IF Apple make some adjustments in firmware/software yes, equally don't expect miracles.
Much is going to be depend on your usage & workflow if the CPU's inability to run at high boost frequencies is a negative impact or not.
Q-6
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2018 i9 MacBook Pro throttling discussion [merged]
- Thread starter M.Rizk
- Start date
-
- Tags
- macbook performance pro
- Sort by reaction score
IF Apple make some adjustments in firmware/software yes, equally don't expect miracles.
Much is going to be depend on your usage & workflow if the CPU's inability to run at high boost frequencies is a negative impact or not.
Q-6
No idea, honestly, but I would wait for the Artstechnica review, if you don't mind waiting. I expect that to be in-depth.
That's in line with what I did. Returned the 2.6/16/512 and ordered a 2.2/16/1 TB. Higher net cost, but 1 TB was worth it to me.
I'd be inclined to keep it. The MBP sn't going to perform badly, however the i9 is very far off from it's full potential.
The MBP will only really pale in the face of W10 notebooks that are physically larger and have far more robust cooling solution's.
Q-6
Yes, but you went up to a 1TB drive. I would be going 1TB to 2TB and the cost difference is MUCH higher.
The 8750H has been a fabulous performer for my needs. As much as I like the idea of the i9 on paper, it's only ever going to run optimally in a large 17" chassis with adequate cooling & power train.
Q-6
Ultimately, even if the i9 is not performing up to its potential as long as its not performing worse than the i7 (2018) model then I think I will keep it. If the i9 performs worse in things like After Effects renders (my main need) then I think I'd have to return it (as much as I don't want to).
yup waiting for something more conclusive on this as well. basically I want to get the most powerful 2018 macbook I can get hence I went with the i9. as long as it performs ahead of the i7's I'm golden.
Geekbench is a peak burst kinda of benchmark, with several quickly executed tasks run one right after another and not truly useful as an indicator of real world performance in most situations. It's just a number that makes it quick and easy to get a ballpark rating that can be tossed around as that kind of indicator and it really truly shouldn't be.
Cinebench, on the other hand, even though it is run fairly quickly (usually a minute or two on most modern hardware, the more cores the better in most situations) but it is able to provide evidence of system performance under sustained load conditions hence it showing so much fluctuation on these MBPs with the i9 - it's not really the processors that are failing everyone here, it's the sub-optimal cooling system that Apple put into these machines and probably didn't optimize the fan curves to account for such workloads.
Sure, Cinebench as stated is a fairly quick to run indicator of some real world performance for sustained workloads, but if you really want to test it, then the Final Cut Pro or Adobe Premiere Pro video editing/rendering ones will be the best possible way to see if the situation with any given MBP 2018 with an i9 is going to suffer from what appears to be a fairly consistent thermal throttling problem.
I must agree with you on this. I personally think that a firmware update could help the issue but definitely not solve it entirely. Overall though, yes, the i9 should perform slightly better...not to it's promised potential but still slightly better depending on your workflow.
Do you have access to an i7 2018 model? I need someone to benchmark an After Effects test so I can compare render times to an i9.
These initial reports don't look very good, which is too bad since I was highly anticipating working with the i9. While I believe Intel gets at least 50% of the blame here, there's no denying that the current design of the MBP is not helping at all, and that in this case, yes, thinness does indeed interfere with functionality. Yes, they could solve the problem by giving up on that front, but they won't do that at any cost, so... ¯\_(ツ)_/¯
I think it is not unrealistic to expect a firmware update to adjust the fan behavior. Certainly not an ideal solution, but one that would likely avoid the worst-case scenarios.
I'm waiting for the initial storm to calm down and most importantly to put the machine through its paces and see how it suits my workflow before taking a decision. The issue is there, still not going to rage-cancel an important purchase without seeing things with my own eyes.
I think it is not unrealistic to expect a firmware update to adjust the fan behavior. Certainly not an ideal solution, but one that would likely avoid the worst-case scenarios.
I'm waiting for the initial storm to calm down and most importantly to put the machine through its paces and see how it suits my workflow before taking a decision. The issue is there, still not going to rage-cancel an important purchase without seeing things with my own eyes.
I was considering one of these but now I'm holding off.
