Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
As in running the fans at higher speeds? You mean it doesn't do that now?
I watched multiple videos from “famous,” youtubers concerning this issue, and not a single one mentioned fan performance, except on Aaron from Zollotech, who measured dB, a subjective measurement depending on microphone sensitivity and placement of said microphone, but did not include actual fan speeds.

Like I said, too many questions, too little answers, lack of imperical data, and as always... It will get sorted out.
 
  • Like
Reactions: dhruveonmars
Isn't the workload in that test 100% ?
yeah, meaning its really throttled hard.

thats why when he puts it in the freezer (and it runs at base speeds or better) it severly outperforms the 2017, but under normal circumstances its throttled so bad, the quadcore actually fares better, because it doesnt need to throttle so bad.
 
I watched multiple videos from “famous,” youtubers concerning this issue, and not a single one mentioned fan performance, except on Aaron from Zollotech, who measured dB, a subjective measurement depending on microphone sensitivity and placement of said microphone, but did not include actual fan speeds.

Like I said, too many questions, too little answers, lack of imperical data, and as always... It will get sorted out.

placement affects it, but SPL meters are calibrated, so if he measures everything from roughly the same difference that's fine.
 
In the past this has been a too-much-thermal paste issue. Will be interesting to see how it performs with properly distributed paste.
It could be. Some other members have pointed that out, but the general consensus is that members here ignore it could be a first-shipment issue, a fan issue, a software optimisation issue, or the fact that it’s actually a PowerBook G5.
 
  • Like
Reactions: simonmet
I am saying there are possibilities... but since I don't have one and cannot test the hardware to see if the fan has more room. They did do some slight throttling on the iMac Pro early on while the fans were still so quiet as to not be audible.... i.e. they had lots of room to power up.
There are only two options:
  1. The fan speed is software limited to something less than its capability.
  2. The fan speed is not software limited and can operate to its full capability.
If Apple can make a software update which permits a higher fan speed then one is the case. If it cannot then two is the case. Or are there additional options I'm missing?
 
It could be. Some other members have pointed that out, but the general consensus is that members here ignore it could be a first-shipment issue, a fan issue, a software optimisation issue, or the fact that it’s actually a PowerBook G5.

it cannot be software. he ran exact same software setup on 2017 quad core and it was faster (until he put the i9 in the freezer).

I, as a potential (up until now definitive) buyer, really hope he received a lemon and these actually perform as they're supposed to.
 
Like I said, too many questions, too little answers, lack of imperical data, and as always... It will get sorted out.
As have I. I said much, much earlier in this discuss we need to wait for more information.
[doublepost=1531942886][/doublepost]
yeah, meaning its really throttled hard.

thats why when he puts it in the freezer (and it runs at base speeds or better) it severly outperforms the 2017, but under normal circumstances its throttled so bad, the quadcore actually fares better, because it doesnt need to throttle so bad.
My "hypothesis" was based on something less than 100% (I used 80%)
 
  • Like
Reactions: kildraik
it cannot be software. he ran exact same software setup on 2017 quad core and it was faster (until he put the i9 in the freezer).

I, as a potential (up until now definitive) buyer, really hope he received a lemon and these actually perform as they're supposed to.
The profile for the computer is not necessarily (and often not) the same for each computer.
 
  • Like
Reactions: kildraik
As have I. I said much, much earlier in this discuss we need to wait for more information.
[doublepost=1531942886][/doublepost]
My "hypothesis" was based on something less than 100% (I used 80%)
you propose capping cpu usage? then single core performance suffers.
[doublepost=1531943157][/doublepost]
The profile for the computer is not necessarily (and often not) the same for each computer.

either way, apple is shipping an unfinished product, if its software side (although i really cant see how this can be true), its also bad.
not as bad obviously as poor design, but still bad.

out of the box, 2018 hexacore should outperform 2017 quad doing the same task.
 
A couple more data points:
- David Lee is planning on re-running his test that initially uncovered the problem, on an i7 "soon": https://twitter.com/Dave2Dtv/status/1019545699627646977
- John Poole (of @geekbench) wasn't able to experience throttling on an i7 using a compiler stress test: https://twitter.com/jfpoole/status/1019655007740653568, and plans to perform a similar test on an i9 soon, to see if he sees throttling for his test: https://twitter.com/jfpoole/status/1019662855145058304

Hopefully, more clarity to come soon on the relative performance of the i7 vs i9 version wrt throttling...
 

if they undervolt it (with a firmware update), they may get better performance. but then its false advertising...
 
you propose capping cpu usage? then single core performance suffers.
I'm not proposing anything, just merely wondering. With that said there are scenarios where software cannot take advantage of the full capability of a CPU. Multithreading is one such area. An example:

I have a Z620 which is capable of running 32 threads (dual 8-core w/hyperthreading). Handbrake is unable to utilize all 32 threads. On a comparable 24 thread (dual 6-core w/hyperthreading) system handbrake is able to utilize all 24 cores. Let's assume 24 threads is the limit for Handbrake. The 32 thread system is only utilized to 75% of its potential.

