How hot does the i9 run vs i7? (15" MBP 2018)

[Shadow Puppets]

Title says it all really.

I'm in the market for a 15" MBP, and trying to choose between specs. Definitely want 32GB RAM, but unsure on which processor to pick. 85% of the computers work load will be basic stuff... emails, browsing, light audio editing, Football Manager , but the rest of the time I'll be doing some more intensive video editing (4K) and rendering.

I know, in theory, the i9 is the way to go and the more future proof, but how hot does it run? Particularly, does it run notably hotter even when doing simple tasks? I don't want to be sat with a heater on my knee while I'm doing fairly basic workflows.

Interested in your thoughts on this.

 

[spintronics]

I can't find reliable info about this anywhere. I can't choose between the i7 2.6 and i9. Really afraid of the temperature especially with external display connected. It's okay to run hot under heavy load but I'm more interested what is happening during browsing or watching Netflix on a 4K external display. Have you found any reliable info about this?

 

[Webster's Mac]

I can't find reliable info about this anywhere. I can't choose between the i7 2.6 and i9. Really afraid of the temperature especially with external display connected. It's okay to run hot under heavy load but I'm more interested what is happening during browsing or watching Netflix on a 4K external display. Have you found any reliable info about this?

The main heat/throttling issue was due to a bug in the firmware and was patched. Under normal load, it probably won't run much hotter than the i7.

 

[spintronics]

The main heat/throttling issue was due to a bug in the firmware and was patched. Under normal load, it probably won't run much hotter than the i7.

This reddit thread worries me. Is the dGPU always make the laptop uncomfortably hot? Now I'm more concerned about all the 2018 MBP 15 configs than the difference between the i7 and i9 cpus. I have no problem with high temperature under heavy load but it's strange when the computer in idle connected to external display or when just it's keeping open some IDEs.

 

[jerryk]

This reddit thread worries me. Is the dGPU always make the laptop uncomfortably hot? Now I'm more concerned about all the 2018 MBP 15 configs than the difference between the i7 and i9 cpus. I have no problem with high temperature under heavy load but it's strange when the computer in idle connected to external display or when just it's keeping open some IDEs.

No mystery here. For high res monitor, More pixels to push = more work. And for IDEs a lot of them are doing things like constantly looking for file changes that trigger a build of the project. Yet another reason to use a simple IDE like VSCode, if you can.

 

[spintronics]

No mystery here. For high res monitor, More pixels to push = more work. And for IDEs a lot of them are doing things like constantly looking for file changes that trigger a build of the project. Yet another reason to use a simple IDE like VSCode, if you can.

Oh, I see, thanks. You have a 2018 15, right? Do you have any experience with 4K external monitors? Is the temperature really that uncomfortable? And if you have 2.6 can you tell me about the idle temperatures as well? It would help me to calm down a little bit.

 

[1096bimu]

Title says it all really.

I'm in the market for a 15" MBP, and trying to choose between specs. Definitely want 32GB RAM, but unsure on which processor to pick. 85% of the computers work load will be basic stuff... emails, browsing, light audio editing, Football Manager , but the rest of the time I'll be doing some more intensive video editing (4K) and rendering.

I know, in theory, the i9 is the way to go and the more future proof, but how hot does it run? Particularly, does it run notably hotter even when doing simple tasks? I don't want to be sat with a heater on my knee while I'm doing fairly basic workflows.

Interested in your thoughts on this.

The i9 isn't any hotter than the i7 and anyone who says so doesn't know how processors work.
They are equally cool when doing simple tasks and they both max out at 100 degrees when under full stress.

Processors with higher frequency ratings can run higher at the SAME POWER level, they are simply faster at NO ADDITIONAL energy cost. And they can do that because the fabrication process naturally produce processors with different qualities, the manufacturer can simply test and pick the best ones to be i9.

If anything, the i9 can complete the same tasks with LESS power and less heat.

 

[Thysanoptera]

Processors with higher frequency ratings can run higher at the SAME POWER level, they are simply faster at NO ADDITIONAL energy cost. And they can do that because the fabrication process naturally produce processors with different qualities, the manufacturer can simply test and pick the best ones to be i9.

