Poll: M295X GPU heat in Valley benchmark

[iBighouse]

I ran it at the default resolution first and it didn't get above 100 and voted in the poll based on that. I then noticed you asked for the higher res, so I re-ran the benchmark. It got to 101.

I have the 1TB SSD if it makes any difference and 32GB of RAM.

 

[vagos1103gr]

i let it run 20 min the max temp i got was 100 degrees. fps 17.9 score 745
i7 r9 m295x 32gb ram.

 

[odedia]

So far no one reached the 107 degrees Celcius mark that is considered the throttling part. I wonder if the benchmark is hitting the GPU high enough? I would imaging it is...

This is the time where I would really appreciate an Anandtech review.

 

[gymrat2k]

Consistently 105-106, occasionally 107 degrees celsius.
22 degrees ambient. 2303 RPM.
i7, M295X, 32 GB, 1 TB SSD.

Edit: Sorry, that was with 2560x1600 resolution. With 2560x1440 it is 104-105 degrees consistently.

 

[andy9l]

So far no one reached the 107 degrees Celcius mark that is considered the throttling part. I wonder if the benchmark is hitting the GPU high enough? I would imaging it is...



This is the time where I would really appreciate an Anandtech review.

I got a video of some form of throttling in Windows. Only took 90 seconds of a mediocre game.

Regardless, all my research points towards the fact that absolutely no piece of hardware should consistently run at 100C+ for hours on end. Simple as that.

https://forums.macrumors.com/posts/20572294/

 

[odedia]

I got a video of some form of throttling in Windows. Only took 90 seconds of a mediocre game.

Regardless, all my research points towards the fact that absolutely no piece of hardware should consistently run at 100C+ for hours on end. Simple as that.

https://forums.macrumors.com/posts/20572294/

I think on the Windows side it's a driver issue at the moment. Makes no sense that it is relatively ok under OSX and so high on Windows.

I would probably stay away from the current version of those drivers.

Off topic: I want to start a separate thread with similar benchmark for the CPU, not the GPU. I've noticed my CPU is constantly around the 60-70 degrees Celsius and I wonder if it's only me. Can anyone recommend a good benchmarking app for the CPU that would allow us to check the HEAT of the CPU?

 

[andy9l]

I think on the Windows side it's a driver issue at the moment. Makes no sense that it is relatively ok under OSX and so high on Windows.

It's the same on Windows and Mac. 105C with some spikes above that temperature.

 

[Alesc]

Just out of curiosity I decided to run this benchmark on my 2013 iMac i5 3.4 nVidia 775M with all the same settings you guys suggested.

GPU temp according to the program hovered around 80c (82c max) so the fan never increased from 1200rpm.

That I don't find surprising. What I do find surprising is the results of FPS and score. Shouldn't you guys be in an entirely different ballpark? Did I test something incorrectly?

Image

Image

EDIT : I obviously didnt take part in the poll as I don't have a 5k iMac FYI. Also to point out I've had the GPU get much hotter with games (ESO 95c), so if you guys are seeing throttling with this benchmark I'd imagine you'd certainly see it real world.

I've ran the test on my Late 2013. 20°C ambient temp, 87°C in the benchmark (iStat displayed 74°C...) :

 

[OlexandrIgnatov]

Hello!

rImac, R295X 4GB, 32GB RAM, i7, 3TB Fusion Drive here.

The temperatures never exceeded 103C. Actually it was changing between 102-103 degrees most of the time.

I ran the test two times in a row for about 7 minutes.

Here's the screenshot of a second test.

 

[cynics]

I've ran the test on my Late 2013. 20°C ambient temp, 87°C in the benchmark (iStat displayed 74°C...) :

Image

To note I use TG Pro for temps and the 82c I saw in the benchmark program was also reflected in TG Pro at 82c.

Its curious iStat would display differently unless it using a proximity sensor or an average?

EDIT: Took some screen shots.

Waited for the benchmark program to get to 80c.

I couldn't get out of the benchmark program fast enough to get a live temp so I reset TG Pro maximum recorded temps before running the benchmark. So you can see it reports the max it got to was 80c.

And of course the fan still putting along at <1200rpm.

 

[WilliamG]

Just out of curiosity I decided to run this benchmark on my 2013 iMac i5 3.4 nVidia 775M with all the same settings you guys suggested.

GPU temp according to the program hovered around 80c (82c max) so the fan never increased from 1200rpm.

