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.
Native Geekbench 5 scores are impressive for first gen beating out the MacBook Pro 13's scores
- Thread starter pshufd
- Start date
- Sort by reaction score
Was this "native" Geekbench running as a new MacOS app or an iOS app? Does that make a difference?
this persons videos aren’t for me
in case you feel that way (and this isn’t his original reporting anyway)
the results were 1098 single and 4555 multicore
I wonder how much faster the per core will be on the first arm macs? I know everyone says this is a “2 year old” chip, but as implemented for the iPad Pro its 3 months old and presumably represented their best implementation of a slate device (which is closer to the way a computer processor would be implemented). If it were trivial to just use an A13 but add cores wouldn’t they have done that?
Going to 5nm,
For the MacBook i’ll speculate 4 high power cores and 4 low and we get 1225 single core and 5000 multi which would be close to a ryzen 3 desktop.
for MacBook Pro I’ll speculate 8 high and 4 low and 1200 single and 8000 multi which would basically be like ryzen 7 desktop
Get me that MacBook for $1199 with likely 8gb of ram and it’s the perfect portable for iOS like entertainment using my AirPod pros, browser based apps, iMessage, zoom calls with my AirPods and to remote in to my windows office machine. I’m sure there are a lot of people like me who use this as a second or third machine who are going to snatch these up to complement their iPhones.
in case you feel that way (and this isn’t his original reporting anyway)
the results were 1098 single and 4555 multicore
I wonder how much faster the per core will be on the first arm macs? I know everyone says this is a “2 year old” chip, but as implemented for the iPad Pro its 3 months old and presumably represented their best implementation of a slate device (which is closer to the way a computer processor would be implemented). If it were trivial to just use an A13 but add cores wouldn’t they have done that?
Going to 5nm,
For the MacBook i’ll speculate 4 high power cores and 4 low and we get 1225 single core and 5000 multi which would be close to a ryzen 3 desktop.
for MacBook Pro I’ll speculate 8 high and 4 low and 1200 single and 8000 multi which would basically be like ryzen 7 desktop
Get me that MacBook for $1199 with likely 8gb of ram and it’s the perfect portable for iOS like entertainment using my AirPod pros, browser based apps, iMessage, zoom calls with my AirPods and to remote in to my windows office machine. I’m sure there are a lot of people like me who use this as a second or third machine who are going to snatch these up to complement their iPhones.
It's the first one I ran into. I have since run into another one. If you have one you like, post a link.
I'm just happy that it's in the ballpark. That it's slightly better is a plus. What I am hoping for is better thermals and lower power consumption for an equivalent amount of compute processor.
I'd still want at least 32 GB. I'd be fine with a Mini if this thing uses half the power of the i7.
It is not trivial just to add A13 cores. Since they knew the much better A14 based cores are right around the corner, the effort of making an A13X was not worth the effort.
BTW, your estimates are laughably low. A13 in a phone already does 1350 in single core.
An A14 based core in something with much more thermal headroom would score much higher than the A13.
[automerge]1595824146[/automerge]
Title is deceptive. They benchmarked the DTK which is not first gen. Its a developer transition kit using an ipad pro processor and will not be sold to general consumers. Apple already stated first gen macs will be a lot faster.
Last edited:
My inferences are based on what has been demonstrated, as you say yourself you can't just add cores. Single core performance needs to decrease but it's a worthwhile trade off. They also have to squeeze a GPU in there that matches or exceeds a discrete desktop GPU. You lose battery life and temperature benefits if you have 8 high power cores running at 3ghz. Whatever ~30% efficiency benefit is available for arm ISA vs x86-64 ISA my ryzen 9 3900x gets pretty warm in a desktop case using 7nm TSMC at ~4 ghz.
I'd wait until i saw something myself before i determined as reasonable a guess as my own to be "laughably" wrong, for a number of reasons.
Last edited:
No it doesn't. As demonstrated by countless Ryzen and Intel products. Their highest performance chips are the 8 or higher core models. There is a thing called turbo boost. You should look it up.
My estimate is the absolute lowest for a macbook pro replacement for single core is 1500 which is a modest 10% increase over the iphone 11. A reasonable estimate is 1600.
When running multicore, there will be some compromises compared to the boost speed when running single core - but given how efficient the Apple A series is - I doubt that they would lose too much speed - especially if they are now actively cooled.
In any case - there won't be too long until we find out. In September at the iphone 12 announcement Apple will detail how much faster the A14 is compared to the A13. I wouldn't be surprised if it is >20% especially since they are moving to a higher performance TSMC 5nm node. (1350x1.2 ~1620 for single core).
Last edited:
what tdp are you using in your Apple $ilicon MacBook Pro and how many cores in the package?
Firstly you need to realise that intel's TDP number are more or less meaningless for the chip under load. It is only meaningful if turbo boost is disabled which no-one ever does. For example, the 2020 13" Macbook Pro pulls 65W under load:
MacBook Pro 13 2020 in Review: Apple's subnotebook only gets the mandatory update
Notebookcheck reviews the new 2020 Apple MacBook Pro 13 w/ 4x Thunderbolt 3. Except for the new keyboard, however, the update is pretty boring.
While the 12.9" ipad Pro draws 15W
Apple iPad Pro 12.9 (2018, LTE, 256 GB) Tablet Review
Notebookcheck.com reviews the new iPad Pro 12.9. The third-generation iPad Pro is equipped with a powerful A12X SoC, 4/6 GB of RAM and up to 1 TB of storage. The iPad Pro is now made from recycled aluminium, while its complete redesign seems like an attempt by Apple to justify its exorbitant...
Given theses 2 chips have approximately the same performance, Apple is at a minimum already 4X more efficient with a 2018 chip vs a 2020 Intel chip. The A14 core in the first macs will be much more efficient than the A12 core.
Given those power figures, I believe Apple can deliver a chip with 8 high performance cores and 4 efficiency cores and keep it under 30W max power usage (which is still a lot less than the supposedly 28W TDP chip in the 13" 2020 macbook Pro which can pull 65W).
would you say you make a lot of assumptions about what a person you disagree is or isn’t familiar with, or just a few?
Well, you did say adding more cores will decrease the single core score. This is patently false.
that’s not exactly what I said nor meant. In any case, you are making a projection based on the assumption that
1. The reason to not use the A13 cores in an iPad/dev kit is that it wasn't worth apple's efforts.
2. You are also assuming it is desirable to use substantially all the benefit of the 5nm process for cpu speed (as opposed to battery life, faster memory transfer and a more powerful gpu)
but i'm happy to have provided you a good laugh with an expectation that is different that yours. i guess we'll have to wait and see.