Thats a pretty blank statement, which while holding some truth, is misleading in general, it may lead people to believe that magically i9@2.9 needs the same power as i7@2.2. While the difference is probably that i7 could do stable 4.6Ghz instead 4.8.

First - are you sure this is the same silicon? They have different cache organization. Of course Intel could degrade i9 to i7 by disabling partially cache and not only downclocking, but even then to take advantage of higher binned part you'd need to run with lower voltage for a given frequency. And so far I've seen no evidence that this is true in case of Apple and MBP. More, given how they messed up the simple throttling and fan curves I'm pretty darn sure they run it at stock voltages. And in this case the i9 will be hotter because of larger cache and higher voltage required to reach higher clocks. Aaaand - the differences between different binned parts would show up at the higher end of frequencies - since you can't run either higher than around 3.3GHz anyway fully loaded, binning is really a moot point.

The only person who could answer such question is somebody who works at Intel and knows their binning testing procedures and results for this particular chip. In case of MBP for all intents and purposes these are identical CPUs. Tell me what you want to prove and I will prepare a test suite that will show that i9 is cooler/hotter/faster/slower/louder/quieter, has better or worse battery life. Whatever you wish.

 

[matthewadams]

I have no problem with high temperature under heavy load but it's strange when the computer in idle connected to external display or when just it's keeping open some IDEs.

Well, for starters I found it weird that that reddit poster was talking about 10-15W usage of the dGPU. I get roughly 4W on my readings.
For what it's worth (this is with 1080p ext display connected via a Thunderbolt Dock charging/powering the 2018 MBP 15 2.6" 32GB in clamshell mode)

CPU Temp in idle with no monitor & power connected is roughly 48C (at 26C room temp).

Overall I haven't seen a massive difference in temperature compared to my 2013 rMBP.

 

[leman]

This reddit thread worries me. Is the dGPU always make the laptop uncomfortably hot?

Using an external monitor causes the dedicated GPU to be permanently on, which will of course make the laptop run hotter.

I have no problem with high temperature under heavy load but it's strange when the computer in idle connected to external display or when just it's keeping open some IDEs.

Why would you even care? Its not that you are using the computer on your lap when its connected to a dGPU...

 

[maflynn]

The i9 isn't any hotter than the i7 and anyone who says so doesn't know how processors work.

Actually they are running hotter, and the proof is in the pudding. The i9s throttle more (not just on the MBP) and run hotter. I may not know how processors work, but I can see first hand they run hotter

 

[Premal212]

Actually they are running hotter, and the proof is in the pudding. The i9s throttle more (not just on the MBP) and run hotter. I may not know how processors work, but I can see first hand they run hotter

So my 15inch 2018 i7/32/512 runs a lot hotter than my old 2016 13 i5 16gb. I don't mind the CPU running hotter, thats the least of my worries, its the keyboard that is too hot, I always leave my fans on 3000RPM or above, but the heat around the WERT453SDF section, is hot. I'm running at around 20% CPU usage right now. With my CPUs running at 50c.

I'm getting a replacement 15inch as this one arrived damaged, I'll do a side by side comparison of the two MBPs, see if there is any difference in them. There shouldn't be, I'm hoping there is, and the replacement is cooler.

 

[1096bimu]

Actually they are running hotter, and the proof is in the pudding. The i9s throttle more (not just on the MBP) and run hotter. I may not know how processors work, but I can see first hand they run hotter

It doesn't throttle more, in fact it throttles less because it runs at a higher frequency even when throttling.

Thats a pretty blank statement, which while holding some truth, is misleading in general, it may lead people to believe that magically i9@2.9 needs the same power as i7@2.2. While the difference is probably that i7 could do stable 4.6Ghz instead 4.8.

First - are you sure this is the same silicon? They have different cache organization. Of course Intel could degrade i9 to i7 by disabling partially cache and not only downclocking, but even then to take advantage of higher binned part you'd need to run with lower voltage for a given frequency. And so far I've seen no evidence that this is true in case of Apple and MBP. More, given how they messed up the simple throttling and fan curves I'm pretty darn sure they run it at stock voltages. And in this case the i9 will be hotter because of larger cache and higher voltage required to reach higher clocks. Aaaand - the differences between different binned parts would show up at the higher end of frequencies - since you can't run either higher than around 3.3GHz anyway fully loaded, binning is really a moot point.

