2010 Macbook Pro 17" i7 heat question

Hi, I'm new to Macs and this forum but I wanted to post that last night I ran Handbrake to encode a video. According to smcFanControl my MBP was running at 200F, fans at 5000 rpm. Activity monitor showed Handbrake using 348% of the CPU.
 
Cinebench test on an i7 15"

I just ran Cinebench on my two week old i7 MacBook Pro 15" with fairly similar results to the PCAuthority test. On the OpenGL test, the CPU hit 80 degrees C with fans running at minimum (~2000rpm). The CPU test had the CPU temp peaking at 100 degrees C before the fans kicked in to ~4000rpm and brought the temp down to about 95 degrees C. Based off sound alone, I don't think that 4000rpm is the max for the fans, so maybe the fan speed is based upon the heatsink temperature which never got above 60?

Attached are screenshots from my second run. The first is right after the fans start to ramp up with CPU temperature peaking at 100. The second is from right before the test ended. The fan speed got up to ~5000 rpm and brought the temp down to 90 (not the hypothetical endpoint, but the lowest right before the test ended). I'm betting with a longer test, the temp would drop even lower.

Conclusion: On the PCAuthority test, the benchmark had just started and the fans had not gotten up to speed. This is supported by the relatively low temps of all the other components from their screencap. Are they trying to slander Apple?
 

Attachments

  • Screen shot 2010-04-26 at 3.35.15 PM.png
    Screen shot 2010-04-26 at 3.35.15 PM.png
    18.2 KB · Views: 2,049
  • Screen shot 2010-04-26 at 3.35.40 PM.png
    Screen shot 2010-04-26 at 3.35.40 PM.png
    17.7 KB · Views: 2,041
joelcsf said:
After 10 days of 10+ hours use per day, on battery and plugged in, I have not had this issue.

I did replace my HDD with an SSD - don't know if that might sway things, but I can't imagine the HDD would generate that much more heat?
Click to expand...


oh yea , it does make a difference , i noticed that on my desktop my HDD gets really warm but my SSD doesn't.
 
MBHockey said:
What do you mean not good? It's fine if it gets up to that momentarily, and then the fans come on. That's how computers work!
Click to expand...
It's not fine if the temperature gets so high that the CPU's internal thermal protection might get triggered before the fans kick in (which seems to be what happened in PCA's test).

IMHO Apple needs to lower the target temperture at which the fans start to kick in.
The original C2D MBP had similar heat issues and Apple solved it with a SMC update. On my "old" MBP the fans start to kick in at ~80°C.
 
Comparison ?

How hot will this thing run compared to my tibook ??
I cant find a temperature monitor which works, and the fan in this thing is either on or off, with the bottom casing getting burning hot :0 before the lonely 2cm fan trys to make it a little cooler :D
Also, does anyone know the age these mbps vaguely get to without something breaking. Only thing that has ever been replaced through an actual fault and not cos i wanted it on the tibook was the battery *:)

Thanks in advance

Btw, ive just ordered a standard 15' i7 for £1250 w/printer applecare and remote :D
 
sth said:
It's not fine if the temperature gets so high that the CPU's internal thermal protection might get triggered before the fans kick in (which seems to be what happened in PCA's test).

IMHO Apple needs to lower the target temperture at which the fans start to kick in.
The original C2D MBP had similar heat issues and Apple solved it with a SMC update. On my "old" MBP the fans start to kick in at ~80°C.
Click to expand...

It all depends on the response time though. It has to first get to the threshold temperature. Even if that's your proposed 80 C if the CPU temp shoots past that quickly there's still an opportunity to take a screen shot during those few seconds. If they posted a video where the temp remained at 100 C for a minute I would believe them, but they didn't. All of the other components are cool. Too cool for the CPU to have been at 100 C for any length of time.

Read the whole thread, specifically post 19. Their test is misleading at best.
 
MBHockey said:
Read the whole thread, specifically post 19. Their test is misleading at best.
Click to expand...
Read the whole thread. In post 20 I even quoted post 19... :rolleyes:

I'm not trying to defend PCA (why should I?) with their strange screenshots, taken at the worst possible moment and showing the wrong CPUs.
I'm just hinting at potential problems with the MBPs thermal management since post 19 confirmed possible temperature peaks. Even if everything stays just within specs (105°C according to Intel) it's still not good in terms of longevity of the components.


[edit]: Thread numbers don't match anymore because of thread merging
 
Follow up to my previous post

I just ran 10 tests on Cinebench back to back to try and get a steady reading. Starting from an icy computer with all the components around 20C, the CPU temp held nearly steady at 97C, with the heatsink holding at 59C and the fans never reaching 4000rpm. I'm guessing the fans are tied to the heatsink rather than the CPU, which seems weird. According to Intel, the max recommended temp is 105C with shutoff at 130C, but I'd prefer my processor to be a bit cooler. The good news is that the enclosure is always MUCH cooler than my previous PowerBook G5, even in the middle of these tests. The other good news is that the benchmarks actually improved with each CPU test, starting at 2.30 points and peaking at 2.50 pts on the final run.
 
