I cancelled my 10-core order and bought the 8-core. And not because of the ground breaking 20% news. Amazon has the base 8-core for $50 off and I just don't need the 20% for $200. I want to be a power user so bad, but deep down I know I'm not.
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.
8-Core 14-Inch MacBook Pro Around 20% Slower Than 10-Core Models in Multi-Core Benchmark
- Thread starter MacRumors
- Start date
- Sort by reaction score
What counts as a real-world task? To the extent that the task is not INTRINSICALLY parallelizable, of course the CPU can't change that. But what the benchmark shows is that the SoC does not make things worse:
- large caches means the CPU's are not fighting each other for cache space
- large bandwidth means they aren't fighting each other for memory access
- not tested (but surely present) is that locking primitives are as fast as they can reasonably be.
The only thing really missing for an essentially perfect machine for running highly threaded code is the machinery for lock elision, and that will come in time -- it's surely on Apple's list of things to do, but likely lower priority than SVE and the rest of ARMv9.
The happiest people are people who know themselves honestly!
except the high perf cores are more useful and representative of hard workloads ,thus being more helpful when perf is needed,besides few software are truly optimised to share the work between many cores .granted 2E cores = 0.7P(since the 4E in m1 equal to 1.33P core )the difference is bigger than 31%
ig its pretty easy to calculate actually,it has double the P cores in the 8c,so 50% more powerful ,with the 10c being 100% more powerful
(ofc it doesnt work exactly that way)
What you say is true -- but also misleading.
The issue is not "normal everyday tasks" it is -- do you have that one particular task that you engage in, and care about enough, to be willing to pay to run it fast?
For some it might be video/photo editing, for some it might be compiling, for me it's running Mathematica (once Wolfram ship a properly M1'd version, not the pathetic rush port they have right now!)
A piece of hardware may be used only once a week -- but if it brings you joy when it is used, it may be worth the cost...
I’m more interested to find out how general efficiency pans out. With Logic Pro X beavering away, connected to a large external monitor, the last thing you want is fans whirring away to get rid of heat.
I think a lot of people suspected this all along. The newsworthy revelation here is that 8 turns out to be 80% of 10. It will take time for that to sink in.
Well, you're a cat. Even being able to do that basic math for your purchase is pretty impressive. Mine just scratch my favorite chair and bite my feet while I'm sleeping.
Not only is the 8-core M1 Pro 20% slower than the 10-core M1 Pro, but even more impressive, the 10-core M1 Pro is 25% faster than the 8-core M1 Pro 
It will be interesting to see the benchmark for the M1 Max 10-core/32 GPU for the 14" and 16". Also, the benchmark of the High Power mode on the 16" M1 Max 10-core/32 GPU. I am really interested if the High Power mode on the 16" M1 Max runs a higher clock rate then normal run mode.
Last edited:
This is smart. A lot of people are buying $3000+ computers to complain about things online. People like to think they’re future proofing, but it’s not a bad idea to buy something cheaper and plan to purchase something new in 3 years and sell the old one for a decent amount. That way you can stay up to date and enjoy buying something new sooner rather than later.
Binning is common practice in the industry CPU/RAM/etc.
When you buy an Intel 2.3GHz i7 it is the same chip as the 3.0GHz i7 but one of them can pass the tests for 3.0GHz and the other one doesn’t. They also disable components on chips when one is does not pass their tests but the rest do. This is common with memory chips.
In the case of the cores in the Apple SOCs, if one of the cores didn’t pass their tests it doesn’t mean that the rest of the chip is lower quality. It just means that there was a problem with one of the cores and it is disabled. The rest of the cores passed the tests on both versions of the chips.
Just remember Amazon does not have Apple's '"no questions asked" 14-day return policy. The only allowed reasons for returning an opened computer to Amazon is that it doesn't start or it arrived damaged. (This is different from Amazon's normal return policy). So you need to be sure you like it.
This is the reason I seldom buy newly-released Apple products from Amazon and other similar resellers.
