Q6600 B3 stock @ 71C prime95, too hot?

[neosapien]

I don't want a 9700. I only grudgingly bought the 9500, and would have preferred to stay with stock cooling. I'm not overclocking. I'll just hope that the 9500, combined with undervolting (see my post above asking about the latest undervolting methods) will be enough.

 

[VirtualLarry]

71C is within the range of acceptable 24/7 temps to me. It's on the upper end of that range to be sure, but nothing to be overly concerned about. My E2140 @ 3.28Ghz 1.425v (BIOS) hits 71C core temp on occasion running Prime95 Small FFTs or SoB client.

 

[neosapien]

Holy jumping jesusfrogs.

After disabling Speedstep and C1E, I've been undervolting my chip (see my Notes) and it is stable for 5 minutes doing small FFTs in prime95, at 1.09375V in Gigabyte EasyTune, or 1.026V in CPU-Z / Speedfan (load temps dropped from 71C to 55C). I haven't done a several-hour-prime95 test because all this crazy amount of undervolting is getting me worried.

Is it really safe to be undervolting a q6600 B3 with a VID of 1.325V, down to 1.09375V? If I continue on the course of trying to find the voltage at which the PC crashes so I can bump it up another notch and everything is stable again, will my CPU die or something? I've never heard of a q6600 running at stock speeds at 1.026V.

 

[serialnumber2012]

my OEM system includes a q6600 b3 1.32v. Bleh. it also includes a Vigor Gaming Monsoon II TEC cooler --as u can see here:

http://www.anandtech.com/casec...howdoc.aspx?i=2943&p=5

...it slightly outperforms the TR ultra 120 extreme -- and yet with nothing altered and everything at stock settings, i get about 69c @ load on ORTHOS.

for zhits and giggles i attempted a 1333 fsb for a 3ghz OC. everything booted up stable into vista. running orthos went as follows:

Started small fft's
coretemp readings during the first 10 seconds:

45 - 55 - 59 - 64 - 66 - 73 - 75 - 77 - 77 - 79

followed by: "OH @#$%" [CLICK CLICK CLICK CLICK STOP STOP STOP STOP]

back to stock settings in bios..... i want my OEM to replace my cpu....for free......

 

[F1shF4t]

Originally posted by: serialnumber2012
my OEM system includes a q6600 b3 1.32v. Bleh. it also includes a Vigor Gaming Monsoon II TEC cooler --as u can see here:

http://www.anandtech.com/casec...howdoc.aspx?i=2943&p=5

...it slightly outperforms the TR ultra 120 extreme -- and yet with nothing altered and everything at stock settings, i get about 69c @ load on ORTHOS.

for zhits and giggles i attempted a 1333 fsb for a 3ghz OC. everything booted up stable into vista. running orthos went as follows:

Started small fft's
coretemp readings during the first 10 seconds:

45 - 55 - 59 - 64 - 66 - 73 - 75 - 77 - 77 - 79

followed by: "OH @#$%" [CLICK CLICK CLICK CLICK STOP STOP STOP STOP]

back to stock settings in bios..... i want my OEM to replace my cpu....for free......

My B3 with the TR120UE gets to 80C on 4 instances of P95 (thats at 3.15ghz). No big deal the cpu has a Tjunction of 100 and throttles at 95C. If it was a G0 stepping then I would worry. At stock it gets to about 71C. This is Australian summer so room temp ranges from like 26 - 30C when I have my aircon off.

 

[BonzaiDuck]

As much as I've been kicking myself for moving too fast last July and then finding myself with a B3 stepping, I feel better now.

But I never even used the processor with the stock cooler. For that, I could tell you that I also believe that 71C (load) is "in the ball-park" and not too high.

But with my motherboard ducting and a U-120-Extreme -- the VCORE set to 1.4183V and room ambients as high as 79F -- the upper range of the four core temperatures never exceeds 65C at either 2xORTHOS or PRIME95 25.4 loads. I can run it at 3.0 GHz at 1.32V, and of course, the temperatures are significantly lower. I don't really expect to break 62C at 79F ambient under load.

If Everest is reporting the VID under "OverClock->CPU-Properties->Core-Voltage," it is 1.288V. I guess that's pretty good.

There was a time when AigoMorla and I were fencing over whether you needed water-cooling to OC a B3-Q6600. Now I think we agree that you can go a lot farther with water, but either way, you venture into a range of voltage that might not be so healthy.

Even so, I think it was an AnandTech article that suggested that a "safe" voltage for the B3 was 1.46+V.

I . . . . . dunno . . . . . though . . . . .