This is exactly the same behaviour as my Core2Duo 13" MBP and a couple of others I have seen. So I don't think your i7s are faulty. Nice to see someone doing real objective tests.

The Intel developer documentation says the CPU should not exceed 105 C, and recommends that it its operating temperature should be 95 C. Apple seems to take use these numbers verbatim and designed a fan control algorithm which lets the CPU reach, but not exceed, 105 C for a brief while, then pulls the CPU back down to around 95 C.

I guess Apple's thinking is that most high CPU loads are transient, so there's no point in winding the fans up for a few seconds of high temperatures. If the high load is not transient, the fans wind up and pull the CPU down to the Intel recommended temperature.

I firmly believe that the Mac uses the CPU temperature (not the heatsink) as its setpoint. if I wind my CPU up and watch the temperatures and fan speeds I see the fan speeds hunt to keep the CPU at 95 C. The fan speed changes from about 2500 to 3500 to achieve this. Also the heatsink temperature can change depending on eg GPU load, but the CPU sits at the same temp, about 95 +/- 2 C.

Whether it's good for the CPU or not I wouldn't like to say. But rest assured your i7 MBPs are not faulty - it looks like their thermal management works exactly the same as the millions of C2D uMBPs out there.

I noted a rumor recently that Google was negotiating with Intel to have its server CPUs re qualified to a higher temperature (presumably 100 C?) so that Google could reduce its server farm cooling requirements. It seems Intel & customers take these temperature recommendations deadly seriously and Google and Apple at least want to design right up to them. I guess therefore Intel knows exactly what it's doing and has qual'ed its CPUs for thousands of hours at 95 C. But that's a guess.

So I think the original 100 C article was designed to mislead to maximise exposure. They could have done the same test with a C2D, an i5 or an i7 any time in the last couple of years.
 
MBHockey said:
What do you mean not good? It's fine if it gets up to that momentarily, and then the fans come on. That's how computers work!
Click to expand...

My old C2D never gets above 70 degrees when I do several days of constant simulations stressing the cpu 100%. ( not a macbook though )

I'm a bit worried for these temperatures to be honest if they get to 100 degrees, even if it's just for a moment :eek:

Anyone knows how high the temperatures of the 13 inch MBP can get?
 
@wirelessmacuser, StupaTroopa

Could you try the same test on your i7? I'm interested in more data on this.
I use the tried and true
-open terminal
- yes > /dev/null
- one for each core. I guess you'd need to run 4x instances of it, one for each virtual core.
but I guess Cinebench would be just as good. You'll need to daisy chain lots of instances together to get the required 15-20 min runtime.

If you watch the temps in iStat Pro you'll see the magic >100 C spike after a couple of minutes. Then it settles after 10-15 mins.

In particular I'm interested in the delta-T between CPU and heatsink. On a couple of C2Ds I've seen, delta-T is about 20 C (eg CPU at 95 C, heatsink at about 75 C). If you do the thermal calculations for the bulk silicon, the thermal paste and the copper spreader than seems to be roughly the right answer for a 35 W TDP. I've seen one case where changing the paste to well-applied AS5 made little difference to this delta-T. but StupaTroopa on here said his delta-T was more like 35 C which seems quite high.

I don't even want to mention bad thermal paste because IMHO it's mostly voodoo. But I'd be intrigued by your results if you have 20 mins to spare.

cheers guys
 
I did heat test on my new 15 inch i7 using Handbreak. After few minutes of using fans are at 2000 rpm and temperature spikes up to 100C, and then both fans kick in and they spin around 4600 rpm. Temperature goes down to 81-85C and stays there for the rest of the process. So fans start spinning faster at around 100C CPU temperature. Just a thought if you doing something CPU intensive maybe its good idea to use programs like SMCFancontrol so CPU temperature does that high up.
 
Ghost11 said:
I did heat test on my new 15 inch i7 using Handbreak. After few minutes of using fans are at 2000 rpm and temperature spikes up to 100C, and then both fans kick in and they spin around 4600 rpm. Temperature goes down to 81-85C and stays there for the rest of the process. So fans start spinning faster at around 100C CPU temperature. Just a thought if you doing something CPU intensive maybe its good idea to use programs like SMCFancontrol so CPU temperature does that high up.
Click to expand...

I tried the same thing on my C2D and got very very similar results. An observation & question....

