3570K and 75C gaming temps

[tigersty1e]

I'm fine with taking my cpu up to 90C with prime stress testing.

But what about 24/7 gaming load temps into the 75C region? I'm at about 65C with 4.4 and just looking at how much overhead I have.

 

[Don Karnage]

What sort of cooling do you have?

 

[Clinkster]

What are you asking exactly? If 75C is acceptable for 24/7 gaming (whatever that means)?

 

[tigersty1e]

What sort of cooling do you have?

antec 620. i'm not sure how my answer would change your response to my first question though.

is 75c fine for gaming temps?

 

[Don Karnage]

antec 620. i'm not sure how my answer would change your response to my first question though.

is 75c fine for gaming temps?

Ofcourse it is and i misunderstood the question.

 

[ShintaiDK]

Tjmax is 105C. So essentially everythign below is fine. At the 95-100C area I would look on how to improve cooling tho, for a hot summer day.

My 3570K hits 80C with a 100% linpack load on the stock cooler.

 

[Idontcare]

I'm fine with taking my cpu up to 90C with prime stress testing.

Intel is fine with you taking your CPU up to 105C.

Why would you arbitrarily decide the engineers at Intel don't know what they are doing and instead institute your own thermal limit that is even lower and more conservative

105C is the limit not because things go downhill at 106C, the 105C number is there because it is already a very conservative number placed well below the actual temperature value at which the rate of degradation becomes problematic on the timescale of the CPU's warranty and Intel's own expectations of in-field fails.

You can run your CPU at TJmax (105C in your case) all day, every day, for years and years. It is 105C, and not 90C, for a reason.

 

[ShintaiDK]

Talking about temps. I think I remember my old TNT2 to hit 130-135C. And it kept working until it got too outdated. Not because it stopped working.

The problem is, as I see it. That people tend to put human scale into temperature for silicon. Chipsets easily hit 100C for example. You just dont know it because no sensor tells you it. But if you got a sensor. You might be omg!!! Water boiling temp!

 

[Idontcare]

Yeah, nowadays the upper temp spec is more likely to be instituted not because of concern for the silicon itself but because of the fact that the product is also spec'ed for a maximum TDP the wattage rating.

Since temperature plays a huge role in determining the power consumption that is due to leakage, if they want to clock the chip at 3.5GHz and have it still fit within the power spec then they must define that upper temperature limit so that the power spec if not violated.

This is why AMD chips have a lower maximum temperature, if they spec'ed to allow a higher max temperature then they'd have to lower clockspeed so that the chip did not violate its own power-rating if and when it reached those upper temps.

Combine that with the fact that as temperatures rise so too does the minimum required operating voltage for stability and you have a one-two knockout on power-consumption as temperatures rise.

Not an issue if you spec your chips to have the power-budget, but if you spec your chip as a 95W chip then it needs to have its maximum temperature spec capped accordingly so it doesn't violate that when operating at its rated clockspeed and corresponding Vcc.

(and there is the reliability/degradation aspect to elevated temps, but that is already captured on the spec definition side of the business...it is 105C and not 90C and not 120C for reasons relating to both power-consumption at 105C as well as degradation at 105C)

 

[bononos]

Intel is fine with you taking your CPU up to 105C.
.......
You can run your CPU at TJmax (105C in your case) all day, every day, for years and years. It is 105C, and not 90C, for a reason.

But won't 105C be harmful to the other components? Did the bumpgate problem have something to do with high temps? Would the pcb buckle up or capacitors pop open?

Talking about temps. I think I remember my old TNT2 to hit 130-135C. And it kept working until it got too outdated. Not because it stopped working.
.....

I remembered the tnt2 very well because Nvidia's reference card I think did not have a fan and it ran hot. I'm fairly certain my tnt2 died because of heat since it would work until it heated up. Overheating wasn't too uncommon with the tnt2.

 

[Ferzerp]

But won't 105C be harmful to the other components? Did the bumpgate problem have something to do with high temps? Would the pcb buckle up or capacitors pop open?

There are very exacting socket and surrounding component specs for a reason.

 

[Idontcare]

But won't 105C be harmful to the other components? Did the bumpgate problem have something to do with high temps? Would the pcb buckle up or capacitors pop open?

Bumpgate was not caused by the max operating temp, it was caused by the number of thermal cycles.

Very much the same concept of why incandescent light bulbs failed and why your parents would tell you to not play with the light switch.

Turn the light on (it gets hot, but steady-state) and leave it on and the light will last for decades.

Flip the switch on and off rapidly and after a while the filament fails under the stress cycling created by thermal expansion and contraction.

That was the problem with bumpgate, lowering the max upper temperature was merely a bandaid to an otherwise critically under-engineered problem.

If Intel has failed to properly characterize the capability of its products to endure thermal cycling then to be sure the upper temp spec will only exacerbate the problem, but lowering the upper temp spec will not negate such a problem either.

The mobo makers are required to design their products such that they can properly operate in whatever conditions the other components are operating.

If a mobo maker creates a mobo that supports IB but the mobo fails if the IB CPU itself gets to 105C then that is a problem created by, and owned by, the mobo maker.

Nothing prevents the mobo maker from making shoddy mobo's, the onus is on the consumer to avoid known shoddy mobo makers if you intend to push the components to the limit. And this is true of all electronics, right? We don't buy a 750W rated PSU for $15 and actually expect it to function if tasked with supporting a 700W load. We know we bought a shoddy PSU for $15 so we keep its load under 300W.

