*Official Retail G0 Q6600 Overclocking Thread*

[twilyth]

I should probably start a thread for this, but I'll float it here first. I had my first Q6600 running at 3.4ghz. Thought I had maxed out because core temps were reading about 60c degrees. So I left it that way for over a month. Then I decided I would see if I could squeeze a little more out of it. I had to increase voltage to around 1.375 or so to get it stable, but it went up to 3.6ghz with no other problems. However now the individual core temps are in the low 40's but temp2 in speedfan is now over 50c and goes over 60c if I run WCG and Prime95 at the same time.

Is it possible that the thermal paste has settled and become more efficient? Pretty sure I read about this, except I thought you needed multiple heat-up/cool-down cycles for that and the rig has been on continuously. Anyway, if that's what's going on, where is the hot temp2 reading coming from?

[chuckbam]

Thats because it is an overclocking thread not a performance thread. There are a lot of performance threads around the time a new cpu is released.

I don't see the difference between CPU OC and CPU performance information.
Just to clock a CPU w/o a measurement of benefit is incomplete. What if the CPU Clock increase decreases or has no performance gain. That information is not important for a CPU OC thread?

That is why benchmarks are always included in an OC article. To find a CPU OC that is justified, you need to look at all parameters. There is more to CPU overclocking than lapping.

[twilyth]

According to this post Temp2 is NB. So maybe I just need to plant a fan on the NB heatsink?

[sesdave]

I should probably start a thread for this, but I'll float it here first. I had my first Q6600 running at 3.4ghz. Thought I had maxed out because core temps were reading about 60c degrees. So I left it that way for over a month. Then I decided I would see if I could squeeze a little more out of it. I had to increase voltage to around 1.375 or so to get it stable, but it went up to 3.6ghz with no other problems. However now the individual core temps are in the low 40's but temp2 in speedfan is now over 50c and goes over 60c if I run WCG and Prime95 at the same time.

Is it possible that the thermal paste has settled and become more efficient? Pretty sure I read about this, except I thought you needed multiple heat-up/cool-down cycles for that and the rig has been on continuously. Anyway, if that's what's going on, where is the hot temp2 reading coming from?

If you can get 3.6 on 1.375V (actual?) you have a very nice CPU For a start use coretemp 95.4 beta or PC wizard 2008 to read your coretemps cos speedfan reads 15c under actual unless there is a new version which has been calibrated to use tjunction of 100c. Did you use prime95 v25.5 which runs a test on all 4 cores -what do you mean by temp2 - CPU or Core2 temp. Some pastes get better with curing like AS5 - I use MX-2 which doesnt need to cure and is as good as AS5.

[Devious]

I don't see the difference between CPU OC and CPU performance information.
Just to clock a CPU w/o a measurement of benefit is incomplete. What if the CPU Clock increase decreases or has no performance gain. That information is not important for a CPU OC thread?

That is why benchmarks are always included in an OC article. To find a CPU OC that is justified, you need to look at all parameters. There is more to CPU overclocking than lapping.

I agree, performance is important, no point overclocking past the point of diminishing returns. But in previous threads it has been shown that with performance has always increased with higher clocks (except gpu limited games) so people want to clock as high as they can as long as it is stable.

This thread gives info to people about what batches are clocking well and which are not, people obviously want to know this info or this thread wouldn't be 61 pages long.

[twilyth]

If you can get 3.6 on 1.375V (actual?) you have a very nice CPU For a start use coretemp 95.4 beta or PC wizard 2008 to read your coretemps cos speedfan reads 15c under actual unless there is a new version which has been calibrated to o use tjenction of 100c. , Did you use prime95 v25.5 which runs a test on all 4 cores -what do you mean by temp2 - CPU or Core2 temp. Some pastes get better with curing like AS5 - I use MX-2 which doesnt need to cure and is as good as AS5.

