201–210 of 328 posts
I know it's my PTSD, the words make sense but I really hope this is not code for "we can't get much more performance out of our architectures, so we'll focus in selling "efficiency".. "
I really hope I'm wrong, since the power consumed by this chips has gotten a bit out of hand ( not as much as the GPUs.. )
The last time a chip company talked about "Efficiency" in this way, we got +10 years of 5% increments in performance and 50% in price increases for 2 and 4 core CPUs.. I know it's my PTSD, the words make sense but I really hope this is not code for "we can't get much more performance out of our architectures, so we'll focus in selling "efficiency".. " I really hope I'm wrong, since the power consumed by this chips ha…
Earlier quoted context omitted.
nested virtualization is missing on m1/m2 on both MacOS and Asahi - it's unclear if this is a hardware or software limitation.
Asahi said it's present in M2 hardware, will likely be supported in future MacOS. https://social.treehouse.systems/@marcan/109838053800961073 Asahi Linux is introducing support for some brand new Apple Silicon features faster than macOS.. M1 has a virtual GIC interrupt controller for enhanced virtualization performance. Linux supports it, macOS does not.. M2 introduced Nested Virtualization support. The patches for s…
Curious what will enable this in M2 vs M1. Looking at https://developer.arm.com/documentation/102142/0100/Nested-v... it appears to indicate nested virtualization is in Armv8.3-A and both M1 and M2 are ARMv8.5-A according to https://en.wikipedia.org/wiki/Apple_M1 and https://en.wikipedia.org/wiki/Apple_M2
Will be interesting to see what other arm cpus gain this with the Linux patches.
Have you used (non-nested) virtualization at all on Asahi? If so what is your experience with performance and overall thoughts so far?
Earlier quoted context omitted.
> Honestly I don't understand why there's not something like a 256 core ARM laptop The high power ARM cores aren’t that small. If you took the M2 and scaled it up to 256 cores, it would be almost 7 square inches. You can’t just scale a chip like that, though, so the interconnects would consume a huge amount of space as well. It would also consume over 1000W. The latest ARM chips are great, but some times I think the…
7 square inches would also include an enormous GPU and tons of accessories. The actual cores are about .6/2.3 mm², and local interconnects and L2 roughly double that. So with just those parts, 256 P-cores would be about 1.5 square inches, and 256 E-cores would be about half a square inch. And in practical terms you can fabricate a die that's a bit more than a square inch. Of course it wouldn't use 1000 watts. When yo…
AMD has done extremely well with multi-chip(let) modules. Zen cores & zen clusters (on Core Chiplet Die, CCD) are wonderfully small, and a huge amount of the regular stuff cores do is relegated to the IO Die (CCX), which is not as cutting edge. But wow there's a bunch of power burned on interconnect between CCDs and CCX. And now AMD's new southbridge, Promontory 21, made by Asmedia, is another pretty significant powe…
Looking forward to this. It seems lately you can only get I’m pretty sure people will suggest x86 options that fit into that description but again it looks like you have to scrape the internet for stuff like intel 1[2,3][1-9]00t or wathever the AMD alternative is.
People criticising the AI talk, remember that they achieved great CPU performance improvement by adding AI to the look aside buffers. Looking forward to even more efficient Zens, my next laptop will definitely be an AMD
Per my usual update schedule, I'm looking down the line to at least a Zen5(+?) upgrade in a few years time, so I hope they improve this kind of things in the future. However that's entirely up to OEMs deciding to make a good product, and a bit out of AMDs grasp.
They could score a quick win for power efficiency by making the Ryzen 9 5900 and Ryzen 9 3900 12-core 65W CPUs generally available instead of only selling them to OEMs.
If your production process and binning doesn't create enough supply for these parts, you only sell them to OEMs, because they can silently discontinue, or limit availability of these models when parts can not be found. However it'll be a PR disaster if the parts have spotty availability in retailers, and will cause wild rumors to spread, from "AMD doesn't want you to have this CPU" to "AMD is going bankrupt".