exchange it to a one of the lower i7 CPU's and you should have a more stable run, still experience some throttling but not as much as the i9
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
Ah! Yes I think you’re right. My 300GB worth of files and apps and whatnot on my old mac would be good on the 512GB SSD.
Thanks
Honestly if Apple is smart this shouldn't actually be that hard to "get on top of."
Step 1: Release a PR statement saying that "A small number of 2018 15" MBP's were shipped with a prerelease/misconfigured thermal management firmware"
Step 2: Update the firmware to manage thermals more better (slower clock-speed ramp up on heavy multicore workloads / more aggressive default fan curve / higher temperature threshold / ideally maintaining at least base clocks under heavy load). At this point it's already been "taken care of," even if the machines now sound like hair dryers under load.
Step 3: If possible quietly work on incrementally optimizing thermals to better manage/mitigate heat/power draw, while reworking the whole chassis for 2019.
And here is one more comparison:
https://9to5mac.com/2018/07/18/how-macbook-pro-throttles-with-final-cut-pro-x/
https://9to5mac.com/2018/07/18/how-macbook-pro-throttles-with-final-cut-pro-x/
Soo, https://twitter.com/jfpoole/status/1019740723002961925
"If the CPU is at 800MHz, the CPU isn’t throttling, the CPU is idle. The test isn’t using the CPU but rather the on-chip hardware encoder."
Interesting...
You are being completely disingenuous with your posts in trying to trying to associate this workflow with the issues discussed in this thread. I'm not denying your result, but it's obvious to most people here that your claimed 3x improvement over the 2017 model cannot be a processor-limited test. There's no processor benchmark on this planet that puts the new 6-core chips that far ahead and you've not posted any traces or evidence showing the sustained maximum CPU usage that's likely to trigger the throttling issues reported. I'm sure software and/or other non-hardware issues are contributing to such dramatic differences.
I get that that you're trying to demonstrate a typical-usage "whole system test", but tests like that are incredibly variable and that's why benchmarks exist. I don't think you should be handing out advice to someone who may be interested in or have processor-dependent workflows.
Last edited:
Right, interesting that CPU from MacBook Pro 2017 is not in idle during export!
Ah, just like benchmarks that show that the i9 is a few thousand points higher than the i7 but at the end of the day, do we trust these benchmarks? Just like benchmarks that we are all getting varying results from? Sure let’s have some more of those.
I simply presented and I already stated that I am far from technical with CPU benchmarking, is a different angle of things and my personal usage in response to those who are making absolute claims about the 2017 i7 being more powerful than the 2018 15 inch models. Did I claim that I am a tech guru and you should all bow down to my opinions? No, these are forums, whether you believe me or not or even those YouTube videos for that matter, everyone should take everything with a grain of salt.
So, I’m not claiming anything else than what I have tested, now maybe your time is better spent informing absolutists who make claims when they themselves don’t even have the new MacBook Pros rather than someone who went through the effort of doing a simple test based on my workflow to show “hey, it’s okay. Throttling yes, but still worked for me”.
Thanks and bye.
[doublepost=1531986310][/doublepost]
Does this mean we go back to certain software optimization possibilities?
Will you people cut it out with the Emoji/Watchbands nonsense? It makes absolutely no sense.
Yeah I can carry a pound or a kilogram, I can carry 10kg extra, doesn't mean I want to. I really don't get your argument. Just because I can carry more, doesn't mean it's comfortable to do so. I could carry a coat in the middle of a hot summer day, but I wouldn't feel nice doing so.
You're ridiculously avoiding the point. That has nothing to do with anything. If my clothes weigh, I don't know, half a kg - does that change the fact that I don't want 3kg hanging on my shoulder?
There are so many factors here, like the combined weight, not individual weight of each item, then the distribution of weight, etc. - but at the end of the day, carry 3kg around your shoulder all day, then carry 2kg around your shoulder the whole day and tell me which one was a more pleasant experience?
So, your argument is: because hikers and backpackers carry 10 kilos on their back while hiking, it's not a problem for me to do so while going around the town? Hey, why not make a 10kg laptop then? Imagine how many hardware you could put in there! Because a normal adult of average strenght can carry 10 kg.
You are seriously confusing capability with comfort.
You're really missing the point completely and this discussion is pointless. You're just wrong. I rarely say it to people, but in this case - you're so wrong, it's not even subjective.
[doublepost=1531987149][/doublepost]
Did you miss breakfast
Q-6
Having a different view is fine. Overusing the same jokes from the Internet echo chamber - that's just spam.
You're digging yourself into a whole, posting random meaningless numbers without explaining them properly. Taking current Geekbench 4 multi-core figures, the best 2018 model (2.9 GHz i9) is 57% faster than the worst 2017 model (2.8 GHz i7). That's still a significant improvement thanks to extra cores and no-one is saying otherwise, but Geekbench is not a terribly long or sustained test that can demonstrate the effect of throttling that's been observed, and that's the real issue.
What's clear however, is that 57% in that test is still nothing like 300% nor can it ever be.
Last edited:
I saw a post over at the [H]ardForum earlier with someone that had the idea of the Blackmagic eGPU being introduced at the same time with the new MBP lineup because Apple knew ahead of time, by a number of months, about the throttling with the i9 and knowing that people would be expecting to use the new MBP hardware for intensive video rendering jobs. The idea the person presented was Apple kinda packaged the eGPU with the MBP in terms of the marketing as a knee jerk reaction to handling the throttling: obviously if one uses the Blackmagic eGPU for doing the rendering work that would leave the internal CPU/GPU in the MBP running much cooler.
It's not as crazy as one might think, even if it does smack of a conspiracy theory like nature. I wouldn't put this past Apple for one second as it just can't be possible the engineers at Apple weren't aware of the throttling but maybe because they decided on a date to release the new lineup they had to come up with that last minute knee jerk solution of the eGPU.