I wouldn't keep it. It will perform well as you stated but I wouldn't keep it because you are paying a premium for something that isn't delivering what you're paying for. And probably never will. It's just wasted money that could go toward, RAM or SSD storage or Applecare.
This is why we need more tests using real world applications I believe. Would be an interesting development to see what Cinebench will "upgrade/update" and an explanation from them would be nice as well.
Last edited by a moderator:
I can understand from a performance perspective why this might be true on some levels, but the fact that the benchmarks in use, optimized or not, are pushing the CPUs to max performance and because of that the cooling system is sub-optimal and the throttling is occurring at excessive levels killing performance rather drastically means whether or not the software in use is optimized or not has no bearing on the cooling system's total lack of an apparently ability to cool the i9 adequately enough to provide performance before that excessive throttling is taking place.
*phew*
tl;dr It's not as if Maxon is going to put out a new Cinebench executable with "i9 support" in the next few days that's going to a) make the benchmark more accurate on the i9 (which it would, of course)and b) suddenly make the operational temps while that benchmark is executing suddenly drop down to much lower levels at the same time. That's just not going to happen.
Simple short and sweet:
The MBP 2018 powered by the i9 has some rather significant thermal cooling problems and they're not going to be solved by just a software solution, not even a major change to the fan curves. The only thing that's going to "fix" this adequately at this point would be a somewhat permanent throttling of the potential speeds of the i9 itself and if that comes along boy that's really going to cause problems 'cause people are paying for that performance Apple is bragging so much about but so far can't seem to deliver based on the available testing results. And with each hour of each day that passes, more such testing results are just providing more concrete evidence that yes, Virginia, there is a problem.
Is it possible that apple will fix it somehow? For example by software update? (so processor/os behaves better).
I don't know if I should go with i9 15"... if apple does not fix the problem I will probably go with i7 13".
I don't know if I should go with i9 15"... if apple does not fix the problem I will probably go with i7 13".
This is what happens when you are not "letting the power drain roam freely" when the system is under full load ... I hope Apple will figure this out on their own.
Are those decent benchmark results for a i9 on macOS? Could someone with an i7 post/point me to some some mac cinebench r15 numbers? Are the 300 bucks more from the next lower CPU worth it given those i9 numbers?
I was scouring the interwebz for reference benchmarks, but only found Windows ones, which were obviously faster.
Are those decent benchmark results for a i9 on macOS? Could someone with an i7 post/point me to some some mac cinebench r15 numbers? Are the 300 bucks more from the next lower CPU worth it given those i9 numbers?
I was scouring the interwebz for reference benchmarks, but only found Windows ones, which were obviously faster.
Last edited:
Come and join us in the merged thread discussing the issue.
To be honest, it’s a little too early to tell right now if Apple will “fix” the problem. They might tweak a bit here and there but might not totally eliminate the issue.
But what I can tell you is this, performance honestly might depend on your workflow. Based on my my quick tests, I did get better results with my workflow (won’t go into detail, you can search my other posts where I detail what I did), other people not so lucky but in the long run I would still place my money on the i9 for quick bursts of speed, elongated load on the CPU could be another story though.
To be honest, it’s a little too early to tell right now if Apple will “fix” the problem. They might tweak a bit here and there but might not totally eliminate the issue.
But what I can tell you is this, performance honestly might depend on your workflow. Based on my my quick tests, I did get better results with my workflow (won’t go into detail, you can search my other posts where I detail what I did), other people not so lucky but in the long run I would still place my money on the i9 for quick bursts of speed, elongated load on the CPU could be another story though.
Here's my 2015 MBP 2.8 i7 under heavy load during a Premiere export. Nothing new really... fans are currently at max. It got down to about 75C range for most of the encode so this is about as good as this laptop gets in Premiere under load even when it's "cool."
Also included is what happened during a 90sec 4K project export in FCPX. CPU running much hotter but also around the base clock.
Also included is what happened during a 90sec 4K project export in FCPX. CPU running much hotter but also around the base clock.
Attachments
Last edited:
It seems that if you speed up the cooling fan, it might be better. Did anyone used a manual fan controller for 2018?
Nice. If the manual fan controller is available, does 2018 is better than 2017?