Again this is merely a matter of curiosity.
 
  • Like
Reactions: Ploki
no.
the original video has 2017 quadcore and it outperforms the i9 due to throttling.
[doublepost=1531940718][/doublepost]
They should've designed the cooling around it, or use a different chip (ryzen?), simple as that.

I hope the base i7 8th gen I got on the 2018 MBP does not have these throttling issues
 
It could be. Some other members have pointed that out, but the general consensus is that members here ignore it could be a first-shipment issue, a fan issue, a software optimisation issue, or the fact that it’s actually a PowerBook G5.

Again, if my 2016 2.7GHz can't maintain base clock at 100% and goes down to 1.6GHz in less than a minute, how could it be a first shipment/lemon/etc issue?

My 2016 does the same thing, fact. So what do you think happens when you replace stuff with much more power hungry internals but you don't change the cooling system at all?

I can't believe how any of this is new to people. Yes Macbook Pros throttle. There's videos about them throttling all over youtube - here's one.

Do your own testing, Intel Power Gadget is free.
 
My understanding is that the no quibble 14 day return is as per UK Consumer Rights Act 2015, and refers to unopened and unused products. In other words, unless the machine was not as described, or was faulty, Apple has no obligation (or promise) to provide you with a full refund.

See: https://www.which.co.uk/consumer-rights/regulation/consumer-rights-act

The 14 day return period is valid even if you open and use the item. However, it must remain as new and saleable condition when returned.
In fact you can keep it longer. You actually need to inform Apple of your intent to return within 14 days. You then have a further 14 days to actually return it to them....

I just returned a system I had purchased just before the update - thats how I know...
 
My understanding is that the no quibble 14 day return is as per UK Consumer Rights Act 2015, and refers to unopened and unused products. In other words, unless the machine was not as described, or was faulty, Apple has no obligation (or promise) to provide you with a full refund.

See: https://www.which.co.uk/consumer-rights/regulation/consumer-rights-act

No, Apple's return policy goes beyond that. Why not just read Apple's policy instead of wasting your time quoting Which? I mean that in the nicest way.

Obviously just don't damage the MBP while you're testing it out and there's no worries returning it for a full refund if you decide to.
 
Hey, my Lenovo runs renders for hours no sweat.
But then you have other issues. My friend (Window Dev) is constantly fighting with his Windows and Microsoft. In fact, he got so fed up Windows update that for work (not gaming) he runs his Windows with everything installed on VMWare on a bare-bones Windows so that every time Microsoft bricks his computer he can just roll back and continue. Then the last update Microsoft caught on (1803) and made it so that he was no longer able to log in. Then, of course, there is just plain the inattention to detail and they are constantly using different standards (different font sizes in sidebar, slightly different shade to certain standard elements, changing standards but never fixing to make sure you change things to that new standard in a consistent way, hiding things in plain odd places (and sometimes moving them around between minor updates)... all of which might seem minor but once someone points it out to you your (or my) OCD side takes over and it becomes very irritating. I have worked in software development long enough that you sometimes are so focused on making the hard bits work perfectly, that you sometimes forget to go back and focus on the small bits... that combined with compressed schedules because someone in marketing thinks that the date should be some random date for no reason without consulting development and give assurances to customers that it will be out on that date (which often will freeze the date because now they are making dependencies on it).

If it turns out to be a teething problem, I will be happy. Luckily I am sort of waiting on everything for my current refresh... though I suspect by the end of August I will have what my next Linux build will be (I have not decided if it will be a Threadripper or the 9900K; the motherboard; or how much graphics performance I will install)... then I will just be waiting on hopefully learning more on the next Mac Pro this Fall or early Winter so I can make a decision on my next Mac.
[doublepost=1531945693][/doublepost]
Test it. Report back.
If you do test it, and it throttles even slightly... try installing a fan controller (it might not be customized exactly - but usually it will work even if it just throws the fans into permanent high mode until you reboot it :eek:
 
Again, if my 2016 2.7GHz can't maintain base clock at 100% and goes down to 1.6GHz in less than a minute, how could it be a first shipment/lemon/etc issue?

