start of the demo
its 76-78fps at stock and 96-98 overclocked, this is with a 8800 GTS 512 G92

"Phil Weldon" wrote in message
m...
'Random' wrote, in part:
I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz by
having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but I
noticed the PSU got very hot, so I decided to just run it at stock. It
was just a cheap 550w PSU.
_____

Your reported temperatures for the CPU are quite good; very low, in fact,
for air cooling with a Q6600 @ 1.4 volts and 3.2 GHz. The 1.4 volts seems
a bit high, especially for a mere 3.2 GHz.

"Quite hot" is not a very useful term - it really isn't a precise enough
basis on which to make decisions. Certainly not enough to warrant
changing components. Check through prior posts in this newsgroup for
threads on overclocking a Q6600, especially those by 'Ed Medlin'.

You haven't mentioned WHAT thermal paste you used. But never mind, there
is VERY little difference in the performance of one compared to another,
even butter does pretty well for the short run. There probably was no
reason to replace your NorthBridge heatsink. A good check to now:
Overclock the FSB to 400 MHz, but set the multiplier for the CPU to lower
than normal; this should give a differential diagnosis as to whether the
problem is with the new NorthBridge heatsink (or its installation.) Or
you could reinstall the original NorthBridge heatsink. It could be that
the new heatsink is not particularly flat and smooth. Or you could have
incorrectly applied thermal paste, or incorrectly installed the heatsink.

You haven't posted your motherboard either, a very important bit of
information for diagnosis. Nor your ambient room temperature, also
important. Nor your highest successful overclock speed test in your new
configuration. Nor what core voltages you have tried. Nor, in fact, why
you violated the first rule; if it ain't broken, don't fix it. So, among
other things, you might just try restoring everything to your original
successful overclocking setup.

Phil Weldon

Peanut butter (creamy) also works quite well but get a bit "crusty" after a
few weeks....:-) Like Phil said, your temps are very good for that vcore.
3.4+ should be possible using that high a vcore, depending on other factors
like your NB chipset ect. With an Asus Striker Extreme (NV 680i chipset) I
get 3.2Ghz with only 1.2vcore and no temperature issues at all with liquid
cooling. Properly installing your thermal paste is very important. Over at
the Arctic Silver website they give a good example on applying it correctly
for the C2 Quads. I don't think their compound is much, if any better than
plain Radio Shack compound, but the instructions are good and do work.
I am getting ready to start a project a bit early because of a SATA
controller issue on this system in which I have lost 2 channels out of 6 and
one took out a HDD with it. This has been an ongoing issue with the Asus
680i boards for me. I just ordered an i7 920 processor and an Asus P6T
Deluxe MB and will see how far the Nehalem will go with just air cooling. I
am going bottom of the line with the processor for right now at under $300
at Newegg. I am seeking overall performance on this build and ordered 6 gigs
of Corsair XMS3 DDR3 PC1600 for 3 channel operation and will use Vista
Ultimate 64bit as my OS. I will recycle my two 8800 GTX cards in SLI and
will probably move a single 8800 GTS I still have over to this system when I
put in a spare P5N32-SLI MB if it shows signs of any more issues. I am
shooting for mid-3Ghz for the 920 if possible. I guess someone has to be the
"goat" with these processors....:-). I would have rather waited a few
months, but with the issues I am having, I don't trust this system much
longer. It has been an excellent performer, but has been disappointing as
far as my standards go with reliability.

Ed Medlin

I've had this processor for a while now, its the one with G0 stepping so
overclockable. It used to run it fine at v1.40 core voltage at 3.2ghz
by having a 400 fsb and 8 multiplyer, it passed Prime 95 for 2 hours but
I noticed the PSU got very hot, so I decided to just run it at stock.
It was just a cheap 550w PSU.
_____

Your reported temperatures for the CPU are quite good; very low, in fact,
for air cooling with a Q6600 @ 1.4 volts and 3.2 GHz. The 1.4 volts
seems a bit high, especially for a mere 3.2 GHz.

"Quite hot" is not a very useful term - it really isn't a precise enough
basis on which to make decisions. Certainly not enough to warrant
changing components. Check through prior posts in this newsgroup for
threads on overclocking a Q6600, especially those by 'Ed Medlin'.

You haven't mentioned WHAT thermal paste you used. But never mind, there
is VERY little difference in the performance of one compared to another,
even butter does pretty well for the short run. There probably was no
reason to replace your NorthBridge heatsink. A good check to now:
Overclock the FSB to 400 MHz, but set the multiplier for the CPU to lower
than normal; this should give a differential diagnosis as to whether the
problem is with the new NorthBridge heatsink (or its installation.) Or
you could reinstall the original NorthBridge heatsink. It could be that
the new heatsink is not particularly flat and smooth. Or you could have
incorrectly applied thermal paste, or incorrectly installed the heatsink.

You haven't posted your motherboard either, a very important bit of
information for diagnosis. Nor your ambient room temperature, also
important. Nor your highest successful overclock speed test in your new
configuration. Nor what core voltages you have tried. Nor, in fact, why
you violated the first rule; if it ain't broken, don't fix it. So, among
other things, you might just try restoring everything to your original
successful overclocking setup.

Phil Weldon

Peanut butter (creamy) also works quite well but get a bit "crusty" after
a few weeks....:-) Like Phil said, your temps are very good for that
vcore. 3.4+ should be possible using that high a vcore, depending on other
factors like your NB chipset ect. With an Asus Striker Extreme (NV 680i
chipset) I get 3.2Ghz with only 1.2vcore and no temperature issues at all
with liquid cooling. Properly installing your thermal paste is very
important. Over at the Arctic Silver website they give a good example on
applying it correctly for the C2 Quads. I don't think their compound is
much, if any better than plain Radio Shack compound, but the instructions
are good and do work.
I am getting ready to start a project a bit early because of a SATA
controller issue on this system in which I have lost 2 channels out of 6
and one took out a HDD with it. This has been an ongoing issue with the
Asus 680i boards for me. I just ordered an i7 920 processor and an Asus
P6T Deluxe MB and will see how far the Nehalem will go with just air
cooling. I am going bottom of the line with the processor for right now at
under $300 at Newegg. I am seeking overall performance on this build and
ordered 6 gigs of Corsair XMS3 DDR3 PC1600 for 3 channel operation and
will use Vista Ultimate 64bit as my OS. I will recycle my two 8800 GTX
cards in SLI and will probably move a single 8800 GTS I still have over to
this system when I put in a spare P5N32-SLI MB if it shows signs of any
more issues. I am shooting for mid-3Ghz for the 920 if possible. I guess
someone has to be the "goat" with these processors....:-). I would have
rather waited a few months, but with the issues I am having, I don't trust
this system much longer. It has been an excellent performer, but has been
disappointing as far as my standards go with reliability.

Ed Medlin

I fixed it today, I got my new grease so I applyed it to the northbridge and
CPU. I noticed the northbridge heatsink was running hotter. There must not
have been enough contact or the grease wasn't doing a good job. Tested it
in Prime 95 and everything worked!

So overclocked it to 3.2 using 400fsb and 8 multiplier and that worked too
at 1.3500v yet processor 2 failed quite early, the other processors carried
on and the machine didn't lock up. Tried 1.3750v and it ran for 25mins
before processor 2 failed so tried 1.4000v and it ran for about 90mins
without any problems before I decided to end the test. CPU-Z reports the
Core Voltage as 1.312v

Good luck with the build!