Howard Oakley, author of several Mac-native utilities such as Cormorant, Spundle, and Stibium, did some digging to find out why his M1 Mac felt faster than Intel Macs did, and concluded that QoS is the reason. QoS is short for Quality of Service — and it’s all about task scheduling.
There’s a very common tendency to equate “performance” with throughput—roughly speaking, tasks accomplished per unit of time. Although throughput is generally the easiest metric to measure, it doesn’t correspond very well to human perception. What humans generally notice isn’t throughput, it’s latency—not the number of times a task can be accomplished, but the time it takes to complete an individual task…
Apple’s QoS strategy for the M1 Mac is an excellent example of engineering for the actual pain point in a workload rather than chasing arbitrary metrics. Leaving the high-performance Firestorm cores idle when executing background tasks means that they can devote their full performance to the
userInitiatedanduserInteractivetasks as they come in, avoiding the perception that the system is unresponsive or even “ignoring” the user.
MacDailyNews Take: Only Apple, vertically integrated and in full control of the entire widget from software to hardware can deliver such performance to users.
We wouldn’t trade our Macs and Apple’s vertical integration for a Dell or any other Windows PC even if our Macs weren’t less expensive than comparable PC boxes… We want and need our tech to work well. — MacDailyNews, October 24, 2006
Check out Oakley’s deep dives on Mac QoS here and here.
Back in the MDN Day, we used to just say, “Is it snappy?”
When the 40 core Macs using Apple silicon appear, we might need a different word instead of ‘snappy’.
And how will you without Jobs to give it to you?
Yes! I read this yesterday on Ars and thought it fascinating and educational. I learned something. QoS and the latency issue were well considered here. I’d never quite thought about it all quite this way. Excellent article and a great take by Salter.
Man I love my M1 Mini (1 TB and 16 MB Unified memory and looking forward to my receipt of an M1 iPad Pro this Friday! I am a little nervous about life expectancy. Since nothing seems to be serviceable inside the new Mini….how many years before it becomes a paper weight?
The massive L2 cache is what makes the M1 so snappy compared to the paltry Intel offerings. The scheduling helps a little to keep it fed.
True
This is exactly what I was talking about some months back in light of a report from a expert on the differences between fundamental chip design concepts of M1 opposed to Intel. It’s not just about clock speeds it’s the fundamental approach to Arm based chips generally and Apples in particular. It’s going to be tough for Intel to make the fundamental changes to its core design to be able to compete as increasing core numbers in their design increasingly creates inefficient overheads that limit overall performance in ways that Apples design does not through the more efficient throughput of information and as we are seeing in these leaks big increases in core numbers with little effect on efficiency. Meanwhile faster clock speeds in the chase for performance just create more heat and energy demands as the other side of the coin to gain that speed.