The new iMac is a lot noisier!

[snap01]

Yes. I was editing my RAWs in LR last night and even after about 30 minutes int he Develop Module, the fans would kick in as I moved around quickly from photo to photo. The RPMs would go from 1800 to 2400 depending on how quickly I was going from pic to pic. That really surprised me as Photos would not do that once it generated previews from the same RAWs. That definitely confirmed the the i7 heats up quickly and is not the kind of experience I'm after.

In the develop module, Lightroom displays the preview initially but then reprocesses the RAW file to give the most accurate display of the image. So the previews don't reduce the computation load and LR is likely doing more processing than Photos (unless photos also reprocesses the image).

I've been working in LR with my i7 and I've only heard the fans kick in once or twice over about two days of usage. Once was when I switched to a folder I hadn't viewed in a long time (i.e. no previews) and I assume LR started generating them in the background. This is was a little strange to me because the fans haven't kicked in on other imports I've done. I think the reason is that my existing folder had a bunch of large PS files (with multiple layers) that LR needed to render.

I think the fans also kicked in when I was doing a large export. But it doesn't always kick in on export. The first test I did when I got my iMac was to export ~120 large raw files and I didn't hear the fan at all.

 

[propower]

Well, I went back to my old Hex core MP and even added a 4K screen to it and in the end I now think I would rather modify my workflow a little and stick to the original plan of new iMac for now. Even it is an interim one that gets replaced with one of the newer models coming. I put the MacPro on the local market for a couple of weeks and this may be a bit bleaker than I had hoped. Once there are new models out there - I fear that will sink considerably lower.

SO - now I have a i5/3.8/580 stock model to try out . For audio these things matter most - cores, random access, CPU Speed. I will mimic a mostly SSD system by keeping the install small an making a partition of the Fusion drive that is ~1.95TB - and leave that blank. Should mimic a mostly SSD iMac. I will do my usual round of YES and Audio tests but mostly I am just going to work this one for a few weeks and if it flies will order one of the i5 variants with just SSD storage. I am not much of a video guy but I am happy to run any simple standard test for the curious. I believe the 7600 and 7600X are the same Silicon - just different bins (could not confirm this though). I anticipate that with Turbo OFF this should run a smidgen cooler than the 7600. With Turbo ON - a smidgen more. In the end I have a hard time seeing the real value in the 7600X wheras the 580 has a more clear advantage - though I am not sure when that matters! LOL...

 

[dsc888]

In the develop module, Lightroom displays the preview initially but then reprocesses the RAW file to give the most accurate display of the image. So the previews don't reduce the computation load and LR is likely doing more processing than Photos (unless photos also reprocesses the image).

I've been working in LR with my i7 and I've only heard the fans kick in once or twice over about two days of usage. Once was when I switched to a folder I hadn't viewed in a long time (i.e. no previews) and I assume LR started generating them in the background. This is was a little strange to me because the fans haven't kicked in on other imports I've done. I think the reason is that my existing folder had a bunch of large PS files (with multiple layers) that LR needed to render.

I think the fans also kicked in when I was doing a large export. But it doesn't always kick in on export. The first test I did when I got my iMac was to export ~120 large raw files and I didn't hear the fan at all.

Thanks for the explanation. And that does make sense that the fans spin up for on demand previews. But it's a noise I never encountered with my 2012 rMBP or my 2010 MP so it's something I immediately noticed and didn't like. I have iStat and I saw the RPMs jump between 1800 and 2400 as I quickly thumbed through my 900+ RAWs. Again, it is normally a noiseless event on my laptop and low, constant hum with the MP.

 

[EugW]

Well, I went back to my old Hex core MP and even added a 4K screen to it and in the end I now think I would rather modify my workflow a little and stick to the original plan of new iMac for now. Even it is an interim one that gets replaced with one of the newer models coming. I put the MacPro on the local market for a couple of weeks and this may be a bit bleaker than I had hoped. Once there are new models out there - I fear that will sink considerably lower.