Math
The first seemingly legitimate Geekbench 5 result for the base model 14-inch MacBook Pro with an 8-core M1 Pro chip has surfaced, and it reveals that the 8-core model is, as expected, ~20% slower than 10-core models in terms of multi-core performance. The 10-core model has 8 performance cores and 2 efficiency cores, while the 8-core model has 6 performance cores and 2 efficiency cores.
The benchmark result lists the 8-core 14-inch MacBook Pro with a multi-core score of 9,948, which is around 20% lower than the average multi-core score of around 12,700 for 14-inch MacBook Pro models configured with a 10-core M1 Pro or M1 Max chip. Keep in mind this is only a single result, so additional results are needed for certainty.
For single-core performance, the 8-core M1 Pro chip has approximately the same score as the standard M1 chip, the M1 Pro chip, and the M1 Max chip.
For multi-core performance, the 8-core M1 Pro chip is about 30% faster than the standard M1 chip, which also has 8-cores (4 performance, 4 efficiency).
Geekbench Scores
The base model 14-inch MacBook Pro with an 8-core M1 Pro chip is priced at $1,999 in the United States. Customers can upgrade this configuration to a 10-core M1 Pro chip with a 14-core GPU for an extra $200, raising the total price to $2,199.
- M1 (8-core) Single: 1742 Multi: 7582
- M1 Pro (8-core) Single: 1767 Multi: 9948
- M1 Max (10-core) Single: 1764 Multi: 12380
The new 14-inch and 16-inch MacBook Pro models became available to order on Monday and have started shipping to some customers ahead of a Tuesday, October 26 launch. In addition to M1 Pro and M1 Max chip options, the notebooks feature mini-LED displays with ProMotion for up to a 120Hz refresh rate, additional ports like an HDMI port and an SD card slot, MagSafe charging, longer battery life, and a notch housing an upgraded 1080p webcam.
Article Link: 8-Core 14-Inch MacBook Pro Around 20% Slower Than 10-Core Models in Multi-Core Benchmark
If you’re not a “power user,” I’ll bet the regular 13 inch M1 would be fine for you
Agreed. It has been enough for me but i went for the 14 inch base anyway because I want the 120hz mini led screen. And I want to be able to connect to two monitors when docked. I know that’s possible via display link but it’s not a great solution.
they'll play smoothly on a £250 Google Play 4a phone.
or even a £280 Chromebook that's 4 years old.
Disappointed that single-core performance is the same across M1, M1 Pro and M1 Max.
Improving single-core performance will have the most benefit for most tasks. And it is why Intel still focusses on improving it.
Sure, more and more software is being (re)written to take advantage of multi-core processing, but (1) not many tasks are, (2) even when they are, the code is imperfect (cache contention, locks, etc)
I think Apple is risking a backlash on the £3500 Max laptops when people find they perform like the M1 in single-core tasks.
Improving single-core performance will have the most benefit for most tasks. And it is why Intel still focusses on improving it.
Sure, more and more software is being (re)written to take advantage of multi-core processing, but (1) not many tasks are, (2) even when they are, the code is imperfect (cache contention, locks, etc)
I think Apple is risking a backlash on the £3500 Max laptops when people find they perform like the M1 in single-core tasks.
Well, the "best value" or "smartest purchase" Mac anyway
Sales don't seem to follow much logic.
https://i-blogs-es.cdn.ampproject.o...m1-max_cpu-performance_10182021/1024_2000.jpg.
A lot if compare watio usage. Mind they can scale it as much as they can as it looks like. Think in a 70w M1 all the core you can add. I9 couldnt beat that as i9 at 70w is what they sell right now with 8 cores and apple is selling just 30w socs
Until my workload demonstrates I can regularly utilize all that power, my better judgment says to save the $200-300 and go with the base option.
It doesn't work like that. Apple Silicon has faster cores (performance) and slower cores (efficiency).
Each performance core in the entire M1 lineup has an impact of around 1,540 in Geekbench.
Each efficiency core has an impact of around 350 in Geekbench.
The only thing that surprises me so far is that the score of the 10-core should be closer to 13,000.
The 10-core M1 Pro really should be almost 30% faster than the 8-core.
The article mentions the 8-core variant is 20% slower. If you rephrase the article, the 10-core variant is 25% faster than the 8-core.
Last edited: