Adding to that it’s an example of third party developers working to update/rewrite complex software to run native on M1
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.
Apple Silicon M1 Chip in MacBook Air Outperforms High-End 16-Inch MacBook Pro
- Thread starter MacRumors
- Start date
- Sort by reaction score
Adding to that it’s an example of third party developers working to update/rewrite complex software to run native on M1
And given Apple gets accused of monopolistic behaviour whenever they promote their own software, it’s probably a wise idea to let the reviews show the FCP performance - which Apple stated was 3x.
I did not realise they had an adapter for 8086/68000 either, so learningThanks for the info - didn't realise there was an adaptor for it.
I assumed it was a rad-hard version of the part. There's mixed lore spread around the net on this when you search "processor space shuttle", some specific enough to indicate that they moved to a 386 when the shuttle updated to the glass cockpit.
I eventually tracked down a reference to the story that first put this into my memory:
For Parts, NASA Boldly Goes . . . on eBay (Published 2002)
NASA is trolling Internet, including Yahoo and eBay, to find replacement parts for electronic gear that no one makes anymore; agency recently bought load of outdated medical equipment so it could scavenge Intel 8086 chips--variant of chips that powered IBM's first personal computer, in 1981...
According to that story, the 8086 is used for diagnostic systems, not necessarily the flight computer. My guess is that story got summarized as "the Space Shuttle uses an 8086", and was written down somewhere easier to find than NASA's own documentation and cited endlessly from there.
Regardless, the 4Pi seems to be the actual flight computer and the AP-101 series is based on even older tech than the 8086. Designed in 1972, ~0.5 MIPS.
My actual point still holds: people misjudge what problems are hard. Landing the space shuttle isn't the proper reference point to judge cutting edge computing, and estimating and correcting the color balance of a 720p image is probably much more computationally challenging. It would take the AP-101B 2 seconds just to touch each pixel once.
There are benchmark scores up for the various chips that they are competing against. You can judge for yourself. The single advantage is single core performance. The trend in software has been going away from single core optimization in favor of much more efficient multithreadded processing. There are multiple outlets that focus on hardware comparisons and this is one of the biggest take aways of the last few years.
Low single core performance with a lot of cores isn’t really the trend anymore. We have realized for the most broad application support you want super fast single cores that are then packaged in a multi-core setup. Throwing a lot of slow cores at something was maybe the trend 10 years ago, but we learned pretty quickly that this only worked for extremely specialized workloads like video/audio processing or 3D rendering.
The M1 should outperform the vast majority of PC systems for the vast majority of applications because it has incredibly fast single core performance and enough cores to spread the load and avoid saturation. The additional specialized silicon for machine learning and background low-power processing should give additional massive boosts to specialized tasks like those we once tried to spread across dozens of slow cpu cores.
I doubt whether it will be especially difficult for Apple to build higher wattage versions next year that take this same high performance core and include 8, 12, or even more of them in a single package. This should make for some truly eye-watering iMac Pro performance.