SO - now I have a i5/3.8/580 stock model to try out . For audio these things matter most - cores, random access, CPU Speed. I will mimic a mostly SSD system by keeping the install small an making a partition of the Fusion drive that is ~1.95TB - and leave that blank. Should mimic a mostly SSD iMac. I will do my usual round of YES and Audio tests but mostly I am just going to work this one for a few weeks and if it flies will order one of the i5 variants with just SSD storage. I am not much of a video guy but I am happy to run any simple standard test for the curious. I believe the 7600 and 7600X are the same Silicon - just different bins (could not confirm this though). I anticipate that with Turbo OFF this should run a smidgen cooler than the 7600. With Turbo ON - a smidgen more. In the end I have a hard time seeing the real value in the 7600X wheras the 580 has a more clear advantage - though I am not sure when that matters! LOL...

7600K. There is no 7600X but actually there is a 7640X. However, that’s a 112 W TDP chip.

Looking forward to your tests! I wonder if under full load in Logic or RAW previewing in Photos / Lightroom if the fan will hit 1500+ rpm.

 

[dsc888]

Looking forward to your tests! I wonder if under full load in Logic or RAW previewing in Photos / Lightroom if the fan will hit 1500+ rpm.

I too am very curious to find out. I don't want a fan that ramps up randomly with multiple previews.

 

[snap01]

Thanks for the explanation. And that does make sense that the fans spin up for on demand previews. But it's a noise I never encountered with my 2012 rMBP or my 2010 MP so it's something I immediately noticed and didn't like. I have iStat and I saw the RPMs jump between 1800 and 2400 as I quickly thumbed through my 900+ RAWs. Again, it is normally a noiseless event on my laptop and low, constant hum with the MP.

Interesting. On my 2013 rMBP the fans come on all the time in Lightroom. And it's loud and noisy but it's an i7 too :-> I really dislike having the fans come on in my laptop but I thought it was unavoidable. At the time I got my laptop, it was my only computer so I went with i7 for power.

Right now (on iMac i7) I'm having Lightroom generate a bunch of previews for a morning shoot (about 150 - 40MB RAWs). If just tab out and do something else while previews are generating and the fan doesn't come on. However, If I stay in Lightroom and try and do other stuff then the fan comes on. So it must be right on the cusp of too much heat to dissipate. This is in a relatively warm room.

 

[czacha]

Hi! I went through all 23 pages .. 23 pages of noise and heat I'm about to order 2017 iMac 27'' 5K and for the most of time I was going to get i7 / R580 with 512SSD. But after this thread I'm not sure what to do .. like most of you after the same lecture Right now i'm using hackintosh with i7 6700K .. but I miss iMac

Most owners of iMac i5 says - fans stays idle even at stress. Honestly I'm not sure if that's good. Hitting 80*C and no fans somehow doesn't makes me feel comfortable. When I had 2011 iMac I used SMC Fan Control and set own rules so fans could do their work and keep CPU at normal temperature while generating 4000 or 5000 1:1 previews in Litgroom or exporting 500 JPEGs from RAW files. I didn't want to wait for CPU to hit high temperature isntead of that I prefered to forced fans to spin faster than someone from Apple set them to.

But you guys would prefer hotter CPU than hear some fan noise.. well that's interesting afterall Always though that at some point CPU reduces it's speed to protect from overheating - am I wrong? So my question is - let's say that 40*C is idle temp for i7 with 1200RPM. So if you set 1400RPM (or even 1600RPM) for 55-65*C will this prevent from hitting higher temps or at least will delay reaching top temp and RPM in time so for longer it will work with turbo boost? That worked with my old iMac. It was a bit louder but after all it suppose to be workstation which should work fast .. at least in my case. I'm a professional photographer so I'm looking for fast computer. I know that Lightroom isn't fastest tool and at the moment it can't utilize power of i7 .. but maybe in 2 or 3 years something will change? I plan to use iMac for at least 5 years just like I did with old one.

Whish there were more vidoes with fan noise 'cause at the moment all we have is one video on youtube... that's poor I must say.