The only person who could answer such question is somebody who works at Intel and knows their binning testing procedures and results for this particular chip. In case of MBP for all intents and purposes these are identical CPUs. Tell me what you want to prove and I will prepare a test suite that will show that i9 is cooler/hotter/faster/slower/louder/quieter, has better or worse battery life. Whatever you wish.

All I sad was that the i9 can run higher at the same power level, I never claimed how much higher can it run.

For processors with the same architecture and same core count, it's what Intel usually do. the i7 parts simply have some cache disabled, which also increases yields because it doesn't matter if some chips have faults in the cache, just disable them and sell it as an i7.

Now does the i9 run at a lower voltage at the same frequency as an i7? It would have to in order to maintain a higher sustained frequency. We just don't know whether Intel or Apple individually calibrate each chip, or they just have the same setting for all i9 chips.

 

[Thysanoptera]

For processors with the same architecture and same core count, it's what Intel usually do. the i7 parts simply have some cache disabled, which also increases yields because it doesn't matter if some chips have faults in the cache, just disable them and sell it as an i7.

Assuming faulty cache is the only binning criteria you would end up with identical statistical distribution of thermal capabilities between i9 and i7 populations. Thus - because i9 has larger volume of active cache it would heat up more at the same frequency.

Now does the i9 run at a lower voltage at the same frequency as an i7? It would have to in order to maintain a higher sustained frequency. We just don't know whether Intel or Apple individually calibrate each chip, or they just have the same setting for all i9 chips.

No it doesn't have to, this is not a linear function. And since you don't know, it would probably be wise to refrain from writing in all caps that i9 can run higher clocks at the same power level, don't you agree?

 

[1096bimu]

Assuming faulty cache is the only binning criteria you would end up with identical statistical distribution of thermal capabilities between i9 and i7 populations. Thus - because i9 has larger volume of active cache it would heat up more at the same frequency.

No it doesn't have to, this is not a linear function. And since you don't know, it would probably be wise to refrain from writing in all caps that i9 can run higher clocks at the same power level, don't you agree?

Why make that assumption though?
I do know higher end processors are more power efficient because I have been building and overclocking custom towers for so long. So yes I do know that.
Also cache takes so little power it's almost irrelevant even on mobile chips.

 

[buran-energia]

I don't think that an i7 running at 2.6 GHz will produce the same heat as an i7 running at 2.2 GHz. I remember overclocking my desktop, keeping a static voltage, upping the clock speed and looking at power consumption with watt-o-meter during a stress test. The higher the clock, the more power it consumed.

So no, I don't think a cpu running at 2.6 GHz will produce the same heat as the same CPU running at 2.2 GHz.

I'm curious what's the difference between 2.2 and 2.6. Both will throttle eventually, but I think the only difference is the one with the highest turbo boost will perform slightly faster during very short tasks.

 

[1096bimu]

I don't think that an i7 running at 2.6 GHz will produce the same heat as an i7 running at 2.2 GHz. I remember overclocking my desktop, keeping a static voltage, upping the clock speed and looking at power consumption with watt-o-meter during a stress test. The higher the clock, the more power it consumed.

So no, I don't think a cpu running at 2.6 GHz will produce the same heat as the same CPU running at 2.2 GHz.

I'm curious what's the difference between 2.2 and 2.6. Both will throttle eventually, but I think the only difference is the one with the highest turbo boost will perform slightly faster during very short tasks.

I never said they do, the difference won't be that significant.

When you hand-tune your desktop CPU, it's not difficult to make your CPU run at 2.6 while consuming power that would otherwise be used by a stock processor at 2.2, given you have a high quality chip. Power goes up linearly with frequency but exponentially with voltage. So all you gotta do is undervolt the CPU.

After you undervolt, the CPU uses less power on all frequencies, so if you are thermally or power limited (like in a MacBook), you can run at a higher frequency compared to a stock processor.

and i9 will be better quality chips compared to the i7 so they will do the same thing. It's just that factory settings will not be as tight as a hand tuned undervolt, and mobile chips are already top of the line so there won't be much room to be better either.

I suspect the i9 will be less than 5% faster than the i7 when thermal limited.

 

[MadCar]

I've had the i7 2.2 now for a month ish and I find it no hotter than my old 2012 i7 to be honest. It gets in some situations nearly the performance of an i9 and for a significant saving.

