apple M1 Pro

[Magnum]

I'll explain what that means. It means that you're providing meaningless facts as arguments in favor of and trying to justify your purchase. You're disregarding any other modern hardware and only compare it to either an older apple device with known issues or a 15 year old CPU. You run an ancient engine and praise M1 for a single result that's just outside your expectations. This is on topic because your post is nothing but an advertisement for apple.

first, it was NOT trying to justify my purchase. I don't need to do that since I bought it of my own free will. I posted the result because the NPS represents a pretty amazing number for a laptop. May be faster ones today. But probably not by a lot. Wasn't intended to be (nor was it) an advertisement for Apple. It was simply MY PERSONAL ESPERIENCE with the machine, nothing more, nothing less... Talk about a "troll?" pot, kettle.

Dear Bob, any progress in porting Crafty to M1 (implementing maybe new technologies to better exploit the new ARM chip) ?

Kind regards, Alex
One of your best fans since '90 (you are really a pionieer in open source chess engines )

I hope to see Crafty on Apple M1 and Apple M2 soon.

[Werewolf]

There appears to be a variety of people who claim Apple is well underway with a 48 core M2 Extreme for the (early) 2023 Mac Pro.

I imagine Crafty would get insane nps on that.

[smatovic]

If it is UMA across dies and CPU+GPU+TPU it would be a fun thing to program a chess engine for, memory centric design, or alike, rumor has it that M2 Extreme could get an external memory controller for TB of RAM.

--
Srdja

[Modern Times]

Even with the existing M1, M2 variations you wonder what could be possible with say Lc0. For example, in the Windows/Intel/AMD world the DX12 version of that came along some time after the standard CUDA version. Is something like that possible for MacOS and the M1/M2 family ?

[smatovic]

Even with the existing M1, M2 variations you wonder what could be possible with say Lc0. For example, in the Windows/Intel/AMD world the DX12 version of that came along some time after the standard CUDA version. Is something like that possible for MacOS and the M1/M2 family ?

Apple M-series silicon uses the Apple Metal shader language for GPU and GPGPU, I guess they need an Ankan for the M-series to code a Metal Lc0 backend, my take would be to sponsor team Lc0 with some M1 devices, but I doubt that Apple has serious interest in chess engines, M1 has 1024 to 2x4096 stream cores on die, and still no Nvidia-like TensorCores but a dedicated neural-engine, hence not a real Nvidia killer in regard of Lc0 IMO, wattage left out.

--
Srdja