That I don't find surprising. What I do find surprising is the results of FPS and score. Shouldn't you guys be in an entirely different ballpark? Did I test something incorrectly?

Image

Image

EDIT : I obviously didnt take part in the poll as I don't have a 5k iMac FYI. Also to point out I've had the GPU get much hotter with games (ESO 95c), so if you guys are seeing throttling with this benchmark I'd imagine you'd certainly see it real world.

It's already been stated (by me at least), that the 295 is hugely disappointing. There are many out there that deny it, but.... that's how it is.

That said, some apps show an improvement for the AMD card vs the Nvidia card, and vice versa. This might be the vice versa. In any case, yes, the 295 is a pile of poop. Just the way it goes. I'll probably upgrade my 5K next refresh to the "real" version.

I know it sounds harsh, but I'm not using rose-tinted glasses when it comes to electronics. The screen is gorgeous and I don't regret getting the 5K, but I'm not going to kid myself - I'm disappointed in the GPU performance (and heat) MASSIVELY.

 

[andy9l]

It's already been stated (by me at least), that the 295 is hugely disappointing. There are many out there that deny it, but.... that's how it is.

That said, some apps show an improvement for the AMD card vs the Nvidia card, and vice versa. This might be the vice versa. In any case, yes, the 295 is a pile of poop. Just the way it goes. I'll probably upgrade my 5K next refresh to the "real" version.

I know it sounds harsh, but I'm not using rose-tinted glasses when it comes to electronics. The screen is gorgeous and I don't regret getting the 5K, but I'm not going to kid myself - I'm disappointed in the GPU performance (and heat) MASSIVELY.

One of the most accurate and honest posts I've seen on the subject. 100% agree.

I am returning my current 5K now in order to get the next, "real", one. I cannot justify buying two in 12 months, and I can't be bothered with the hassle and risk of selling one...if it even lasts long enough before failing!

I've been spending time backing up and erasing my current one - man I'm going to miss that screen.

 

[Brookzy]

Is this really a big deal?

The sound of the fan is somewhat annoying but reasonable when you consider the GPU is pushing four times the pixels.

The throttling seems very rare. I've not experienced it when gaming and most posters on this thread haven't hit 107 degrees.

And if the heat does damage the components then everyone has a minimum of one year's warranty. If the components were going to fail due to heat alone I'm sure it'll fail in a year or less.

I'm sure Apple engineers know enough of what they're doing to know what temps their GPUs hit and if it's safe.

 

[melrich]

Consistent at 98 - 99 temp and between 15 - 26 fps.

i7, 295.

 

[andy9l]

Is this really a big deal?



The sound of the fan is somewhat annoying but reasonable when you consider the GPU is pushing four times the pixels.



The throttling seems very rare. I've not experienced it when gaming and most posters on this thread haven't hit 107 degrees.



And if the heat does damage the components then everyone has a minimum of one year's warranty. If the components were going to fail due to heat alone I'm sure it'll fail in a year or less.



I'm sure Apple engineers know enough of what they're doing to know what temps their GPUs hit and if it's safe.

No-one knows for sure yet. Potentially, it is.

That said, it sounds like it wouldn't really bother you if it was, and you are happy with your product. If that's the case, just enjoy it!

How do you know you've not experienced throttling? If you're running a game at 100 FPS. When it throttles, it will only drop to 90 or so. As a human, you're unlikely to notice that. You need to measure the clock speed/FPS to see it.

So then you could argue, "so if you can't notice it, who cares". That's fine, if that's your view then the 5K iMac is perfect for you. For me, though, I don't appreciate paying for something I don't get. I paid for that GPU, and I get to use its power for about 60 seconds. In a year or two, that missing performance will be needed more and more. Not cool.

I have managed to get my M295X to 108C on Windows without really trying, and 106C on Mac. The temperatures are consistently maintained and fall outside of the operating temperature of the card. Consistently running electronics outside of the safe temperature, at least in the past, has not led to positive things.

Sure, I could keep it, let it fail and use my AppleCare to get components replaced. But besides the hassle, that doesn't solve the problem. I will still hit 95-100C and hear the fans when watching some 1080p movies, I will still be throttled when doing any intensive task, I still won't be getting the advertised performance. The 2013 iMac is outperforming it in this case. I'm not happy with that.

