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.
Is 2.6GHz 6‑core a better option?
- Thread starter scorpio187
- Start date
- Sort by reaction score
Umm, I see. I guess as long as the 2.2 serves it’s purpose for you then sure. I have a company issued 2.2 i7 and 2.6 i7 with the same 32GB RAM and 1TB drive but I always go back to my personal i9 as I get slightly better times running processes with RAW image file handling as well as the occasional video edit here and there.
Try out the Corona Benchmark software and let us know what you get.
One thing that has surprised me is how all over the board the MBP scores are. The i9/32/1TB that I had last, scored particularly poor in both Cinebench and Corona, nothing even in the ballpark of what Mr. Romos has posted. Scores where poor with both HS and Mojave. Had it achieved anything remotely close to what Mr. Ramos posted, I might still own it today.
I posted mine elsewhere, will try to find them later, but iirc they were in the upper 800’s Cinebench and around 2100k rays. Crappy for an i9
Just run that Corona benchmark on my 2.2. 2:59, 2708k rays/sec. I don't get how that benchmark works, higher rays/s but longer render time than @mr.anthonyramos
Me either. Prime 95 and Furmark and probably better CPU/GPU stress tests, but won't run on MacOS. Unigine's Valley also, but again won't run on MacOS
Not really a fair comparison as my primary notebook has a lot more thermal head room - i7 8750H/32GB/1070GTX - ambient temp 25C/26C
Pity Apple didn't put more focus on the cooling solution, we can see the 8th Gen hex cores can really pull hard with adequate cooling and reach their full potential.
I doubt the likes of the i9 can ever reach full potential in a very thin & light chassis such as the MBP's, however I do believe Apple could have done a better job, allowing the CPU's to take greater advantage of the Turbo Boost technology.
Q-6
Last edited:
That's as unfair as it gets
You could quite literally fit two MBP in this one. If anything I'm surprised MBP is only 10% slower.
That's as unfair as it gets
It's larger as I wanted the 17.3" display, yet still under 3Kg and 12 seconds faster in this benchmark alone, which is not bad for a 8750H versus a 8950HK
Q-6
It's larger as I wanted the 17.3" display, yet still under 3Kg and 12 seconds faster in this benchmark alone, which is not bad for a 8750H versus a 8950HK
Q-6
No its not bad at all, but i thought you were comparing it to the 2.2 MacBook Pro.
Irrelevant benchmark (for me anyway), since we all know MacBooks are running hot and on the edge, and have been since I can remember. It's not the question whether MBP allows the hexa-cores to stretch their legs, its the question of whether 2.2 can outperform the 2.9 in the macbook chassis.
TBH I think the performance will be very close. The 2.2GHz is likely the best option, the 2.9GHz probably does outperform the 2.2GHz, equally it's not going to be significant due to the limited cooling of the MBP chassis.
The i9 is a brutal chip and really needs a very solid cooling solution, with some OEM's now not utilising it for their thin & light notebooks which to me is a sensible move. It's really far better suited to workstation class or big gaming notebooks, in thin & light's mostly a marketing ploy.
Q-6
Last edited:
like the others said, the higher the clock the faster you hit throttling. sure the 2.9 is 10-15 percent faster but for how long? 10 mins before throttling kicks in? and with 1031 and i9 getting 1056 i wouldn't spend that much money on that small of a gain.
Well if 2.9 maintains 2.9 GHz base clock while throttled and 2.2 maintains 2.2 (not turboed to 2.9) then for longer periods 2.9 is better.
I use mostly logic - i wont be rendering for hours, meaning a little more headroom per core can save me from an overload when it might otherwise happen.
Oh well, cant return it anyway. I hope it gives me a little edge when doing prolonged work.. if not, I've learned my lesson i guess.
Syntethic benchmarks means nothing, doesn't reflect the real world usage.
You can watch as many YouTube videos as you want, the 2.9 outperforms the 2.2 and 2.6 and some even shows that runs cooler and with better battery life than the 2.2 and 2.6.
The i9 has the base clock speed at 2.9.
The 2.2 and 2.6 would need to turbo to get there, which means more heat and draining more battery.
At the end of the day, buy whatever you feel you need, they are all excellent choices and with very little difference between them.
You can watch as many YouTube videos as you want, the 2.9 outperforms the 2.2 and 2.6 and some even shows that runs cooler and with better battery life than the 2.2 and 2.6.
The i9 has the base clock speed at 2.9.
The 2.2 and 2.6 would need to turbo to get there, which means more heat and draining more battery.
At the end of the day, buy whatever you feel you need, they are all excellent choices and with very little difference between them.
