Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A CPU that runs entirely on GPU

github.com

Post reply on HN
Print view View original

131–140 of 144 posts

Re: A CPU that runs entirely on GPU

#131

Post by fulafel »

st_goliath wrote: post #76
bob1029 wrote: post #28

A fun experiment but I wonder how many out there seriously think we could ever completely rid ourselves of the CPU. It seems to be a rising sentiment. The cost of communicating information through space is dealt with in fundamentally different ways here. On the CPU it is addressed directly. The actual latency is minimized as much as possible, usually by predicting the future in various ways and keeping the spatial ex…

> I wonder how many out there seriously think we could ever completely rid ourselves of the CPU. It seems to be a rising sentiment. This sentiment is not a recent thing. Ever since GPGPU became a thing, there have been people who first hear about it, don't understand processor architectures and get excited about GPUs magically making everything faster. I vividly recall a discussion with some management type back in 2…

The field explored this direction before in vector computers with high bandwidth memory (Cray etc).

Re: A CPU that runs entirely on GPU

#132

Post by layla5alive »

jerf wrote: post #112
spot5010 wrote: post #80

Earlier quoted context omitted.

I don't think we get rid of the CPU. But the relationship will be inverted. Instead of the CPU calling the GPU, it might be that the GPU becomes the central controller and builds programs and calls the CPU to execute tasks.

But... why? How do you win moving your central controller from a 4GHz CPU to a multi-hundred-MHz single GPU core? If we tried this, all we'd do is isolate a couple of cores in the GPU, let them run at some gigahertz, and then equip them with the additional operations they'd need to be good at coordinating tasks... or, in other words, put a CPU in the GPU.

Surprise: there are already CPUs in the GPU - they're called things like "Command Processor" (but not only) - they're often tiny in-order ARM or RISC-V cores.

Re: A CPU that runs entirely on GPU

#133

Post by fragmede »

sdwr wrote: post #124
mamaluigie wrote: post #115

Earlier quoted context omitted.

Get used to it. The modern day solution for everything right now is to throw AI at it. Hmmm... I need to measure this piece of wood for cutting, let me take a picture of it and see what the ai says its measurement is instead of using a measuring tape because it is faster to use the AI.

That honestly sounds great! If it works...

It does? Throw the picture at ChatGPT and see what it does with it

Re: A CPU that runs entirely on GPU

#135

Post by jagged-chisel »

koolala wrote: post #50
jagged-chisel wrote: post #46

“A CPU that runs entirely on the GPU” I imagine a carefully crafted set of programming primitives used to build up the abstraction of a CPU… “Every ALU operation is a trained neural network.” Oh… oh. Fun. Just not the type of “interesting” I was hoping for.

Isn't it interesting it doesn't instantly crash from a precision error? That sounds carefully crafted to me.

Interesting, yes. Still not the kind of interesting I was expecting.

Re: A CPU that runs entirely on GPU

#136

Post by volemo »

mghackerlady wrote: post #106
volemo wrote: post #38

Earlier quoted context omitted.

Except on my stupid iPad “Pro”. :(

iirc theres an app on the app store that's basically a small alpine container

Well, there's iSH and a-Shell but they don't have GUI capability and are somewhat limited in other ways. There's also UTM, but without weird hacks you can only get SE version which is very slow.

Re: A CPU that runs entirely on GPU

#138

Post by alienbaby »

fc417fc802 wrote: post #107
chris_money202 wrote: post #53

Earlier quoted context omitted.

I don’t think the remaining issue is memory capacity. CPUs are designed to handle nonlinear memory access and that is how all modern software targeting a CPU is written. GPUs are designed for linear memory access. These are fundamentally different access patterns the optimal solution is to have 2 distinct processing units

GDDR has high bandwidth but limited capacity. Regular RAM is the opposite, leaving typical APUs memory bandwidth starved. Both types of processor perform much better with linear access. Even for data in the CPU cache you get a noticable speedup. The primary difference is that GPUs want large contiguous blocks of "threads" to do the same thing (because in reality they aren't actually independent threads).

I always understood the main difference between CPU and GPU is that CPU's are specialised to handle branching, where GPUs are not.

Re: A CPU that runs entirely on GPU

#139

Post by theamk »

sdwr wrote: post #124
mamaluigie wrote: post #115

Earlier quoted context omitted.

Get used to it. The modern day solution for everything right now is to throw AI at it. Hmmm... I need to measure this piece of wood for cutting, let me take a picture of it and see what the ai says its measurement is instead of using a measuring tape because it is faster to use the AI.

That honestly sounds great! If it works...

It works great!

(At least 90% of the time.. the other 10% it will be slightly off, and your items will come out crooked. But don't worry, there is a tiny gray disclaimer about AI making mistakes and that you need to double-chrck it, so it's not AI's fault)

Re: A CPU that runs entirely on GPU

#140

Post by clocksmith »

Proof that you are a genius:

```lean

  inductive HumanNeed where
    | retailArithmetic
    | genericLinkedInPost

  inductive IndustrySolution where
    | commodityALU
    | frontierAutocomplete

  def optimal : Need → IndustrySolution
    | .retailArithmetic => .commodityALU
    | .genericLinkedInPost => .frontierAutocomplete

  def latency : IndustrySolution → Nat
    | .commodityALU => 1
    | .frontierAutocomplete => 248000

  theorem superbowl_ads_have_not_improved_superdope_adds :
    latency (optimal .retailArithmetic)
```
Post reply on HN