Looking at the lawsuits and failing AMD cards in the 2011 iMac/MBP suggests otherwise, with regards to your comment about Apple engineers.

When I bought the iMac I said to myself, "anything short of perfect, and it's going back". My returning of it is me keeping that promise, nothing more.

 

[Alesc]

To note I use TG Pro for temps and the 82c I saw in the benchmark program was also reflected in TG Pro at 82c.

Its curious iStat would display differently unless it using a proximity sensor or an average?

EDIT: Took some screen shots.

Waited for the benchmark program to get to 80c.

Image

I couldn't get out of the benchmark program fast enough to get a live temp so I reset TG Pro maximum recorded temps before running the benchmark. So you can see it reports the max it got to was 80c.

Image

And of course the fan still putting along at <1200rpm.

Thanks a lot for your feedback! To view the live temp from other apps during the bench, I do cmd+Tab: it works and you can see the live temp.
I've reported the difference to Bjango, the editor of iStat, I keep you posted.

 

[thundersteele]

One of the most accurate and honest posts I've seen on the subject. 100% agree.

Not very honest imho. Poster is "massively disappointed" with the machine, had the option to return it for a full refund most likely (Apple holiday extended return period), and still keeps the machine. So it must really be a terrible experience to own this machine...

Also the purpose of this thread was to compare benchmark, we already have at least one other thread to whine about the temperature of the M295X...

 

[andy9l]

Not very honest imho. Poster is "massively disappointed" with the machine, had the option to return it for a full refund most likely (Apple holiday extended return period), and still keeps the machine. So it must really be a terrible experience to own this machine...



Also the purpose of this thread was to compare benchmark, we already have at least one other thread to whine about the temperature of the M295X...

No.

William said he is massively disappointed with the M295X, not the entire machine. The heat is entirely relevant as that's causing the throttling. Hence the benchmarking. Hence this thread.

 

[Brookzy]

I still won't be getting the advertised performance.

When I bought the iMac I said to myself, "anything short of perfect, and it's going back". My returning of it is me keeping that promise, nothing more.

Fair enough, if you'll notice the occasional slight throttling then send it back.

But, what is the 'advertised performance'? AFAIK, the graphics chip in this was unreleased before the 5K, so we don't have another computer to compare it to. I'm sure Apple quoted a figure for processing power somewhere - now, if after throttling you get less than Apple's quoted performance, then that's false advertising and we should all be pissed. But as it is, I don't think Apple has misled anyone.

 

[cynics]

Is this really a big deal?

The sound of the fan is somewhat annoying but reasonable when you consider the GPU is pushing four times the pixels.

The throttling seems very rare. I've not experienced it when gaming and most posters on this thread haven't hit 107 degrees.

And if the heat does damage the components then everyone has a minimum of one year's warranty. If the components were going to fail due to heat alone I'm sure it'll fail in a year or less.

I'm sure Apple engineers know enough of what they're doing to know what temps their GPUs hit and if it's safe.

I would say its a pretty big deal.

This benchmark shows that at a given resolution (2560x1440) a 2013 mid level iMac has similar performance, with a lot less noise at a fraction of the price.

Thats not right IMO.

While I agree the screen itself is infinitely better I feel at the price point and knowing its an "upgrade" the retina iMac should be better then that.

This is becoming very frustrating. First I wait forever for a new Mac Mini and end up being left down (for my uses). Now I look into a retina iMac and I'm let down again (again for my uses). Oh well.

 

[odedia]

I've noticed the temperature stay pretty much the same even at 1600x900 resolution with no anti aliasing and all check boxes off. The frame rate jumps to around 90 fps but the heat is still at similar levels.

 

[kgapp]

i7 4.0, 295, 16gb ram, 1tb SSD

FPS: 17-30
Temps: flip flopped between 99 and 100 (iStat showed 101)
fan: 2300 rpm


I do not believe that the previous posters comparing the performance to a last generation iMac at the same resolution is a realistic comparison. Both machines are running at 2560X1440 but with the iMac that is a retina resolution which obviously means a whole lot more pixels are being pushed around. While they may yield a similar FPS that single statistic is clearly not telling the whole story.

If two vehicles both had the same 0-60 time but one weighed 2000 lbs more and was more luxurious and contained higher tech would you say that there is no increase in power in the heavier car? The 'performance' may be similar but the 'power' required to achieve that performance is certainly quite different.