It's at least at 1.375 but might be one or 2 ticks higher - definitely under 1.4. Speed fan seems to be accurate since i have the temp alarm set to 60c and it goes off whenever speedfan indicates temp2 is above that

Temp2 is speedfans designation for readings from a particular sensor. It seems that it's using the sensor for the north bridge. I'm running the memory at a little under 1066 but I guess the increase in voltage is what's causing the problem - although I'm still a little unclear on the relationship between vcore and the NB. I've read some stuff about NB but it's still pretty fuzzy.

btw - using AS5 for the paste.

[sesdave]

It's at least at 1.375 but might be one or 2 ticks higher - definitely under 1.4. Speed fan seems to be accurate since i have the temp alarm set to 60c and it goes off whenever speedfan indicates temp2 is above that

Temp2 is speedfans designation for readings from a particular sensor. It seems that it's using the sensor for the north bridge. I'm running the memory at a little under 1066 but I guess the increase in voltage is what's causing the problem - although I'm still a little unclear on the relationship between vcore and the NB. I've read some stuff about NB but it's still pretty fuzzy.

btw - using AS5 for the paste.

speedfan is not accurate for coretemp on a quad as far as I know unless the latest beta has resolved it. Up to now it works on a tjunction of 85c unless you calibrate it so it will be -15c out. Try this and post what the temps are http://www.cpuid.com/pcwizard.php - install,run then minimise - it will post temps in top of screen. I run 3.6 with 1.4v real - idle 35c - load 62c using air cooling. I also stick a little 40mm fan on the NB from habit on the P5W - probably doesnt need it but every little helps-lol!

[Hans Gruber]

My point of the Tomshardware.com review was to show the performance of the MSI M35 NEO-2 FR/FIR compared to the Gigagbyte board. I also wanted to point out the limits of their testing as well.

[sesdave]

My point of the Tomshardware.com review was to show the performance of the MSI M35 NEO-2 FR/FIR compared to the Gigagbyte board. I also wanted to point out the limits of their testing as well.

No worries Hans- it was taken that way. I was just poking a bit of fun at the apparent limitations of the test, Very sensible but not really going for it like we would like to see -lol!. You cant really see what a MB or the CPU will do unless they push it to the limit and identify weaknesses. e.g check this thread out as an example - its a living review of the Abit x38 QUADGT - its just started but day by day it will let you really know what the MB is capable of , how to configure it and its quirks . http://www.clunk.org.uk/forums/revie...ng-review.html Clumk also did one on the Asus Maximus . these are worth reading if you are in the market for an x38 MB.

[~aoe~]

speedfan is not accurate for coretemp on a quad as far as I know unless the latest beta has resolved it.

The latest beta resolves this issue. Running 4.34 beta 34 and temps are reported correctly without the need to apply offset (in advanced options).

[sesdave]

The latest beta resolves this issue. Running 4.34 beta 34 and temps are reported correctly without the need to apply offset (in advanced options).

about time - lol! I have got to like PC wizard so I will stick with that.

[Hans Gruber]

Good to hear the the newest Speedfan works correctly now. I can make and fix any hardware issue. Software issues are my weakness. I couldn't get Core Temp to work on my Vista 64 bit, but Speedfan works great.

I'm not that bad with getting software to work. I got riva tuner to work with Vista 64.

The MSI P35 NEO-2 FR has received rave reviews from many different sites before Toms's Hardware review. The FIR version has firewire and a few more features but it's essentially the MSI P35 platinum minus the fancy heatpipes.

I also agree Tomshardware doesn't push processors enough, that's why I'm here. You guys push stuff to the extreme so I don't have to test as much. I can just copy what the experts are already doing.

[Devious]

[Family man 72]

Impressive clocks there Devious and really nice temps Is cpu lapped and what kind of wc kit you have?

[Devious]

Impressive clocks there Devious and really nice temps Is cpu lapped and what kind of wc kit you have?

Hi, CPU isn't lapped, WC is PA120.3 with D-Tek fuzion w\nozzle kit.