PS. Sorry for my english

 

[ZMacintosh]

I have the i7, no heat or noise issues thus far and I did a huge raw photo import, and even maxed the resolution, so far one of the best iMacs I've be owned yet

I will mimic a mostly SSD system by keeping the install small an making a partition of the Fusion drive that is ~1.95TB - and leave that blank. Should mimic a mostly SSD iMac.

Might as well break the fusion set up might be easier and more fluent with your testing

 

[propower]

A quick YES test with external monitors ON in the end...
In general for a single iMac screen user pretty much would never get Fan Noise with this top tier i5 though temps will get up into the upper 70deg zone at least. Clearly more GPU in addition to maxed out CPU brings temps up a bit further. 100% CPU plus driving external monitors via a Thunderbolt 2 chain got me in the 80s. Turning Turbo off brought the CPU from 4GHz to 3.8GHz but saved about 10degC. All of this in a well controlled 22degC room.

I5 3.8GHz Turbo ON (quick test)
1 YES = 19.5W CPU 4.05 GHz, 54degC
2 YES = 29W CPU, 13W GPU, 4.02 GHz, 62degC
3 YES = 40W CPU, 13W GPU, 4.0 GHz, 70degC
4 YES = 46.4W CPU, 14W GPU, 4.0 GHz,77degC

Turbo OFF on 4 YES —> 39W, 3.8GHz, 71deg
Turbo OFF + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 41W, GPU 29W, 3.8GHz, 76degC

Turbo ON + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 51W, GPU 29W, 3.8GHz, 86degC

FWIW - putting the Fan to 2000 at 1/2 load made a measly 4degC change in CPU temp. In my work - no point in that. Also makes me wonder why there is so little change for nearly doubling fan speed! Has anyone ever run these without the memory cover on to increase air flow. I have been wondering about that and putting a SilenX fan to pull air out of the case over an open memory compartment. Thermally the airflow path (especially on the exhaust) seems a bit meager on these.

First impressions for audio work is strangely zero difference between this one and the base model i5 3.4GHz --- wasn't expecting that! When does the extra oomph of the 575 or 580 make a difference?

 

[dsc888]

A quick YES test with external monitors ON in the end...
In general for a single iMac screen user pretty much would never get Fan Noise with this top tier i5 though temps will get up into the upper 70deg zone at least. Clearly more GPU in addition to maxed out CPU brings temps up a bit further. 100% CPU plus driving external monitors via a Thunderbolt 2 chain got me in the 80s. Turning Turbo off brought the CPU from 4GHz to 3.8GHz but saved about 10degC. All of this in a well controlled 22degC room.

I5 3.8GHz Turbo ON (quick test)
1 YES = 19.5W CPU 4.05 GHz, 54degC
2 YES = 29W CPU, 13W GPU, 4.02 GHz, 62degC
3 YES = 40W CPU, 13W GPU, 4.0 GHz, 70degC
4 YES = 46.4W CPU, 14W GPU, 4.0 GHz,77degC

Turbo OFF on 4 YES —> 39W, 3.8GHz, 71deg
Turbo OFF + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 41W, GPU 29W, 3.8GHz, 76degC

Turbo ON + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 51W, GPU 29W, 3.8GHz, 86degC

FWIW - putting the Fan to 2000 at 1/2 load made a measly 4degC change in CPU temp. In my work - no point in that. Also makes me wonder why there is so little change for nearly doubling fan speed! Has anyone ever run these without the memory cover on to increase air flow. I have been wondering about that and putting a SilenX fan to pull air out of the case over an open memory compartment. Thermally the airflow path (especially on the exhaust) seems a bit meager on these.

First impressions for audio work is strangely zero difference between this one and the base model i5 3.4GHz --- wasn't expecting that! When does the extra oomph of the 575 or 580 make a difference?

Wow! Thank you so much for your detailed test and results. Now I'm wondering if I should've ordered the top-tier I5 with a better GPU instead. I will only use the main display and so I am very unlikely to get the processor hot enough for the fan to come on except if I encode video in the few times that I will edit video. I'm mostly a PS/LR photo guy. Decisions. Decisions. I honestly have no use for the 580 over the 575 but it's nice to have the extra GPU power. Too bad Apple's online order checker is down at the moment...
[doublepost=1499634826][/doublepost]

Interesting. On my 2013 rMBP the fans come on all the time in Lightroom. And it's loud and noisy but it's an i7 too :-> I really dislike having the fans come on in my laptop but I thought it was unavoidable. At the time I got my laptop, it was my only computer so I went with i7 for power.