There will always be some compromises as an early adopter to new technology. The Retina iMac is no exception to that general rule. If someone is not prepared to accept those compromises for early access to new technology than like anything else it is always best to wait until a second or third generation product is released. I am sure that most of those complaining likely already knew this to be true before they even clicked the 'buy' button for a Retina iMac.

 

[Alesc]

I would say its a pretty big deal.

This benchmark shows that at a given resolution (2560x1440) a 2013 mid level iMac has similar performance, with a lot less noise at a fraction of the price.

Thats not right IMO.

While I agree the screen itself is infinitely better I feel at the price point and knowing its an "upgrade" the retina iMac should be better then that.

This is becoming very frustrating. First I wait forever for a new Mac Mini and end up being left down (for my uses). Now I look into a retina iMac and I'm let down again (again for my uses). Oh well.

Is this bench very reliable? All graphic cards (775, 780, 290 and 295) are in a nutshell, that's weird. And what is weirder: the 775 with core i5 performs better than the 780 with i7...

Both machines are running at 2560X1440 but with the iMac that is a retina resolution which obviously means a whole lot more pixels are being pushed around.

I disagree: both machine moves 2560x1440px in that case, they are the same bases. If you could ran the bench at retina 5k resolution, I seriously doubt the FPS would be the same.

There will always be some compromises as an early adopter to new technology. The Retina iMac is no exception to that general rule. If someone is not prepared to accept those compromises for early access to new technology than like anything else it is always best to wait until a second or third generation product is released. I am sure that most of those complaining likely already knew this to be true before they even clicked the 'buy' button for a Retina iMac.

I agree with you on that part.

 

[kgapp]

I disagree: both machine moves 2560x1440px in that case, they are the same bases. If you could ran the bench at retina 5k resolution, I seriously doubt the FPS would be the same.

I believe that this is completely flawed logic. My understanding is that on a Retina iMac when you are running in 2560x1440 you are still moving around 14.7 million pixels (even if you are not in 5120 x 2880 resolution). That is how it is able to achieve a significantly crisper 'retina' display.

If you weren't using all 14.7 million pixels on the display than 3 out of every 4 pixels would be black. That is certainly not the case.

 

[cynics]

i7 4.0, 295, 16gb ram, 1tb SSD

FPS: 17-30
Temps: flip flopped between 99 and 100 (iStat showed 101)
fan: 2300 rpm


I do not believe that the previous posters comparing the performance to a last generation iMac at the same resolution is a realistic comparison. Both machines are running at 2560X1440 but with the iMac that is a retina resolution which obviously means a whole lot more pixels are being pushed around. While they may yield a similar FPS that single statistic is clearly not telling the whole story.

If two vehicles both had the same 0-60 time but one weighed 2000 lbs more and was more luxurious and contained higher tech would you say that there is no increase in power in the heavier car? The 'performance' may be similar but the 'power' required to achieve that performance is certainly quite different.

There will always be some compromises as an early adopter to new technology. The Retina iMac is no exception to that general rule. If someone is not prepared to accept those compromises for early access to new technology than like anything else it is always best to wait until a second or third generation product is released. I am sure that most of those complaining likely already knew this to be true before they even clicked the 'buy' button for a Retina iMac.

One of the "upgrades" (I clearly quoted that before) is the 290 to 295. As posted on here has dangerously similar performance. Both have retina screens.

Save me the car analogies they never work. Yours is completely fictitious. Car manufacturers take this into account, if more luxury items are included with an upgraded motor it still offers premium performance. Any A to S Audi, any AMG upgrade from Mercedes, And M package upgrade from BMW. Car manufacturers aren't that daft they would sell you a more powerful machine with more luxury items at a premium price and it have the same performance as the cheaper model....enter Apples business strategy? Strange.

I feel people like you and I know better then to be early adopters (unless you are fine with the product offered) but your average customer doesn't.

This is just an iPad 3 all over again (I still have one). Pretty screen put on tech that can hardly support it. Without fan assisted cooling devs have to literally dumb down the graphics in games so it would run right. Even native Apple apps like maps in 3D flyover run like poop. They burned me even worse because I wasn't an early adopter so I bought it and just out of return Apple dropped the iPad 4. Pisses me off thinking about it.

Anyway thats before I knew I had to do tons of research before buying an Apple product. Now I know better.

EDIT: Also like to point out when I'm comparing my iMac to Macs with lower resolutions mine still benchmarks noticeable faster.