[lockeh]

anyone heard of this stepping Q6600

L725R190 SLACR

[Eldonko]

Hi, CPU isn't lapped, WC is PA120.3 with D-Tek fuzion w\nozzle kit.

Those temps are killer, are you benching in cool air? Anything other than CPU in loop? I have apogee and have 10C more than that at 1.36v and 3.76G

[Devious]

Those temps are killer, are you benching in cool air? Anything other than CPU in loop? I have apogee and have 10C more than that at 1.36v and 3.76G

Welll, its a cool l day today, there is definately nothing else but my ambient and my PA 120.3 rad w/fans

[Obijuan83]

Q6600 @ 4.0GHz 1.40volt
Transcend AxeRam 800 @ 1112 5-5-5-5-25 2.12volt
Maximus Formula (NB without water)
Single Stage By Unseen

[andressergio]

My God what batch you have ObiJuan ?

Regards
Sergio

[LuCH]

Hi everybody, just signed in, and I'd like to tell you how it went with my q6600 g0.
Well, I managed to bring it to 3,6GHz, 400x9 at 1,4V in BIOS (1,35V in CPU-Z)... I haven't tested for stability, tho, since it's just too damn hot for me. Core temp would report some 70°-75° for the hottest core during Prime small-FFT-torture. I guess my HS is not properly seated... AS5 doesn't seem to be curing anything.

This chip is a good one, its VID is just 1,125V and for 24/7 peer2peering I underclock/undervolt it to 233FSB/0.95V. Temps then are in the low 30s.

[jowi69]

Q6600@4366MHz under SS

[rbsmontemor]

New CPU: (New BD entry plz :P)

Q6600 G0
L726A822
Vcore 1.58v
D-tek Fuzion - MCP655 - MCR320
P5K Deluxe



Rbs.

[Chosen.]

737B, 3609MHz @1.328v under 247 100% cpu load (boinc)

Inside a mATX Thermaltake Lanbox Lite

[eva2000]

Wow Chosen very cool cpu.. what's room and case ambient temps ???

Trying my Q6600 L725A903 G0 cpu for the first time on Asus Blitz Formula. It's also first time I've tried air cooling on this Q6600 L725A903.

Full batch details:

Q6600 SLACR
FPO/Batch: L725A903
Version: E10342-001
Pack Date: 07/31/07

To cool Q6600 L725A903 G0 cpu, I decided on using my Thermalright Ultra 120 Extreme with 120x38mm Panaflo 103cfm fan and Artic Cooling's MX-2 thermal paste.

Surprisingly, with Thermalright Ultra 120 Extreme air cooling, I'm getting similar cpu load delta temps (coretemp) as with my Swiftech G4 Storm water cooling setup - air cooling approximately 1.4C delta cpu load temps behind and 0.4C warmer in terms of cpu idle temps above room ambient. But air cooled setup is running Prime95 torture test at less vcore even when room ambients on air are warmer!

@3600Mhz at 1.432v load on Asus Blitz Extreme - Swiftech G4 Storm H20
CPU load to room delta = 57C - 24.5C = 32.5C Delta

vs

@3600Mhz at 1.408v load on Asus Blitz Formula - Thermalright Ultra 120 Extreme Air
CPU load to room delta = 62C - 28.1C = 33.9C Delta

System

• Intel Core 2 Quad Q6600 L725A903 7/31/07 pack date G0
• Thermalright Ultra 120 Extreme + 120x38mm Panaflo 103 cfm fan + MX-2 TIM
• Asus Blitz Formula 1101 bios
• 128MB Gainward FX5200 PCI
• 2x1GB Crucial Ballistix Tracer PC2-8500
• Memory cooling = 120x25mm Spire 96cfm fan on rheobus @100% fan speed.
• 80GB Hitachi 7K80 SATAII
• Pioneer DVR-08 Burner
• Sony FDD
• 620W Corsair HX620 PSU
• WinXP Pro SP2

room temp = ~28.1C

Prime95 Small FFTs Idle


Prime95 Small FFTs Load