Right now (on iMac i7) I'm having Lightroom generate a bunch of previews for a morning shoot (about 150 - 40MB RAWs). If just tab out and do something else while previews are generating and the fan doesn't come on. However, If I stay in Lightroom and try and do other stuff then the fan comes on. So it must be right on the cusp of too much heat to dissipate. This is in a relatively warm room.

Yes. It's the previews that cause the i7 to run hot and the fans to spin up.

 

[T_Oscura]

In general for a single iMac screen user pretty much would never get Fan Noise with this top tier i5 though temps will get up into the upper 70deg zone at least. Clearly more GPU in addition to maxed out CPU brings temps up a bit further. 100% CPU plus driving external monitors via a Thunderbolt 2 chain got me in the 80s.

So it's 20 deg cooler than the i7? That's quite a lot.

 

[propower]

So it's 20 deg cooler than the i7? That's quite a lot.

It is a lot cooler than the i7! The i7 maxed the fans out at 50% load (4 YES test - the i7 can do 8 YES). For the same load (4 YES) the i5/3.8 (Turbo OFF) is at 100% CPU load and 20 deg Cooler. I believe what we are seeing is that for this cooling system each 100MHz is taking us up an increasing larger amount in temp. 3.8GHz (Turbo OFF) on this one to 4GHz (Turbo ON) brought things up 10degC. The i7 runs 4.2GHz with Turbo OFF and has 20degC higher temp and full speed fans. With the i7 and Turbo ON - I can easily see short spikes in activity heating the CPU up very fast and kicking the fans. This is what my old 2013 MBP i7 did with the lid down and ext monitor - drove me crazy!

For my uses the i7 would be something like 25% load usually and likely be 1200 RPM but still running in the 70s to 80s in temp. For any full on task - Video Render etc -- 2200 to max fan every time. It may be simpler to just use the i5 and render any heavy CPU tracks as I go. Ask me in a couple of weeks If this goes as I expect I am once again thinking of saving $$ and going base model (i5 3.4GHz) with the only upgrades being a 512 or more likely 1TB SSD + 16 more gigs aftermarket RAM. The Mid Tier looks nice as well but I have to see a real gain from any of the ones on paper to make it compelling. From my quick reading - for us non-gamers and at the most 4K weekend video editors 4G Video Ram is plenty and the 570 is better than the last versions top end - or my D300 Mac Pro. 575 is faster - but when does it matter?

 

[Smoothie]

Just maxed out all four cores on my i5-7600K. Maximum temperature was 75 deg with ambient temp at 23 deg. Fans stayed at idle. I'm very happy with the result.

 

[EugW]

A quick YES test with external monitors ON in the end...
In general for a single iMac screen user pretty much would never get Fan Noise with this top tier i5 though temps will get up into the upper 70deg zone at least. Clearly more GPU in addition to maxed out CPU brings temps up a bit further. 100% CPU plus driving external monitors via a Thunderbolt 2 chain got me in the 80s. Turning Turbo off brought the CPU from 4GHz to 3.8GHz but saved about 10degC. All of this in a well controlled 22degC room.

I5 3.8GHz Turbo ON (quick test)
1 YES = 19.5W CPU 4.05 GHz, 54degC
2 YES = 29W CPU, 13W GPU, 4.02 GHz, 62degC
3 YES = 40W CPU, 13W GPU, 4.0 GHz, 70degC
4 YES = 46.4W CPU, 14W GPU, 4.0 GHz,77degC

Turbo OFF on 4 YES —> 39W, 3.8GHz, 71deg
Turbo OFF + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 41W, GPU 29W, 3.8GHz, 76degC

Turbo ON + Thunderbolt chain + 2nd/3rd Monitors ON
CPU = 51W, GPU 29W, 3.8GHz, 86degC

FWIW - putting the Fan to 2000 at 1/2 load made a measly 4degC change in CPU temp. In my work - no point in that. Also makes me wonder why there is so little change for nearly doubling fan speed! Has anyone ever run these without the memory cover on to increase air flow. I have been wondering about that and putting a SilenX fan to pull air out of the case over an open memory compartment. Thermally the airflow path (especially on the exhaust) seems a bit meager on these.