 

[PCTC2]

Originally posted by: BonzaiDuck
As much as I've been kicking myself for moving too fast last July and then finding myself with a B3 stepping, I feel better now.

But I never even used the processor with the stock cooler. For that, I could tell you that I also believe that 71C (load) is "in the ball-park" and not too high.

But with my motherboard ducting and a U-120-Extreme -- the VCORE set to 1.4183V and room ambients as high as 79F -- the upper range of the four core temperatures never exceeds 65C at either 2xORTHOS or PRIME95 25.4 loads. I can run it at 3.0 GHz at 1.32V, and of course, the temperatures are significantly lower. I don't really expect to break 62C at 79F ambient under load.

If Everest is reporting the VID under "OverClock->CPU-Properties->Core-Voltage," it is 1.288V. I guess that's pretty good.

There was a time when AigoMorla and I were fencing over whether you needed water-cooling to OC a B3-Q6600. Now I think we agree that you can go a lot farther with water, but either way, you venture into a range of voltage that might not be so healthy.

Even so, I think it was an AnandTech article that suggested that a "safe" voltage for the B3 was 1.46+V.

I . . . . . dunno . . . . . though . . . . .

I think 71C is fine. I kick myself for falling for the Fry's July 4th Sale on Q6600 B3 steppings.... I should've just waited a month. But I've run my Q6600 B3 @ 75-77C @ 3.0GHz 1.38v SET 1.34 BIOS 1.30 Windows for Folding@Home for 3 months straight. I recently ran out of AS5, had to migrate my computer to a new case, could not replace the TIM, and now running my Q6600 @ 80C MAX and it's still chugging along...

But I'm replacing it soon anyways. I doubt I shortened its lifespan by more than a year or so.

 

[BonzaiDuck]

Anyway, SerialNumber_, you say it's an OEM system? . . . . So question here is this: what memory modules, what memory voltage, and what timings? And . . . . what motherboard? Unless the OEM vouched that you could over-clock the system, I don't see how or why they'd replace the CPU. But a 3 Ghz OC for a B3 -- especially with a Vigor Monsoon -- should be easy, unless the PSU is wanting, and especially if the RAM isn't up to it . . . . Or you haven't tweaked the RAM voltage or timings. OEMs often bundle with "just-adequate" power-supplies. IF it came with the MOnsoon, they would've accommodated it to PSU selection. But if you OC the system, well . . . .

Anyway, if it still works at stock settings, there shouldn't be anything wrong with it -- nothing damaged . . . . probably . . . .

 

[BonzaiDuck]

Originally posted by: PCTC2

I think 71C is fine. I kick myself for falling for the Fry's July 4th Sale on Q6600 B3 steppings.... I should've just waited a month. But I've run my Q6600 B3 @ 75-77C @ 3.0GHz 1.38v SET 1.34 BIOS 1.30 Windows for Folding@Home for 3 months straight. I recently ran out of AS5, had to migrate my computer to a new case, could not replace the TIM, and now running my Q6600 @ 80C MAX and it's still chugging along...

But I'm replacing it soon anyways. I doubt I shortened its lifespan by more than a year or so.

Those are the stock DDR2-800 settings that are SPD'd in the Crucials, aren't they? Did you ever try to tighten the latencies to 4,3,4,8?

You could also (probably) run the system at 3.2 GHz (multiplier 9) -- 1:1 -- 1,420 FSB and 710 DDR with timings as tight as 3,3,3,8 and even 1T command-rate. Then reduce tRC to 11.

Let me suggest the IC Diamond TIM. I'm curious as to whether you're using the stock cooler or an aftermarket cooler.

I'm wringing my hands over the gap between CPU's rated to run at 1,066 and memory rated at DDR2-800 or higher. If the processor were spec'd at 1,333 and had the same over-clocking potential, there wouldn't be much of a problem getting to 400 Mhz and 1:1 -- even with the 680i and the need for some tweaks to NB, VTT and VHT voltages.

And further, with this sort of setup, the easiest divider to use is 5:6 @ 3 Ghz, without pushing the RAM beyond 2 x 400 or 800 Mhz. But the most efficient divider barring 1:1 would probably be 4:5.

 

[ArchAngel777]

The bottom line is this: Logically, nearly everyone knows that 80c is fine for this chip. Seriously, if you just put some thought into it and analyze the situation, you will find that the chip can handle 80c. Besides that, the chip will never run that hot anyway in real world applications.

BUT, then comes the emotion side... The one that 'feels' that it is running to hot and makes you feel worried, which causes you to lower the voltage, the overclock and purchase a new heatsink for your CPU. Now you 'feel' better about your system. But you also spent more money, perhaps up to 50% the cost of the CPU.

