Its not A youtube wannabe, its MANY youtube wannabes as you call them showing real world numbers and benchmarks proving this. I understand you and many others bit the bullet and bought the i9 blindly without seeing if it would actually make difference, but that doesn't mean its worth it to others. Even with those numbers your showing, I still wouldnt spend that money for a 2 minute improvement. As I mentioned earlier, no one can tell me those 2 minutes, even 10 extra minutes in a whole workday is worth it.
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.
Max Yuryev tests the MBP 15" CPUs: 2.2 vs 2.6 vs 2.9
- Thread starter sennomulo
- Start date
- Sort by reaction score
Its not A youtube wannabe, its MANY youtube wannabes as you call them showing real world numbers and benchmarks proving this. I understand you and many others bit the bullet and bought the i9 blindly without seeing if it would actually make difference, but that doesn't mean its worth it to others. Even with those numbers your showing, I still wouldnt spend that money for a 2 minute improvement. As I mentioned earlier, no one can tell me those 2 minutes, even 10 extra minutes in a whole workday is worth it.
The vast majority of real world tasks are quick burst loads on the processor. Thin and light laptops are designed with those kind of workloads in mind (design, coding, image editing, office apps). If laptops were designed with rendering or gaming in mind you just can’t make anyone happy. They will just throw bigger and bigger data at it no matter how efficient you try to make the machine.
I have throughly tested and timed my batching workloads. I have shaved 30% off my batching time with Coffeelake and that can mean anywhere around 20 minutes a day....or 52 hours a year.
Anything that can be automated and made faster while consuming the same or less amount of energy should be applauded and embraced, otherwise your carbon footprint will remain higher and your productivity lower.
Nobody should be so stupid to say ‘Gee I really want to argue against faster more efficient technology because some YouTuber says please click my link’’
Do you have the i9? What are you comparing your 30% time savings against? Would you still be seeing a 30% time savings if you would have gone for the cheaper i7 model?
This is why I'm really hoping that we'll get multiple sources benchmarking these kind of things. If it leads to 10-20% differences in these situations between the 3 variants then sure the i9 would be worth it for me. These tests just don't show that, however they do show a lot of variance exists. For instance both the Youtube video and Bearfeats.com show lower Cinebench scores than I have gotten on my i9. I get it to repeatedly go up above 1100 on the first round and stay above 1000 after 5 consecutive runs, now that the ambient temp is lower.
I just hope that people are willing to benchmark these machines in different settings documenting how they have gotten the results, so others can decide whether those situations match their use-cases.
These tests do show that you won't get the full benefit of the i9 and that means I can't recommend the i9. I guess it's similar to picking the amount of ram; if you need 32gb you know that you need it already, else go for 16gb. If you need the i9 you should already know that you need those last couple of percentage points, else go for the i7.
My machine is going back due to other issues (I've talked enough about them in other topics), not due to it's performance, but when I'm buying one in 1 or 2 months from now I'll hope that these tests are out there and I'll make the decision again. If all issues are ironed out by then, they all seem fine products and if things stay as they are right now, I'll pick the 2.6 i7. The $300 I'd save I rather put towards a eGPU solution when performance differences are this small.
Last edited:
and here we are, a bunch of professionals using our i9 MBPs to be professional at stuff.
I'm comparing the 2.6 i7 against the equivalent 2016 and 2017 models. The i9 would give me even more boost at resizing ops (see Photo Zoom result attached previously)
With the cheaper i7 2.2 model maybe 10-15% in real world terms. The clock speed is very low on this model and wouldn't be a leap enough. Also, OpenCL is used for batch processing (in image resizing for example) and that 560X helps here.
If someone is upgrading from 2017 13" models to 2018 i5 models then the time savings would be greater, relatively speaking.
@Macshroomer thanks mate! that's exactly what I was looking for. Real experience from real users. your test makes a lot of sense and it does prove the real capability of this MBP in every day tasks. I would use the i9 in the exact same way you are using it (extreme photo editing, with large files from Leica Q and Leica M) and I know I will be happy with it. Thanks a lot man, your words and tests are what I value the most ... everyone seems talk about air and smoke and only a few really speak about concrete work.
His test is missing data that's the point of this thread though - 2018 i7 vs i9
i9 being noticeably faster than 2017 i7 is something we would have all assumed.
This thread seems to be about if the i9 is worth getting if the i7's were in the 4min ballpark also, as an example.
Barefeats saw the i9 performing up to 15% faster depending on the task. So for some usages it doesn’t seem to be such a bad deal?
http://barefeats.com/macbook_pro_2018_i9_vs_i7.html
Guess it totally depends on what you’re planning to do with the machine.
http://barefeats.com/macbook_pro_2018_i9_vs_i7.html
Guess it totally depends on what you’re planning to do with the machine.
Yes this as well BTW; my i9 constantly runs at 50-55 degrees in idle. I suspect that's mainly do to a reluctance to spin the fans to the point that they become audible, but I find it quite hot to use on my lap. I use Macs Fan Control to do speed the fans up now and adjusted the curve to it spins them at 3200 for cpu and 3000 for GPU and keeps them both at 46-48 degrees without too much noise.
Last edited:
Of course you won't get 20%. It will be somewhere around 5% from the i7 to i9. This has always been the case between mid-tier and hi-tier CPUs... you pay a major premium for only minor return. The base CPU was always the best bang for your buck.
I shouldn't have said 20% sure, but I said 10 to 20%. Let's leave out the 20%, I still don't think a expecting 10% performance difference is unreasonable from a machine that is advertised as going from 2.6 to 2.9 Base speed and 4.3 to 4.8 Turbo boost and requires additional payment of $300 for that difference. Mind you, you pay the $300 just for the CPU, not an additional power supply, cooling or other case adjustments, nothing else is changed, just the CPU and its projected performance.
If you then end up getting zero to 5% difference, how could you advise people the i9 instead of the i7 2.6? With the test results we have now, I honestly can't. Like I said, I hope that these machine's will be tested and benchmarked for all types of different workflows and that the i9 consistently starts to show a 5% performance difference, as then you could argue that for some people that makes sense. Even then you're paying 11% more for the total machine, so whether you value those 5% for an 11% price increase is a good question to ask yourself.
Advising people to spend an additional $300 because it saves you 0-5% that's now the result from most tests is not something that you should be doing I think, even if that means you yourself probably didn't make the best choice by buying the i9. A lot of people come here for honest advise and I just don't see with the current results how anyone should be advised to buy the i9.
Last edited:
The 15% was versus the i7 2.2ghz base model. It was 4.5% versus the i7 2.2ghz model.
That’s probably a difference between 560x and 555x if you are talking about the photo zoom test.
Didn't even think about that, does that use the GPU instead of the CPU?
Barefeats test's basically have the same flaw as the YouTube video that started this thread. I really hope to see benchmarks where everything is kept equal apart from the CPU. Variations in GPU and ram (there are 16gb and 32gb versions used in Barefeats tests as well) allow for other performance differentiators than the CPU.
A 17% price increase is a few hundred dollars.
A 15-30% productively increase can be many thousands of dollars a year extra profit. Or more time to relax.
There's a 30% difference between the 2018 models?
Or are you bringing 2017 models into a thread about 2018 model differences?
Last edited:
What makes you say that? I think he'd earlier referenced comparing 2018s to 2016/17 models, so it's not entirely clear what comparison he's using for that 30% difference statement.