I agree and as I mentioned before, real world tests are a huge difference when it comes to comparing these systems and many members as well as youtubers have proven the fact that if you want that 10-15% improvement especially after all the updates with MacOS, Premiere Pro etc, have shown that results have been getting better and better. FCX has always been a performer anyways on MacOS.
It's just so funny how we get people who try to disclaim real world testing and then push members to do benchmarking, then when benchmarking results come out, state that we need real world testing and the loop goes on.
It goes to say, the evidence is there, but the denial is stronger than the evidence. Sure, cooling system would have made a difference but then at the same time the results are there and if you want the best performance possible, get the i9. For my use, I am literally gaining days over the course of a year as I am saving at least an hour a work day because of the difference between my three MacBook Pro systems.
My problem with all of this is that everybody who gets their hands on both of them, just runs some benchmarks, or real life workloads and declares "i9 is crap" or "i9 is the second coming of i7-920". Without thinking for a moment why. How is it possible that the same silicon is faster / slower at the same frequency in an identical thermally limited environment. For a single core load it is obvious, i9 by about 10% due to difference in max allowed frequency and enough cooling headroom in such scenario. Anything more than one core and I respectfully dare to question any difference between i9 and i7.
If you look at what was posted even on this very board, you'll see big differences between exactly same machines, way more than any theoretical difference between 2.2 and 2.9. Apparently Apple quality control took a giant hit, I could buy the last machine @SDColorado had and create a youtube video showing i9 being 20% slower than base i7, using undisputed evidence, right? And his machine was still "within specs".
Second - fan profiles. I don't have i9, but when I run benchmarks on my 2.2 the fans barely move, or are laggy as hell. Since I've put custom fan profiles my 2.2 is within 5% (plus or minus) of every i9 benchmark/workflow I could reasonably replicate.
So until I hear from an actual Intel or even Apple engineer, or read a paper on a webpage that has address ending with .edu I'm going to live in denial that under sustained multicore load, i9 inside MBP chassis is no better than base 2.2.
If you look at what was posted even on this very board, you'll see big differences between exactly same machines, way more than any theoretical difference between 2.2 and 2.9. Apparently Apple quality control took a giant hit, I could buy the last machine @SDColorado had and create a youtube video showing i9 being 20% slower than base i7, using undisputed evidence, right? And his machine was still "within specs".
Second - fan profiles. I don't have i9, but when I run benchmarks on my 2.2 the fans barely move, or are laggy as hell. Since I've put custom fan profiles my 2.2 is within 5% (plus or minus) of every i9 benchmark/workflow I could reasonably replicate.
So until I hear from an actual Intel or even Apple engineer, or read a paper on a webpage that has address ending with .edu I'm going to live in denial that under sustained multicore load, i9 inside MBP chassis is no better than base 2.2.
Totally depends on your workflow & applications. CPU power consumption is simply not linear, as the processor ramps up to full Turbo the power demand ramps up exponentially with the i9 pulling over 100W with subsequent heating which needs to be dissipated by the cooling solution.
Even the base 8750H can pull close to 90W on a well optimised notebook, generally the higher spec CPU's only tend to offer around 5% performance boost, however the i9 8950HK can offer a significant increase, equally it's far better suited to larger notebooks with a substantial cooling solution which are note always practical to travel with.
Trade off with the likes of the MBP and XPS 15 is portability, saving say a pound in weight, equally such thin & light chassis will struggle to maintain higher Turbo frequencies for any period of time. Right now you simply cant have your cake and eat it until CPU power consumption reduces...
Q-6
I'm aware of all of that, unfortunately audio is a mixed bag of tricks - and CPU overloads happen way before CPU is utilized 100%, i can get to about half 75% utilization before dropouts happen, and that is true for any machine. And every DAW has its own quirks to top it off.
I have since i posted that read, that full-load is pretty much limited by the thermal design so all chips should be at roughly the same frequency.
Bursts/mixed operation is where higher clocked CPUs should have an edge. In Logic, when i load the new benchmark, clock is around 3,5-3,7 GHz. Do you have logic by any chance?
edit2:
afaik, they are the same design but different bins, arent they? so that shouldnt be much of a surprise.
Yep. Same dven from logicprohelp;-)
For the TNM test, I see roughly 3.45 to 3.65 when running 50 tracks. When I run the max number of tracks (72/73 for me) it stabilizes at about 3.2. The initial turbo burst is about 3.9. I've got a bunch of stuff also running (web dev related things).
Power draw is about 45-46W when running the max tracks.
I think that's the same test - I was using his new username on gearlsutz. Pretty sure its TTOZ, since how many prolific posters from Melbourne can there be;-)?ah! hi
Can you link me to the test? I only tried the TTOZ's new test