Comment Re:Why only Geekbench? (Score 1) 174
As with all things, it depends on the workload. Intel (and to a lesser extent AMD) have to focus a lot of their CPU improvements on the AVX side. That is, massively parallel computational workloads the likes of which seen in supercomputers and scientific simulation software.
Those likely will not run as fast on M1 (probably at a fraction of the speed). But MacBooks don't really run those types of workloads (most of the time). Its users do a fairly limited set of things:
1. Compile code (often in Swift using XCode).
2. Browse complex websites (if you think this is a trivial task, look at how much memory Safari/Chrome eats up and how much it heats up your laptop; today's websites are computationally heavy).
3. Work with media/video/3D rendering.
4. Fast and response GUI to facilitate all of this.
M1's CPU's are optimized for #1, #2 and #4 above. And they've thrown a huge amount of resources (literally) on-die to do those well. At the expense of large SIMD pipelines like Intel and AMD do (again, because they need to satisfy their supercomputer customers too).
#3 above, Apple has chosen to offload to accelerators and the GPU. Which, since they control the whole software stack, they can get most apps to do as well within months of release.
That's how they get such wild IPC (for workloads of #1 and 2) increases. And yes, if you even look at the Spec2k, Chrome Octane, PCMark or any other standard PC benchmark that is able to run on iPads, they do have a significantly higher IPC than their x86 counterparts.