Same with your mobo, but don't say you are operating at 90C for the sake of the CPU, recognize you put your $300 CPU into a shoddy $75 mobo and the $75 mobo is what is limiting what you can reasonably do with your $300 CPU.

(just as you wouldn't put goodyear all-season radial tires on your lambo and then attempt to drive your lambo at 200mph......at least I don't do that with my lambo ...but if you did that doesn't mean the lambo has problems going 200mph, rather we would file that under "user error")

 

[Ferzerp]

(just as you wouldn't put goodyear all-season radial tires on your lambo and then attempt to drive your lambo at 200mph......at least I don't do that with my lambo ...but if you did that doesn't mean the lambo has problems going 200mph, rather we would file that under "user error")

I hate having to buy Y rated tires for my car, but yeah, not going to create a situation where I could let that happen.

 

[Termie]

Just an aside here - doesn't this graph suggest that power use scales linearly with temperature, and not with voltage, which is what I thought the common understanding was?

 

[tigersty1e]

Intel is fine with you taking your CPU up to 105C.

Why would you arbitrarily decide the engineers at Intel don't know what they are doing and instead institute your own thermal limit that is even lower and more conservative

105C is the limit not because things go downhill at 106C, the 105C number is there because it is already a very conservative number placed well below the actual temperature value at which the rate of degradation becomes problematic on the timescale of the CPU's warranty and Intel's own expectations of in-field fails.

You can run your CPU at TJmax (105C in your case) all day, every day, for years and years. It is 105C, and not 90C, for a reason.

exactly, the number was thrown out arbitrarily because 75c gaming temps correspond to 90c prime load temps.

my question was about 75c gaming temps.

as far as 104c being safe for "all day, every day"... is another story. it may be, but running an engine to redline all day, every day is within spec. i wouldn't want to do that on my new car.

 

[tweakboy]

It will work the chip will live and work but its lifespan will be shorter. or it can possibly get damaged or bad sector etc.

75c is a great temp for a HSF with that voltage and OC.

Your never going to full load your computer unless you render movies from vegas or premiere etc.

gl

 

[Idontcare]

Just an aside here - doesn't this graph suggest that power use scales linearly with temperature, and not with voltage, which is what I thought the common understanding was?

Power scales with both temp and volts.

The graph you were looking at above can be deceiving because the variation in voltage is very small (the range on the lefthand y-axis) compared to the variation in temperature (the range of the x-axis) and the variation in the power-usage (the range of the righthand y-axis).

 

[Termie]

Power scales with both temp and volts.

The graph you were looking at above can be deceiving because the variation in voltage is very small (the range on the lefthand y-axis) compared to the variation in temperature (the range of the x-axis) and the variation in the power-usage (the range of the righthand y-axis).

Wow. Not that I can entirely follow that, but I see that temperature comes into the equation in the "Poole-Frenkel effect". Wikipedia says it's "a means by which an electrical insulator can conduct electricity." Sounds like leakage to me.

Do you do this for a living or something?

 

[IntelUser2000]

Guys, while the Tjunction might be rated 105C, Intel recommends that the heat sink solution be designed so it stays well under that. In the datasheet they give an example of 10 degrees lower, which is 95C.

In previous Intel processors, throttling of the Base clock happened at 80-85C.

 

[ShintaiDK]

Guys, while the Tjunction might be rated 105C, Intel recommends that the heat sink solution be designed so it stays well under that. In the datasheet they give an example of 10 degrees lower, which is 95C.

In previous Intel processors, throttling of the Base clock happened at 80-85C.

Heatsink needs to be designed for different climates. Summer in Africa, India or winter in Iceland or Russia. Hence why the heatsink in regular temperatures needs a headroom. The specs also uses a certain ambient (Ta) to measure from. Or if someone adds 4 GFX cards in SLI/CF...

 

[IntelUser2000]

That makes sense, but there are reviews that show throttling at 90C. There's the "unsafe to operate" temperature, but also "performance limiting" temperature as well.

Pentium M with Tjunction temperature of 100C throttled at ~85C.

 

[ShintaiDK]

That makes sense, but there are reviews that show throttling at 90C. There's the "unsafe to operate" temperature, but also "performance limiting" temperature as well.

Pentium M with Tjunction temperature of 100C throttled at ~85C.

They do NOT throttle at 90C. Do you want a Linpack screenshot at 104C with +200Mhz turbo on all 4 cores?

 

[IntelUser2000]

They do NOT throttle at 90C. Do you want a Linpack screenshot at 104C with +200Mhz turbo on all 4 cores?

http://www.tomshardware.com/reviews/ivy-bridge-overclocking-core-i7-3770k,3198-2.html

Core Temp 1.0 RC3 reports that our Core i7-3770K reaches 90-100°C (194-212°F) internally when it's overclocked to 4.5 GHz. No wonder the chip's thermal monitor tripped, throttling the CPU. This phenomenon dropped the effective clock rate of our chip to approximately 3.5 GHz, corresponding to the CPU’s nominal frequency.

http://hothardware.com/Reviews/Intel-Core-i73770K-Ivy-Bridge-Processor-Review/?page=3

We were actually able to boot into Windows and run benchmarks at clocks above 4.8GHz, but our CPU was breaking the 90’C mark and throttling to keep temperatures in check.

 

[Magic Carpet]

Perhaps, the board itself triggered a fail-safe?

Had I engineered those... that's what I would have done. No need to burn my efforts in vain.

 

[ShintaiDK]

According to the screenshot they supply. Its running at 4.5Ghz at 101C. Meaning no throttle at 101C.

Plus I have tested with my own 3570K. So I know it doesnt throttle until 105C.