I mean really, did anyone ever expect such a thing from Apple? I know they don't necessarily make the Blackmagic eGPU but even so, the timing of it being released in the same marketing materials with the new MBP lineup... that's not a simple coincidence, people.
Anyway, here's the post, I found it pretty interesting:
https://hardforum.com/threads/new-m...macbook-pro-i7.1964444/page-3#post-1043735592
It's not as crazy as one might think, even if it does smack of a conspiracy theory like nature. I wouldn't put this past Apple for one second as it just can't be possible the engineers at Apple weren't aware of the throttling but maybe because they decided on a date to release the new lineup they had to come up with that last minute knee jerk solution of the eGPU.
I mean really, did anyone ever expect such a thing from Apple? I know they don't necessarily make the Blackmagic eGPU but even so, the timing of it being released in the same marketing materials with the new MBP lineup... that's not a simple coincidence, people.
Anyway, here's the post, I found it pretty interesting:
https://hardforum.com/threads/new-m...macbook-pro-i7.1964444/page-3#post-1043735592
Thanks for sharing. Maybe another thing to consider is that Apple actually is aware that their current line has reached a cap with the current processor creating a balance between heat and balance and they KNOW a redesign will not be available until 2020 hence they bring out the eGPU for those looking to boost their Macs over the next two years?
Does lower clocked CPUs really impact battery life that much? I am doing some research, and it seems that a ton of more casual consumers prefer lower clocked CPUs and stated significant improve in battery life compared to the higher models.
Of course it works that way, just as - oh no, dreaded car analogy incoming - if you drive a car at 55 MPH for an hour it'll use less gas than if you drove it at 85 MPH for the same period of time. More power used aka more speed/processing power = more fuel aka battery life consumed to get that level of performance.
It's one reason why the Intel "U" processors are in most "Ultrabooks" because they tend to sip power when idle and even when maxed out they still don't consume a serious amount of power. The reason most Ultrabooks can only offer a given level of battery life isn't because of their processors, it's because in the effort to keep making them thinner and thinner, the batteries are smaller and smaller. If you took a modern 15W TDP Intel "U" processor and coupled it with a traditional 90WH battery cell you'd be looking at 15+ hours of battery life, or even more.
The quest for thinness in all this computing technology is actually not a good thing in terms of performance. You can have good to great performance, or you can have long battery life, but if you're trying to get both in a super thin form factor, one of those aspects must be sacrificed. Not even Apple is going to make that kind of magic happen.
I see, thanks, looks like I will be switching to the 2.2 i7 model...cooler, more battery life, and longevity is much better than a bit of top performance.
why did apple even bother offering the i9 as an option without having figured out a cooling solution to get the most out of it? I mean it's no news that the i9 is running hot, and that other Laptop manufactures have the same issues?
Okay. You have your opinion, I have mine. Great! That’s what forums are for. Lol.
One interesting point to add (and not sure if anyone has made this before), has anyone tried measuring the i9 with Turbo Boost switched off?
Background: I ultimately got annoyed by the fans on my MBPs always spinning up on even simple tasks based on the CPU's aggressive behaviour with ramping up Clock Speed on tasks that sometimes barely have any benefit from it.
I came across a bunch of tools that allow you to disable/enable Turbo Boost manually:
http://tbswitcher.rugarciap.com
I had done some simple benchmarks with Handbrake converting video at the time and the difference between TB on and off was roughly 5-10fps when converting 1080p.
I know this isn't a solution to a much bigger underlying issue, but I'm curious as to see how this affects the overall performance with aggressive throttling in place.
Background: I ultimately got annoyed by the fans on my MBPs always spinning up on even simple tasks based on the CPU's aggressive behaviour with ramping up Clock Speed on tasks that sometimes barely have any benefit from it.
I came across a bunch of tools that allow you to disable/enable Turbo Boost manually:
http://tbswitcher.rugarciap.com
I had done some simple benchmarks with Handbrake converting video at the time and the difference between TB on and off was roughly 5-10fps when converting 1080p.
I know this isn't a solution to a much bigger underlying issue, but I'm curious as to see how this affects the overall performance with aggressive throttling in place.
Just wanted to share my Mac's situation to compare the latest ones. Mine is late 2016 15" with lowest i7-6700HQ 2.6GHz, 512gb ssd and RP 460.
It runs flawlessly as I expected from Apple for years. I ran multiple cinebench tests consecutively and it always kept its 4-core boost clock speed (3.1 GHz).
Well.. the thing I mentioned before is Intel lies about tdp with recent cpus. My Mac's i7-6700HQ is rated at 45w tdp and it never ever exceed this value during testing and generally stayed at 41-42w.
Also, running the fans always at max rpm didn't improve the score but kept cpu a bit cooler. (with Auto-Fan it reaches easily up to 95C degree hitting sometimes to 98C, with max fan rpm, it didn't pass 89-90C.
Btw, it gets scores between 665-670 always, so it is highly consistent, no fluctuations.
It runs flawlessly as I expected from Apple for years. I ran multiple cinebench tests consecutively and it always kept its 4-core boost clock speed (3.1 GHz).
Well.. the thing I mentioned before is Intel lies about tdp with recent cpus. My Mac's i7-6700HQ is rated at 45w tdp and it never ever exceed this value during testing and generally stayed at 41-42w.
Also, running the fans always at max rpm didn't improve the score but kept cpu a bit cooler. (with Auto-Fan it reaches easily up to 95C degree hitting sometimes to 98C, with max fan rpm, it didn't pass 89-90C.
Btw, it gets scores between 665-670 always, so it is highly consistent, no fluctuations.