Singh Kays is an individual blogger who writes about Apple Silicon M1 performance and power use, video formats and streaming industry topics.
HN commenters cite his M1 analysis, including retain/release speed, while one says his posts before 2021 read better than the later, more florid ones.
> The usefulness of hardware acceleration for video decoding is highly debatable.
No it isn't. There's a reason it's used on 99% of consumer devices. Hardware companies are generally not in the business of adding to the BOM cost for no reason. Linux alone is the outlier.
> It's not always much more energy efficient, but it sometimes is, but less than you'd think, GPUs need power too
"As you can see a GPU enabled VLC is 70% more energy efficient than using the CPU!"
https://devblogs.microsoft.com/sustainable-software/vlc-energy-optimization-with-gpu/
chrome-hw showing 1/4th the power consumption of chrome-sw on the same video on more recent Apple M1: https://singhkays.com/blog/apple-silicon-m1-video-power-consumption-pt-2/
Also hardware decoders have consistent performance, which is not true of CPU-based decoders. This is especially problematic & obvious at high resolutions. Windows & MacOS ultrabooks can do 4k video all day long without an issue. Linux ultrabooks get noticeably choppy at 1440p and 4k is right out.
This is also why you'll find ultra-low end SoCs regularly prioritizing hardware decoders over faster CPUs, notably those in every smart TV & the majority of TV streaming dongles/sticks/boxes. Which really shouldn't be surprising, fixed-function hardware has always been drastically more efficient than programmable hardware, and video has changed nothing about that.
> 2) It increases greatly the complexity of client software that has to implement both…
> EVERY single modern laptop has a video decode IC that is optimized for mobile usage
Mobile as in Laptop or Smartphone? Early Intel CPU requires 2W+ for AV1 4K60 content hardware decode. That is not in any shape or form "optimized for mobile usage" in my book when. A 500mW or 1W Hardware Decoding on SmartTV is not really "optimized for mobile usage" in my book. Or do you want to look at GPU Hardware accelerated Motion JPEG2000 decode? Which has even more absurd number.
And what has sharing iGPU design got to do with Video Decoder IP? ( If you are referring to Samsung and AMD ) The point is A video decoder that uses 500mW on Laptop is still more than 150mW on Smartphone.
>No, you can't. powermetrics on macOS is not granular enough to do that. If you figured out a way, please be explicit on how you did it.
Here. A whole detail article on similar findings.
https://singhkays.com/blog/apple-silicon-m1-video-power-consumption-pt-1/
Its strongest hold is definitely in Podcasting. Right now, thats basically all a podcast feed is, its an RSS feed. This is largely due to the hands off nature that Apple had with podcasting for so long, though Spotify[0] is trying to wrestle this control over the ecosystem in moving away from publishing an open RSS feed.
Its very much alive in this way though, at least for now, and I think its important to recognize the role of RSS in keeping podcast distribution fairly uniform and relatively meritocratic. I think this why it has exploded as a medium, in part. Its also why you get such a fantastic selection of great third party players.
Unfortunately, I think RSS for news is already past its prime, though alot of "indie" sites and blogs still do have feeds it seems.
[0]: https://singhkays.com/blog/how-spotify-is-killing-the-open-podcast-ecosystem/
If this[0] is accurate and I'm reading it correctly, "Total Package Power Consumption over Time" indicates no more than 4W at any time (and often way less). Maybe someone can read it for me?
A Ryzen 7000 U-series CPU by itself has a TDP (this is not power consumption but rather Thermal Design Power) of 15-30W. This seems to approximate power consumption.
This[1] explores comparing the previous generation (6800U) to an M1 Pro[1] and states that the Apple chip is ~2.4x as efficient in single core Cinebench performance while being approximately tied on Cinebench R23 multi-core.
[0] https://singhkays.com/blog/apple-silicon-m1-video-power-consumption-pt-1/
[1] https://www.notebookcheck.net/AMD-Ryzen-7-6800U-Efficiency-Review-Zen3-beats-Intel-Alder-Lake.623763.0.html
My understanding is that the performance is based around Apple silicon being optimized for the NeXT/ObjC model of retain/release, which they got down from 30 nanoseconds on Intel to 6.5ns on M1 (14ns on emulated Intel).
https://singhkays.com/blog/apple-silicon-m1-black-magic/#is-8-gb-ram-on-x86-intelamd-the-same-as-8-gb-on-apple-silicon-m1
IMO 8gb is still barely acceptable for most consumers in 2024, but not for any laptop that you'd expect to get more than say two years of usefulness out of, so… yeah it's about time.
One thing to remember though, is that the M1 chip is super optimized for macOS and perhaps these optimizations do not make it the best fit for any other os:
According to John Gruber:
> retaining and releasing NSObjects is so common on MacOS (and iOS), that making it 5 times faster on Apple Silicon than on Intel has profound implications on everything from performance to battery life.
https://daringfireball.net/2020/11/the_m1_macs
https://www.singhkays.com/blog/apple-silicon-m1-black-magic/