Apple recently made some marketing claims using a 5nm processor against a 4-year-old architecture and we were awaiting the benchmark that we could use to make some solid comparisons. Earlier today, single-core and multi-core scores have leaked out in the latest Cinebench R23 and boy is it a different story than the Apple used in the announcement. The Apple M1 is a very impressive chip – but it is the fastest in the mobility CPU – not by a long shot.
Intel and AMD Mobility x86 CPUs also downgrade Apple M1 to the Cinebench R23 benchmark even with a node constraint
It is also interesting that many people do not take into account that Apple’s M1 is based on the 5nm process – which is a 2 node gain over the Intel 14nm and a 1 node gain over the AMD 7nm. This advantage has no relationship with Apple’s architectural / engineering skills and is simply a function of the process provided by TSMC. However, even with this constraint, Intel 10nm and AMD 7nm mobility chips outsell Apple’s M1 in the recent Cinebench R23 benchmark.
Unlike Geekbench, which runs multiple algorithms and then takes a Geomian (which algorithm-specific optimizations can skew the result if present on an architecture), Cinebench measures the brute force power of a processor and most enthusiasts. This is the benchmark for Comes to the processor. It is a rendering benchmark that is able to take full advantage of all the compute performance of a processor (and all threads) without any inefficiency. The first Cinebench R23s start with a single-core benchmark.
The Apple M1 chip scores 1498 points in the Cinebench R23 single-core benchmark. AMD’s x86 chips, which lag behind 1 node and do not target very aggressive clocks on a single core (preferring a high multicore) score up to 1284. On the other hand, the Apple M1 is easily thrust through Intel 4-core variants. 10nm: Intel Core i7 1185G7 which is about 1 node behind TSMC’s 5nm process. But of course, single-core benchmarks are purely cosmetic because most applications use multithreading. This is where the Cinebench R23 multi-core results come in:
Here, AMD 7nm chips with the loss of 1 node absolutely destroys the Apple 5nm chip. Even Intel’s 14nm processor has more or less the same score and 10nm variants (with half a core!) Within just 1000 points compared to the Apple M1. The Intel TGL 10nm 8 core variant is expected to land in a few months and is expected to quickly climb to the top of this chart, as well as the four-core results shown here.
As we can clearly see with these benchmarks, the Apple M1 is not the fastest mobility chip around a very fast shot. But I want to point out that shifting to M1 makes perfect sense for Apple. Due to its exclusivity contract with TSMC, it has access to nodes earlier than anyone else and can use this node advantage and significantly increase its profit margin without adding too much performance by combining it with the ARM architecture Can do. In fact, for the original code and mobile form factor, the performance differences are inherently negligible and will not matter.
High performance, on the other hand, is another story. I’ve deliberately not added any desktop processor to these benchmarks, but needless to say – even Intel’s 14nm desktop chips destroy the M1. The problem is that the ARM architecture is not very high clock and has high leakage at high power levels (this is also largely true, x86 is not very good in low power scenarios). This means that the bread and butter segment of the x86 market is protected from Apple’s ARM influence – until companies like NUVIA have nothing to say about it.