My 2016 does the same thing, fact. So what do you think happens when you replace stuff with much more power hungry internals but you don't change the cooling system at all?

I can't believe how any of this is new to people. Yes Macbook Pros throttle. There's videos about them throttling all over youtube - here's one.

Do your own testing, Intel Power Gadget is free.

Denial. They're the same people who SWORE all those other folks who said Apple was throttling their iPhones were wrong even after seeing reports with benchmarks. <nod>

As an aside, You or someone You trust should repaste Your unit. Should help some.
 
  • Like
Reactions: ssgbryan



Apple's new 15-inch MacBook Pro can be upgraded to include a 6-core 2.9GHz Intel Core i9 processor that has demonstrated impressive performance, but one YouTuber is warning customers away from purchasing it with claims that the MacBook Pro chassis can't provide sufficient cooling for it to run at full speed.

Dave Lee this afternoon shared a new video on the Core i9 MacBook Pro he purchased, and according to his testing, the new machine is unable to maintain even its base clock speed after just a short time doing processor intensive work like video editing.


"This CPU is an unlocked, overclockable chip but all of that CPU potential is wasted inside this chassis -- or more so the thermal solution that's inside here," says Lee.

He goes on to share some Premiere Pro render times that suggest the new 2018 MacBook Pro with Core i9 chip underperforms compared to a 2017 model with a Core i7 chip. It took 39 minutes for the 2018 MacBook Pro to render a video that the older model was able to render in 35 minutes. Premiere Pro is not well-optimized for macOS, but the difference between the two MacBook Pro models is notable.

Lee ran the same test again with the 2018 MacBook Pro in the freezer, and in cooler temperatures, the i9 chip was able to offer outstanding performance, cutting that render time down to 27 minutes and beating out the 2017 MacBook Pro.

As Lee points out, thermal throttling is in no way unusual and it's seen in all manner of laptops and mobile devices from a range of manufacturers, but he says that "this degree" of thermal throttling is "unacceptable."It's not clear if there's something wrong with the MacBook Pro with Core i9 chip that Lee received, because this kind of throttling is likely something Apple would have tested for and not something that other users have reported at this point.

Because this is just one data point, it's not enough information to reach a conclusion about the i9 chip available for the 15-inch MacBook Pro, but additional testing will certainly follow to shed more light on Lee's video. We have asked Apple for comment on Lee's findings, and will update this post if we hear back.

Article Link: YouTuber Claims 15-Inch MacBook Pro With Upgraded Core i9 Chip is Severely Throttled Due to Thermal Issues

When you do your test, I suggest you quietly go to an Apple Store or online and purchase an i9 MBP on your own instead of receiving a review unit from Apple. That way, you are not getting a unit that Apple doctored up just to past the throttling portion of the review.
 
I'm not proposing anything, just merely wondering. With that said there are scenarios where software cannot take advantage of the full capability of a CPU. Multithreading is one such area. An example:

I have a Z620 which is capable of running 32 threads (dual 8-core w/hyperthreading). Handbrake is unable to utilize all 32 threads. On a comparable 24 thread (dual 6-core w/hyperthreading) system handbrake is able to utilize all 24 cores. Let's assume 24 threads is the limit for Handbrake. The 32 thread system is only utilized to 75% of its potential.

Again this is merely a matter of curiosity.
That is correct, if you have more than 8 cores or threads (can't remember which one it is) it will not perform at maximum... you literally have to basically peg it against a defined set of 8 ... and if you have more power just run multiple copies. I think it is actually in the lib (x264/x265) that Handbrake relies on (and other things). You cannot just run multiple copies without doing that or they will just step on each other and the performance for each copy will suffer.
 
Again, if my 2016 2.7GHz can't maintain base clock at 100% and goes down to 1.6GHz in less than a minute, how could it be a first shipment/lemon/etc issue?

My 2016 does the same thing, fact. So what do you think happens when you replace stuff with much more power hungry internals but you don't change the cooling system at all?

I can't believe how any of this is new to people. Yes Macbook Pros throttle. There's videos about them throttling all over youtube - here's one.

Do your own testing, Intel Power Gadget is free.
So whereas other companies get the accountants breathing down the necks of the engineers, forcing them to cheapen up the specs of their machines, Apple is getting the fashionistas doing the same thing, saying something like, "You have to make it paper thin, and as long as the user gets the impression of something happening fast for a second or two, they won't notice that the rest of the task is no faster". Smoke and mirrors, and of course it fails once subjected to an objective analysis.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.