Live data from Hacker News

Nvidia's Project Digits is a 'personal AI supercomputer'

techcrunch.com

141–150 of 510 posts

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#141
post #93
post #89

Earlier quoted context omitted.

Doesn't really help though if it requires an nvidia kernel.

The Linux kernel license requires Nvidia to disclose their Linux kernel sources and Nvidia open sourced their kernel driver. That said, you can probably boot a Debian or Gentoo system using the Nvidia provided kernel if need be.

It always has been the userspace of the Jetsons which was closed source and tied to Nvidia's custom kernel. I have not heard from people running Jetpack on a different userland than the one provided by Nvidia. Companies/Labs that update the OS don't care about CUDA, Nvidia contributes to Mesa support of the Jetsons and some only need a bit more GPU power than a RasPi.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#142
post #129

Earlier quoted context omitted.

This definitely loses on CPU performance.

I originally thought so, since the previous Grace CPUs used the neoverse N2, which loses to Apple's M4 cores. However apparently 10 of the cores are the Cortex-X925 CPUs, which are a serious upgrade. Basically 10 performance cores and 10 efficiency cores that should be pretty competitive with any current apple CPU.

M4 Max is 12 performance cores and 4 efficiency cores, the former of which are basically the fastest single-core performance you can buy right now, and the latter of which have never been touched by any architecture for their energy class. It seems highly unlikely that what you say is the case.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#143
Smart move for Nvidia to subsidise their ARM CPU and platform business by selling a big GPU packet with a CPU that most users don't really care about. Even if the margin is less than selling the raw GPU power would be (which I doubt), it'll look good on the shareholders conference if other business segments go up steep, too.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#144
post #81

This looks like a successor to the Nvidia Jetson AGX Orin 64GB Developer Kit: https://www.okdo.com/wp-content/uploads/2023/03/jetson-agx-o... I wonder what the specifications are in terms of memory bandwidth and computational capability.

The AGX Orin was only 64GB of LPDDR5 and priced at $5k so this does seem like a bargain in comparison with 128GB of presumably HBM. But Nvidia never lowers their prices, so there's a caveat somewhere.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#145
How about “We sell a computer called the tinybox. It comes in two colors + pro.

tinybox red and green are for people looking for a quiet home/office machine. tinybox pro is for people looking for a loud compact rack machine.” [0]

[0] https://tinygrad.org/#tinybox

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#146
post #6

I feel this is bigger than the 5x series GPUs. Given the craze around AI/LLMs, this can also potentially eat into Apple’s slice of the enthusiast AI dev segment once the M4 Max/Ultra Mac minis are released. I sure wished I held some Nvidia stocks, they seem to be doing everything right in the last few years!

Not only that, but it should help free up the gpus for the gamers.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#147

Would love to see something like this with an ATX form factor, socketed GPU, socketed GPU and non-soldered RAM.

Isn't that just a regular PC with one or more 5090 or equivalent workstation GPU?

There would be unified GPU/CPU memory and an ARM processor that isn't soldered to the board.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#148
post #53
post #21

Earlier quoted context omitted.

This is somewhat similar to what GeForce was to gamers back in the days, but for AI enthusiasts. Sure, the price is much higher, but at least it's a completely integrated solution.

Yep that's what I'm thinking as well. I was going to buy a 5090 mainly to play around with LLM code generation, but this is a worthy option for roughly the same price as building a new PC with a 5090.

It has 128 GB of unified RAM. It will not be as fast as the 32 GB VRAM of the 5090, but what gamer cards have always lacked was memory.

Plus you have fast interconnects, if you want to stack them.

I was somewhat attracted by the Jetson AGX Orin with 64 GB RAM, but this one is a no-brainer for me, as long as idle power is reasonable.

Re: Nvidia's Project Digits is a 'personal AI supercomputer'

#149

How about “We sell a computer called the tinybox. It comes in two colors + pro. tinybox red and green are for people looking for a quiet home/office machine. tinybox pro is for people looking for a loud compact rack machine.” [0] [0] https://tinygrad.org/#tinybox

These look terrible. For 5 times the price you get worse performance.
Post reply on HN