First impressions for audio work is strangely zero difference between this one and the base model i5 3.4GHz --- wasn't expecting that! When does the extra oomph of the 575 or 580 make a difference?

Thanks propower!

Did you do any other 100% CPU tests, and if so, what did you do? How long did it take to get to 86C?

With my 7600 running an h.265 encode in Handbrake, the temp with 100% CPU usage initially maxed at 69C and then went into the low 70s after not too long. However it took a long time to hit 80C and after I quit the Handbrake h.265 encoding test after 10 mins of continuous 100% usage it was at 83C.

 

[EugW]

Just maxed out all four cores on my i5-7600K. Maximum temperature was 75 deg with ambient temp at 23 deg. Fans stayed at idle. I'm very happy with the result.

How long did you run this, and what were you running?

 

[dsc888]

Just maxed out all four cores on my i5-7600K. Maximum temperature was 75 deg with ambient temp at 23 deg. Fans stayed at idle. I'm very happy with the result.

Nice. What apps were you using to max it out?

 

[czacha]

Just maxed out all four cores on my i5-7600K. Maximum temperature was 75 deg with ambient temp at 23 deg. Fans stayed at idle. I'm very happy with the result.

You got 2x higher temp than idle temperature with 100% CPU load and yet no cooling in progress.. I'd say its odd.

From what you and other wrote it looks like Apple knows that cooling fan gives some noise (d'oh!) so they pushed away moment when fan start to do it's job it was made for - cool that hot CPU

I bet you like this but is this safe? Mine Hackintosh i7 6700K goes 55*C max with full load... so I'd say 80*C is rather high temperature Perhaps you should help those fans and let them start spin a bit faster with SMC fan control?

 

[Smoothie]

Nice. What apps were you using to max it out?

I just use the "yes" routine running in the console - one for each of the cores.

 

[dsc888]

I just use the "yes" routine running in the console - one for each of the cores.

Thanks, smoothie. I will give that a try when I get my new machine. My i7 needs no help from me to get the fans going

 

[Smoothie]

How long did you run this, and what were you running?

Just the /dev/null "yes" routine. I ran it until the temps stopped going higher.
[doublepost=1499647295][/doublepost]

You got 2x higher temp than idle temperature with 100% CPU load and yet no cooling in progress.. I'd say its odd.

From what you and other wrote it looks like Apple knows that cooling fan gives some noise (d'oh!) so they pushed away moment when fan start to do it's job it was made for - cool that hot CPU

I bet you like this but is this safe? Mine Hackintosh i7 6700K goes 55*C max with full load... so I'd say 80*C is rather high temperature Perhaps you should help those fans and let them start spin a bit faster with SMC fan control?

Well, there is cooling. The fan is rotating at 1200 rpm and exhausting hot air. Yes, Apple values low noise over high heat, so the fan routine doesn't usually ramp up until the temp gets pretty high and stays there for a certain length of time. But 75 deg. C (which was the max in my test) won't hurt the CPU, which has a Tjunction temp of 100 deg. My homemade PC goes to 70 deg. C when being tortured with Prime 95, and the Intel 4790K CPU is air-cooled with a Noctua cooler. I'm actually surprised the iMac did as well as it did.

 

[EugW]

Just the /dev/null "yes" routine. I ran it until the temps stopped going higher.

Nice. How long was that?

So it seems you got a good 7600K. Judging by the reports out there, the results of the 7600K may be variable. AnandTech had a 7600K that behaved similarly to 65 Watt TDP chips, Toms had a 7600K that did fairly well too, but they could get much higher than the 7600 under very extreme conditions.

