that's where YOU are wrong, apple tells you that their mjesus chip wipes the floor but is proven not to be true, whether you chose to accept that is a different story entirely. Mjesus mops up skylake which is a 4 year old design at this point. It does not "destory" intel's 11th gen nor amd ryzen chips at all.
I Agree, it does not destroy the 11th gen or Zen3 - but it does outperform them - and it does it quite handily if we are confined to lower power envelopes (<30w.)
I know you wont accept it, and that's fine, but the fact remains that Apple Silicon generally has a couple of percent better IPC than Zen3 and 11th gen -> and it does so @ a lower power consumption - Which is the REALLY important part.
We need not discuss this further as you are impervious to articles and facts that PROPER lab tests have shown several times. A geekbench run is worth nothing because that does not take power into account.
A Geekbench score, by itself, does not prove that one system is better than another, but Geekbench scores can be a starting point for discussion if you take context into account.
From all appearances, 'kojack' is uninterested in such context, but I will reply to his post anyway, because I believe that most of the other people participating in this thread (whether enthusiastic or skeptical about the M1) have shown an interest in hard facts.
Again, You are the one impervious to the truth. The 11th gen I9 and the ryzen cpus BOTH out perform the Mjesus chip. I am not sure on the ryzen in battery life but the i9 beats the mjesus by a good margin in battery life. But keep telling yourself what you want to hear. Lots of apple fans do.
There is no such thing as a "Mjesus chip". As for the M1 in the M1-based MacBook Air, it is turning in its scores on a 10-watt thermal budget in an entry-level laptop, in a thin chassis that has no fan. It is unfair to compare its performance to CPUs in much more expensive machines, and to CPUs that have much greater electrical and thermal budgets – and that often have help from fans with their cooling.
What's causing such a stir is how well it does even in such comparisons.
M1 vs. Ryzen CPUs
Some Ryzen CPUs do outperform the M1 in single-core benchmarks. The top single-core score on the Geekbench 5 charts is 2239, and belongs to a machine identified as an "iMacPro1,1 AMD Ryzen 7 5800X 3800 MHz (8 cores)." An AMD Ryzen 7 5800X
has a 142 watt thermal allowance, far too great to be practical for laptop use.
Despite running with both hands tied behind its back, thermally, the M1 in the Air turned in a single-core score of 1690. That is, to say, a low-end, mobile M1 chip delivered 75% of the single-core performance of a power-hungry, desktop-bound, AMD Ryzen 7 at the very top of the Geekbench 5 single-core charts.
M1 vs. mobile Intel CPUs
In a more apples-to-apples comparison, a M1-based 13" Retina Air beat a 13" rMBP with a quad-core, 2.3 GHz, Intel Core i7-1068NG7 by 1690 to 1239 (36%) on the single-core test. On multi-core tests, a M1-based Air scored 7303 to 6862 for a 16" rMBP with an 8-core 2.4 GHz Intel Core i9-9980HK. 6.4% isn't much of a win – until you consider thermal budgets (10 watts vs. 45 watts), fans (none vs. two good ones), and the fact that the chip in Apple's least expensive laptop is beating the chip in its most expensive, highest-end laptop.
If this is what Apple is selling to low-end customers, what is it planning to sell to high-end ones? If Apple chose to do so, they could split the difference on thermal output, put out a M2 with a thermal budget of 27 watts that would run rings around mobile Intel Core i9s, and use the 18 watt savings to extend the 16" rMBPs battery charge life. If I can think of this, the people designing Apple Silicon SoCs certainly can.
M1 vs. AMD EPYC (Zen)
As for AMD's EPYC CPUs, their multi-core performance is impressive. But while they are very energy-efficient, they are not appropriate chips for laptops. Their single-core scores are poor compared to the M1's (e.g., 1094 vs. 1690), and the sheer number of cores that are present in the top EPYC (Zen) chips places their power consumption at a level that is too high for a laptop.
The top-scoring EPYC-based
system had two AMD EPYC 7B12 CPUs with 64 cores each. It got a score of 62368 on the Geekbench 5 multi-core test. That's roughly 8.5 times the multi-core score of a M1-based MacBook Air. The impressive thing about the M1-based Air is not that it lost a multi-core race to a desktop-bound system that had two CPUs, 16x as many cores, and a TDP budget that was, in all likelihood,
24x to 45x higher . That was to be expected. The impressive thing is that it didn't lose by a greater margin.
128-core AMD EPYC (Zen) vs. 96-core Intel Xeon
A
system based on four
Intel Xeon Platinum 8280L CPUs with 28 cores each scored a fairly anemic 874 on the single-core test, and a respectable 45226 on the multi-core one. (The latter score was about 83% as fast as the EPYC system's score on a per-core basis.)
Xeon Platinum 8280L CPUs have a TDP of 205 watts each, so that's 820 watts for four of them. The AMD Thermal Design guide speaks of TDPs ranging from 120 watts to 225 watts. Assuming that the EPYC CPUs consumed no more than 225 watts each, that's 450 watts maximum for two.
Putting it all together, the EPYC-based system probably scored at least 138.6 multi-core points per CPU-watt. The Xeon-based system scored only 51.2 multi-core points per CPU-watt. Any questions?
(Just for kicks, I'd like to note that the M1-based Air scored 169 multi-core points per CPU-watt. It's not playing in the same data center / scientific computing niche as the other two – but the fact that it is energy-efficient, like the EPYC, bodes well for the design of desktop chips suitable for use in the iMac Pro and Mac Pro.)
