Live data from Hacker News

CPU.fail

cpu.fail

21–30 of 33 posts

Re: CPU.fail

#21
post #14

Earlier quoted context omitted.

"Is it time to just write an X86 API on top of GPUs and get rid of CPUs?" How many days are you willing to wait for your computer to boot? GPUs aren't "better" that CPUs, they're different . Between the two, CPUs probably make better GPUs than GPUs make CPUs, but it's a tough call; neither of them are very good at the other!

We've been building personal computers around GPUs for a long time now. The CPU based computer was kind of an oddball IBM "thing". Old MOS (nintendo, commodore) and ARM for example seem to have the CPU serve the GPU in most configurations I've seen.

I didn't say it would be impossible in some abstract sense. Goodness knows even a single execution unit of a modern GPU is more powerful than my first IBM computer, if you just hook it to the right things. I asked how many days you'd be willing for your computer to boot, by which I mean something like your current workstation.

Huge, huge swathes of your normal boot process would be trying to run on a single GPU execution unit, since there would be no parallellization available, and your GPU is terrible at out-of-order dispatch (i.e., basically can't do it last I knew), so all the optimization we've spent the last 50 years putting into our CPUs won't be firing. You'd basically be trying to run your computer on something that would be in the range of 100MHz down to for all I know single-digit MHz-equivalent of your current Intel or AMD CPU (after all the penalties around not using any prediction, not having the proper caches, thrashing like hell in your GPU's memory caches, and all the other effects... I'm not even sure I'm willing to promise you'll never hit code with KHz-equivalent performance; you might just get that NES performance(!)).

Your modern GPU-based computer trying to boot Windows or Linux is gonna craaaaaaawwwwwwllllll. Can it do? With the right work, yeah, probably, but you're not going to enjoy it, or be willing to use it. CPUs are terrible GPUs, but GPUs are terrible CPUs.

How well a computer could run if it was optimized for the GPU is an open question, but I guarantee you that if in some bizarre parallel universe everything was running on GPU-like hardware, but in 2010 suddenly people figured out CPUs and next year the Core Duos were available, people would be flipping out over how awesomely they perform and would be rushing to rewrite huge swathes of code in these new-fangled "in-order execution units".

Re: CPU.fail

#23
post #12

Is it time to just write an X86 API on top of GPUs and get rid of CPUs? Seems like the shortcuts we've been taking to get sequential speed are all blowing up in our faces, and fixes aren't possible without huge performance regressions.

Haha. Yeah, if you think Intel is bad at security, I’m sure you’ll love Nvidia taking over that role...

Re: CPU.fail

#24
Wow, ChromeOS decided to disable hyperthreading entirely? That seems like a pretty drastic mitigation. I wonder if that's just a short term solution or if they're actually planning to leave it that way.

Re: CPU.fail

#26
post #12

Is it time to just write an X86 API on top of GPUs and get rid of CPUs? Seems like the shortcuts we've been taking to get sequential speed are all blowing up in our faces, and fixes aren't possible without huge performance regressions.

The shortcuts that we're talking about here are all pretty trivial in terms of performance impact and in terms of what it takes to fix the issue.

Meanwhile, GPUs are designed to run thousands of threads in parallel, and orders of magnitude slower than a CPU when running only a handful of threads.

Re: CPU.fail

#28
post #14

Earlier quoted context omitted.

"Is it time to just write an X86 API on top of GPUs and get rid of CPUs?" How many days are you willing to wait for your computer to boot? GPUs aren't "better" that CPUs, they're different . Between the two, CPUs probably make better GPUs than GPUs make CPUs, but it's a tough call; neither of them are very good at the other!

We've been building personal computers around GPUs for a long time now. The CPU based computer was kind of an oddball IBM "thing". Old MOS (nintendo, commodore) and ARM for example seem to have the CPU serve the GPU in most configurations I've seen.

I’m not sure what you mean here. The GPU is a specialized parallel compute element, it doesn’t do general purpose calculations with complex branch conditions very well. That’s why it can be so massively parallel. The CPU conversely does comparatively well with branching and arbitrary memory loads.

The examples you give are gaming machines, so the programs on the CPU, especially on older machines, are based around video interrupts to facilitate smooth graphics. It’s a program choice on the CPU.

Re: CPU.fail

#29
post #21

Earlier quoted context omitted.

We've been building personal computers around GPUs for a long time now. The CPU based computer was kind of an oddball IBM "thing". Old MOS (nintendo, commodore) and ARM for example seem to have the CPU serve the GPU in most configurations I've seen.

I didn't say it would be impossible in some abstract sense. Goodness knows even a single execution unit of a modern GPU is more powerful than my first IBM computer, if you just hook it to the right things. I asked how many days you'd be willing for your computer to boot, by which I mean something like your current workstation. Huge, huge swathes of your normal boot process would be trying to run on a single GPU execu…

Last year I got a video of my 75MHz pentium booting windows faster than a web app could load on my macbook pro (1tb ssd, 16GB ram, i7). General performance in 2019 is already horrible. Rethinking everything from the ground up would be quite pleasant imho.

Re: CPU.fail

#30
post #8

> Computer makers Apple and Microsoft and browser makers Google and Mozilla are releasing patches today. Computer makers? Wouldn't that be OS makers? They are patching their OS to prevent leaking...

apple makes macs / macbooks etc, microsoft makes surfaces / surface pros...

Yes, but I couldn't find anywhere if Apple and Microsoft are patching this as a "hardware" fix for specific products.

Almost like saying that the "software maker" John Deere will fix their latest-model Haverster.

Post reply on HN