 

[Smoothie]

Nice. How long was that?

So it seems you got a good 7600K. Judging by the reports out there, the results of the 7600K may be variable. AnandTech had a 7600K that behaved similarly to 65 Watt TDP chips, Toms had a 7600K that did fairly well too, but they could get much higher than the 7600 under very extreme conditions.

I didn't time the test, but it was no more than ten minutes by the time the temps stabilized. I assume the /dev/null test is enough of a stress. All four cores were at 100 percent. I tried it on my old 2009 MBP with a dual core CPU, and it hit 85 deg. with the fan on full blast. I didn't want to let it go any longer than that.

 

[propower]

The Apple algorithm for cooling is - wait till CPU hits ~95degC and ramp fans to keep it more or less there. Been that way on every Mac I have had since 2011. The Older MacPros were so cool running I never heard the fan ramp.

The extra 10degC for running a mirrored monitor chain is always a concern for me with the iMacs. But - from what I have seen the difference in performance and/or heat from mid tier to top tier is very small and can be mitigated in the top tier by turning off Turbo. The only real reason I see for someone to prefer the top tier is the GPU.

 

[EugW]

I didn't time the test, but it was no more than ten minutes by the time the temps stabilized. I assume the /dev/null test is enough of a stress. All four cores were at 100 percent. I tried it on my old 2009 MBP with a dual core CPU, and it hit 85 deg. with the fan on full blast. I didn't want to let it go any longer than that.

Ten mins seems reasonable. Before I read your message, I tried it for 5 mins and then stopped it, since temps had almost stopped increasing.

2.6 mins (71C):

5.0 mins (74 C):

I then saw your message saying it took around 10 mins, so I re-ran it again, this time for the full 10 mins (76C):

Yep, by that time, the temps had basically plateaued. And my result also in a 23C room was pretty much identical to yours. I was at 75-76C. Power utilization 44 Watts.

However, I'm not sure that test is as stressful as a Handbrake h.265 encode. With that test after 10 mins I got the CPU to about 83C, and with considerably more power utilization, at 55 Watts:

BTW, for those who care, the command is this (in the terminal):

yes > /dev/null & yes > /dev/null & yes > /dev/null & yes > /dev/null &

And to stop the test, type this:

killall yes

So our 7600K and 7600 seem to behave similarly. Although the 7600K is a potentially higher power chip, it's quite possible it may be as cool as the 7600, which isn't a big surprise since for heavily multi-threaded applications, it only is about 3% faster in clockspeed, or 2-7% faster in some real world multi-core benches.

However, for the 7600K, it may be more of a lottery. Smoothie seems to have gotten a good one, on par with my 7600, and probably within the 65 W TDP class envelope (even though it's a 91 W TDP class chip), but YMMV. Here are Tom's results under heavy and extreme load. 7600K on left and 7600 on right:

 

[propower]

So our 7600K and 7600 seem to behave similarly. Although the 7600K is a potentially higher power chip, it's quite possible it may be as cool as the 7600, which isn't a big surprise since for heavily multi-threaded applications, it only is about 3% faster in clockspeed, or 2-7% faster in some real world multi-core benches.

However, for the 7600K, it may be more of a lottery. Smoothie seems to have gotten a good one, on par with my 7600, and probably within the 65 W TDP class envelope (even though it's a 91 W TDP class chip), but YMMV. Here are Tom's results under heavy and extreme load. 7600K on left and 7600 on right:

View attachment 707922

The 7600K is unlocked for clock speed. You can run it as fast as you can a) keep it stable, b) keep it below thermal runaway. I can blow up a 125W Mosfet by running it without a heatsink. Or I can thermally cool it and run - 125W (I designed power supplies and audio gear for many years...). I believe the higher rating for both the K parts is about maximum rating even with stellar cooling - or perhaps what the equvalent power for the maximum useable frequency is. The home PC builders mostly pride themselves on 60degC max temp at whatever Overclocking they desire. The rating on the chip I believe means the same - like - maybe it can do 5GHz if you can get the heat out. I do not think it means that this chip is all the sudden 91W at the same freq as the non-K version