211–220 of 855 posts
At what point Apple will put those chips in their servers or sell server chips? It only makes sense for them to take this architecture to the cloud deployments
They could be secretly working on their own cloud platform, with their data centres having a choice between M1, Pro Max ultra instances. $$$$
Is it actually more powerful than a top of the line threadripper[0] or is that not a "personal computer" CPU by this definition? I feel like 64 cores would beat 20 on some workloads even if the 20 were way faster in single core performance. [0] https://www.amd.com/en/products/cpu/amd-ryzen-threadripper-3...
The performance per watt isn’t in the same universe and that matters.
Earlier quoted context omitted.
Not sure what is required of a dev, but as an example, Adobe Premiere pro doesn't take any advantage of >1 CPU, at least on Windows. https://www.pugetsystems.com/labs/articles/Should-you-use-a-...
CPU as in core or socket? These days most CPUs are "many-CPU-cores-in-1-socket" and having X CPU cores over 1 or 2 sockets make a small difference, but software does not care about sockets.
Looks like all the people saying "just start fusing those M1 CPU's into bigger ones" were right, that's basically what they did here (fused two M1 Max'es together). And since the presenter mentioned the Mac Pro would come on another day, I wonder if they'll just do 4x M1 Max for that.
Is it actually more powerful than a top of the line threadripper[0] or is that not a "personal computer" CPU by this definition? I feel like 64 cores would beat 20 on some workloads even if the 20 were way faster in single core performance. [0] https://www.amd.com/en/products/cpu/amd-ryzen-threadripper-3...
My 1st gen 16 core Threadripper is barely faster than an M1 Pro/Max at kernel builds, so a 64 core TR3 should handily double the M1 Ultra performance. But you know, I'm still happy to double my current build perf in a small box I can stick in my closet. Ordered one :-)
Also 1st gen threadrippers are getting on a bit now, surely. It's a ~6 year old microarchitecture.
Would be interesting to get more info on the neural engine. On one hand, I find it fascinating that major manufacturers are now putting neural architectures into mainstream hardware. On the other hand I wonder what exactly it can do. To what degree are you tied into a specific neural architecture (eg recurrent vs convolutional), what APIs are available for training it, if it's even meant to be used that way (not just…
Typical "neural engines" are intended for real-time network inference, not training. Training is highly parallel and benefits more from GPU-like vector processing.
https://developer.apple.com/documentation/coreml/model_custo...
Earlier quoted context omitted.
Large businesses do. I bet in some places you could get grant/loan incentives to replace a PC fleet with these things. Back in the Pentium-4 days, iirc I was able to get almost $250k in grants and $1.5M in subsidized loans to do accelerated refresh of a PC fleet and small datacenter, all through a utility's peak load reduction program.
I don't deny such things happen but it's illogical. If you have 20 people on desktops a small fraction of them will use more energy for microwaving lunch, making coffee or on air conditioning than they will save in aggregate on this nominal reduction in power draw.
So little memory? The now outdated Mac Pro goes up to 1.5TB, only 128GB available here.
Hopefully the next gen will provide more capable and flexible memory controllers, both so they can scale the top end for a full Pro-scale offering, and so there is more memory flexibility at the lower end e.g. the ability to get an M1 with 32+GB RAM, or a Pro with 64+.
So little memory? The now outdated Mac Pro goes up to 1.5TB, only 128GB available here.