It is a known fact that OEM Q6600 systems run plenty hot, hotter than 71 under load. The OEM's warranty these machines. They have more insider knowledge than any of us do as to what their chips are capable of... IF many OEM's are using Q6600's with poor thermals, I don't think you have a lick to worry about, logically... But if you think with your emotions, then you will likely OCD over your temps. Been, there done that.

My advice is to do what is 'reasonable' to lower temperatures. Once you have done that, clock accordingly and enjoy your purchase. If your chip dies, replace it, because by the time you would need to replace it, it will cost pennies on the dollar.

Edit **

For those wondering, I do put my money where my mouth is

1) When I first had my chip, I was using a Tuniq Tower 120 and my chip heated up to 85c for several hours... In fact, I know have well over 30 hours of testing at 85c. (P95 4 threads large FFT)

2) When I installed my SI128 my temps were in the low 80's, several more hours of testing at these temps. (P95 4 threads large FFT)

3) When I was changing items around, I forgot to plug in the fan on my SI128. I went into the BIOS to find the temperature hitting 99/100c (boiling!!!) it ran like this for a few minutes... Surprised it didn't shut down, but the heat sink burned my hand when I touched it.

So, I have a chip that has been through what many would consider 'hell'... Like, people will claim I lost 5 years on my chips life! Pfft, whatever... Don't get me wrong, I don't 'like' running those temps, but A) I don't run any programs that stresses my CPU like P95 does. B) The chip can handle 80c just fine. It does not throttle until 82c (B3).

So, I plan on keeping this rig for a few years... I'll let you all know how she handles 3.6Ghz @ 1.56v... If you read about my chip dieing, I'll be the first to post it... But, I don't think that is going to happen. BTW, I am on water now and temps hit around 58-60 under load now.

 

[aigomorla]

Originally posted by: wileyum
refuse delivery then! u need to get the 9700 with the bigger fan. its worth the extra money!

That wont even cut it.


How come no one believes me when i say these guys are furnaces? Ive been telling members for how many months now on how hot thees girls get when overclocked?

I dont think theres anyone on this forum who owns more quads then i do, yet people still dont believe me when i say these guys are monsters.

Listen people! expecially a B3 overclocked @ 1.4Vcore+ will generate ~200W+ of heat.

200W!!! this is NOT a small number. Its = to a small space heater. Lets see you try to cool a space heater with a small dinky copper sink


2 sinks FTW! also the bear minimum if you want to get decient overclocks.

Tuniq Tower
TR Ultra120 extreme or non extreme.


@archangel:

man you abused the crap out of that chip... i dont think any of my chips have ever seen middle to upper 60's even. Anyhow your making up for the abuse in the leg end. So you'll even out on life span as someone who was on air the entire life on moderate overclock.

Intel engineer once said, you reduce your temps by 10C you effectively double the life of your processor.

 

[ArchAngel777]

Originally posted by: aigomorla
@archangel:

man you abused the crap out of that chip... i dont think any of my chips have ever seen middle to upper 60's even. Anyhow your making up for the abuse in the leg end. So you'll even out on life span as someone who was on air the entire life on moderate overclock.

Intel engineer once said, you reduce your temps by 10C you effectively double the life of your processor.

I agree that you probably double the life by cutting the temperature by 10c. Keep in mind that is constant though.

If my machine runs 40c for 12 hours, 60c for 11 hours and 85c for 1 hour per day, it will not have nearly the impact you think comparing it to, say 38c for 12 hours, 55c for 11 hours and 75c for 1 hour. In other words, the 10c rule only applies if you can keep operation always 10c lower for the same ammount of time.

If I shut my system off when I am not using it (15 hours a day) and use it 9 hours a day, I still outlast someone who runs theirs 24 hours a day that runs 10c cooler.

I think what I like to get accross is people using their brains to make educated decisions. The variables we have are 1) Load versus Idle, 2) Time in each state, 3) Temperature, 4) Voltage and 5) Hardiness of the CPU. Those 5 variables will affect the equation of how long your CPU will last. Additionally, no one knows how long a CPU really lasts in the first place. Intel hasn't released a half-life publicly, that I know of...

But, I do have much experience behind systems running hot. The servers that I administer, some of them (Quad Xeon's) run 'extremely' hot... Yet, even after several years they are running strong with no signs of weakness. Then you have stories of people running their C2D for a year not knowing that the heatsink wasn't properly mounted and the temps were idle at 70c and loaded near shutdown point. Yet, they still work properly and are still running strong today. I ran my AMD 64, 3500+ so hot that the the Arctic Silver cemented the heatsink to the chip. I litterally ripped the chip right out of the socket removing the heatsink. This chip is STILL going strong after 4 years! These chips can take one hell of a beating... Besides that, even if the A64 3500+ failed today, I could replace it for 20-30 bucks... Not a big deal in my opinion.