Handbrake probably spends a lot of its time waiting on the DVD drive. So the CPU load surges up and down. This will be even truer on your faster i7 than my C2D. So I expect we are not taxing the CPU to its thermal limit. Hence the 85 C compared to the 95 C I can get if I run yes > /dev/null once for each virtual core. Certainly Handbrake makes my fan surge up and down quite a lot compared with yes > /dev/null, and the fan revs are lower (about 3700) compared with yes > /dev/null which is currently pushing 4900 rpm. Also I can see in Activity Monitor that the CPU is not pegged to its limit at all times with Handbrake.

What was your heatsink temperature while this was going on? More interestingly, what was the difference between the CPU and heatsink temperature? On mine it's rock solid at 20-21 C.

Could you try the yes > /dev/null test. I think you will need 4 terminal windows each running one instance of it. ctrl-t opens a new terminal tab.
 
More info: I think the fan control algorithm looks at other temperatures as well as the CPU. I was running Handbrake and charging all at the same time, my HDD temp went up to about 37 C as did my Enclosure 4 sensor. I think this sensor is where the Magsafe plugs in, which got warmish.

Now I run yes > /dev/null the CPU won't go above 90 C and the fan is working like crazy - 5000 rpm. Not seen it do this before. I suspect it is trying to get those two temps below some critical value, I suspect 35 C. Apple doesn't want users burning themselves, after all. So if you do want to run a redline CPU temp test, make sure the rest of the computer is cool!
 
Pax said:
I tried the same thing on my C2D and got very very similar results. An observation & question....

Handbrake probably spends a lot of its time waiting on the DVD drive. So the CPU load surges up and down. This will be even truer on your faster i7 than my C2D. So I expect we are not taxing the CPU to its thermal limit. Hence the 85 C compared to the 95 C I can get if I run yes > /dev/null once for each virtual core. Certainly Handbrake makes my fan surge up and down quite a lot compared with yes > /dev/null, and the fan revs are lower (about 3700) compared with yes > /dev/null which is currently pushing 4900 rpm. Also I can see in Activity Monitor that the CPU is not pegged to its limit at all times with Handbrake.

What was your heatsink temperature while this was going on? More interestingly, what was the difference between the CPU and heatsink temperature? On mine it's rock solid at 20-21 C.

Could you try the yes > /dev/null test. I think you will need 4 terminal windows each running one instance of it. ctrl-t opens a new terminal tab.
Click to expand...


I tried handbreak again, and CPU reached 100C and CPU Heatsink 56C. Then fans kicked in with around 4300rmp, CPU temperature dropped to 83C and CPU Heatsink to 53C. I would prefer fans to kick in before it reaches 100C, But for now i think i will control fan rpm myself when i run CPU Intensive applications. When running non CPU intensive application it runs pretty cool around 40C.

Next I tried yes > /dev/null, CPU temperature reached 97C and CPU Heatsink 56C. Then fans started spinning at around 5100rpm and CPU temperature dropped to 88C and CPU Heatsink to 52C. I dont think there is anything to worry about.

And one more thing that i forgot to mention, when i was running both tests I was charging battery in the same time
 
I have a question. I have the 17" core i5 macboook pro with 4gig ram and 500 gig 5400 rpm drive. I downloaded the istat menu app and my cpu pretty much sits at 100 degrees farenheit all the time with the fans never going over 2000 rpm. I can be doing absolutely nothing on my macbook pro and it will be 95-100 degrees farenheit all the time. Do i have a defective macbook pro? I have like 2 days left before my 14 day period is up......The temp goes as high as 130+ when doing anything other than web surfing.
 
Next Tuesday said:
I have a question. I have the 17" core i5 macboook pro with 4gig ram and 500 gig 5400 rpm drive. I downloaded the istat menu app and my cpu pretty much sits at 100 degrees farenheit all the time with the fans never going over 2000 rpm. I can be doing absolutely nothing on my macbook pro and it will be 95-100 degrees farenheit all the time. Do i have a defective macbook pro? I have like 2 days left before my 14 day period is up......The temp goes as high as 130+ when doing anything other than web surfing.
Click to expand...

100 Fahrenheit = 37.7777778 degrees Celsius which is normal. According to intel CPU should handle up to 105 Celsius =221 Fahrenheit.
 
Ghost11 said:
100 Fahrenheit = 37.7777778 degrees Celsius which is normal. According to intel CPU should handle up to 105 Celsius =221 Fahrenheit.
Click to expand...

well right now its at 151 farenheit with just mail, safari photobooth and iphoto open. fans still at 1999 -2000 rpm. thats normal? I see all these posts and youtube video of people saying there macbook pros run at 50-75 farenheit mostly all the time.
 
Next Tuesday said:
well right now its at 151 farenheit with just mail, safari photobooth and iphoto open. fans still at 1999 -2000 rpm. thats normal? I see all these posts and youtube video of people saying there macbook pros run at 50-75 farenheit mostly all the time.
Click to expand...

50-75 Celsius.

The amount of people that confuse these two units on this board is really amazing. :)
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Back
Top