If you REALLY need the performance of the i9 then I personally would question if it would be better to have an iMac / Pro for that requirement. Its a laptop after all and the thermals are getting pushed to the edge and beyond.

I'm more than happy with the decision and how well its turned out.

 

[Thysanoptera]

Why make that assumption though?
I do know higher end processors are more power efficient because I have been building and overclocking custom towers for so long. So yes I do know that.
Also cache takes so little power it's almost irrelevant even on mobile chips.

L2 and L3 cache in Intel CPU's consumes 20 to 30% of the total package power. That's insignificant to you? I guess you didn't notice during your overclocking endeavors.. And besides, why do you overclock the high end CPU's? The whole idea is to get more performance for less money, my 68000 was beating my friends 68EC020, my 8700k beats 8086k. When you buy high end CPU you're buying stuff that's already at the limit.

Power goes up linearly with frequency but exponentially with voltage. So all you gotta do is undervolt the CPU.

In transistor theory, linear with freq, square with voltage. But this is only valid at the same temps. When the temp increases silicon properties take over and leakage becomes dominant, close to Tj the falloff is so big you need a logarithmic scale to notice it's not a vertical line.

Since your knowledge is limited to an anecdotal evidence from your own experience on small amount of samples, here are the conditions to make i9 consume less power than i7:
1. Intel cherry picks i9 from the same waffles as i7, taking into account more than just a faulty cache. That's an assumption, not a fact.
2. Apple fine tunes power delivery to i9 so that it takes less voltage than i7. For this you wrote yourself that you don't know this.
3. Intel bends the laws of physics making larger cache in i9 consume less power than i7.

So just please, stop spreading false information for which you have no evidence whatsoever and which contradicts the known propertierties of CPU's.

I'm done with this thread, going to play Skyrim on my totally non-gaming MBP.

 

[1096bimu]

L2 and L3 cache in Intel CPU's consumes 20 to 30% of the total package power. That's insignificant to you? I guess you didn't notice during your overclocking endeavors.. And besides, why do you overclock the high end CPU's? The whole idea is to get more performance for less money, my 68000 was beating my friends 68EC020, my 8700k beats 8086k. When you buy high end CPU you're buying stuff that's already at the limit.

20~30%? where did you pull that from? The A11 has 8MB of L2, no less than much higher powered Intel chips, which would be impossible if cache took that much power.

Because the whole idea isn't to get more performance for less money, it's to get more performance than money can buy. An overclocked 8086k will still beat your 8700k.

Since your knowledge is limited to an anecdotal evidence from your own experience on small amount of samples, here are the conditions to make i9 consume less power than i7:

The i9 is faster in all the tests and benchmarks.

 

[Matt_9141]

Title says it all really.

I'm in the market for a 15" MBP, and trying to choose between specs. Definitely want 32GB RAM, but unsure on which processor to pick. 85% of the computers work load will be basic stuff... emails, browsing, light audio editing, Football Manager , but the rest of the time I'll be doing some more intensive video editing (4K) and rendering.

I know, in theory, the i9 is the way to go and the more future proof, but how hot does it run? Particularly, does it run notably hotter even when doing simple tasks? I don't want to be sat with a heater on my knee while I'm doing fairly basic workflows.

Interested in your thoughts on this.

Bought an i7 2.6Ghz. This is the best i7 currently out there.

They tried to make me go toward the i9, but I refused, knowing that there was a problem.

My config: (No issues at all, not heating up too much, perfect)

2.6GHz i7
512GB SSD
560X Radeon Pro
32GB RAM. (Goodbye LPDDR3!!!)

This should last you 5-6 years or more!

-Matt

 

[stylinexpat]

The top section run really hot on mine when plugged in

 

[Matt_9141]

The top section run really hot on mine when plugged in
View attachment 789181

Bro thats like normal for MacBook pros. That how the heat is dissipated, cause the fans blow air out near that spot.

 

[IngerMan]

The top section run really hot on mine when plugged in
View attachment 789181

Hot is a relative term. I have iStat that records and displays temperature.

 

[iMacDragon]

The top section run really hot on mine when plugged in
View attachment 789181

The heatsink, and thus CPU/GPU are directly under that part of laptop.