I think what I am here to do is counter all the people who incite fear to their chips are going to die within months. I have seen many people post that on these forums and on others... So, this isn't directed at you Aigo. I see both sides of it... But these chips are sturdy

 

[aigomorla]

Have to agree with you however...

Today's die = smaller = weaker = cant take as much stress/heat as the older and larger dies. :X


Thats why i keep telling my parents, computers and electronics are not being build as reliable as they did back then with much higher stress points. So look at upgrading frequently.


And even now im hearing about people burning yorkies out while running 1.65Vcore on them. And some of them were even phased. :X

Yorkies = QX9650

 

[ArchAngel777]

Originally posted by: aigomorla
Have to agree with you however...

Today's die = smaller = weaker = cant take as much stress/heat as the older and larger dies. :X


Thats why i keep telling my parents, computers and electronics are not being build as reliable as they did back then with much higher stress points. So look at upgrading frequently.


And even now im hearing about people burning yorkies out while running 1.65Vcore on them. And some of them were even phased. :X

Yorkies = QX9650

It probably died to voltage rather than heat though. That is ~40% over the VID... That would be taking a Q6600 with a VID of 1.3 and running it at ~1.82v. I don't even know how you could cool that with phase... I suppose anything is possible... Has anyone ran 1.8v on a QX6850? In any case, it will be interesting to see the life of the 45nm overclocker... You are right though, that these chips are becoming more fragile, but I still think they are quite hardy and will be for some time more.

 

[dsg2003gt]

Im running my G0 Q6600 at 3.4ghtz and 1.3275v in the BIOS, my chip has a VID of 1.125. It is 12 hrs prime stable and Wprime stable. Im basically just browsing the net right now and my temps are 39-36-39-34. I am running the tuniq tower with the fan at about 1200rpm. I have 1 intake fan and 1 rear exhaust fan. I put a side exhaust fan under my graphics card(8800gt) and it helped out by 1-2*.

The Q6600 will reach temps about 59*C when running prime. The highest ive seen the temp under normal use is 55*.

 

[Idontcare]

Originally posted by: ArchAngel777

Originally posted by: aigomorla
Have to agree with you however...

Today's die = smaller = weaker = cant take as much stress/heat as the older and larger dies. :X


Thats why i keep telling my parents, computers and electronics are not being build as reliable as they did back then with much higher stress points. So look at upgrading frequently.


And even now im hearing about people burning yorkies out while running 1.65Vcore on them. And some of them were even phased. :X

Yorkies = QX9650

It probably died to voltage rather than heat though. That is ~40% over the VID... That would be taking a Q6600 with a VID of 1.3 and running it at ~1.82v. I don't even know how you could cool that with phase... I suppose anything is possible... Has anyone ran 1.8v on a QX6850? In any case, it will be interesting to see the life of the 45nm overclocker... You are right though, that these chips are becoming more fragile, but I still think they are quite hardy and will be for some time more.

Also folks are treating 45nm tech like it is merely an extension of the 90nm and 65nm tech.

The introduction of HK/MG xtors really changes a lot of the failure mechanisms and tendancies of these new chips.

An acceptable design margin for overvolting a 90nm or 65nm may not even exist in 45nm chips.

It is not like Intel has had this material for 30 years to know how to design for margin...I would not be surprised if 32nm xtors are a lot more robust than these initial production run 45nm ones just for the sake of the fact that Intel's engineers will have had a few more years experience with knowing how these things break down in the corner cases.

 

[Idontcare]

Originally posted by: dsg2003gt
Im running my G0 Q6600 at 3.4ghtz and 1.3275v in the BIOS, my chip has a VID of 1.125. It is 12 hrs prime stable and Wprime stable. Im basically just browsing the net right now and my temps are 39-36-39-34. I am running the tuniq tower with the fan at about 1200rpm. I have 1 intake fan and 1 rear exhaust fan. I put a side exhaust fan under my graphics card(8800gt) and it helped out by 1-2*.

The Q6600 will reach temps about 59*C when running prime. The highest ive seen the temp under normal use is 55*.

That is a ridiculously wonderful low VID. Congrats! My lowest VID G0 is 1.2625.

It does run incrementally cooler than my 1.275 and 1.2825 VID G0's at the same clocks (9x367) but surprisingly these G0's these days all seem to be pretty consistent about their OC capability versus voltage requirements and loaded temps.