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

I put a datacenter GPU in my gaming PC

blog.tymscar.com

Post reply on HN
Print view View original

71–80 of 199 posts

Re: I put a datacenter GPU in my gaming PC

#71

Post by segmondy »

Gracana wrote: post #48
casey2 wrote: post #9

Some resell group is going to have to make this easier. The shear amount of these cards otherwise heading towards the landfill is staggering. That is if Big Tech don't destroy them to prevent model weights from leaking.

Things like this have started to show up on eBay: https://www.ebay.com/itm/198383386991 2X NVIDIA Tesla V100 32GB NVLink Water Cooled X99 E5-2686v4 AI Workstation PC Item Quantity Intel Xeon E5-2686 v4 CPU 1 2U CPU Cooler 1 Jingyue X99 Motherboard 1 DDR3 Memory 32GB SSD 480GB AMD Radeon R5 240 4K Display Card 1 NVIDIA Tesla V100 32GB SXM2 GPU 2 NVLink SXM2 Dual-GPU Baseboard 1 Corsair Water Cooling System 2 850W Bron…

terrible deal

Re: I put a datacenter GPU in my gaming PC

#72

Post by iugtmkbdfil834 »

rafram wrote: post #62
tgv wrote: post #32

Earlier quoted context omitted.

Because their custom training data contains an emphasis on such verbiage. It doesn't come from the God-knows-how-many TB of web content the model is pre-trained on. There, such phrasing is only a drop in the sea. But the "yes, you're right" phrases, the em dash, etc., come from the later stage, for which content is created according to some (probably overprecise) guidelines.

Right. The overuse of "genuinely" most of all. Seems like they put Claude through a few good rounds of training to always answer questions about its consciousness, thoughts, etc., with something about how it's "genuinely unsure," and as a result, the model learned to use "genuinely" as an intensifier in all sorts of inappropriate contexts.

Oi, I personally use adverbs everywhere. Genuinely, kids these days.

Re: I put a datacenter GPU in my gaming PC

#73

Post by feisuzhu »

knollimar wrote: post #63
matja wrote: post #12

Earlier quoted context omitted.

The latest Raspberry Pi 5 has one 32-bit channel (2x 16-bit subchannels) of LPDDR4X-4267 SDRAM giving 17.1GB/s of bandwidth, 52x less than this GPU. Never mind lacking the CUDA and Tensor cores, so the FP16 performance is 102x less (307 GFLOPS vs 31.4 TFLOPS). So for £200, there's absolutely no comparison for this specific use-case.

Yeah thats what I'm saying. How is it so cheap????

V100 GPUs are e-waste.

Re: I put a datacenter GPU in my gaming PC

#74

Post by segmondy »

The most interesting and perhaps useful for most would be how they control the fan. If you are thinking of doing this, you really want to get those fans under control, they are loud. For anyone thinking of these, v100s idle super high! 25-35watt with nothing loaded and easily 50w when a model is loaded.

Re: I put a datacenter GPU in my gaming PC

#75

Post by markus92 »

viseyth wrote: post #61

Volta (and Pascal, which I'm using) should still be supported with driver 580 as long as you don't use the open modules, and you can use up to cuda 12.9 and cudnn 9.10.2. No need to limit yourself to an old kernel.

It is. We still run quite a few of them in prod and with 580 drivers they run just fine. Very useful GPUs still.

Re: I put a datacenter GPU in my gaming PC

#76

Post by xp84 »

20wenty wrote: post #56
whoamii wrote: post #36

The real question: did your local LLM write this post?

There are many tells aren't there? There was clearly hard human work and experimentation here, but it's a shame the OP let AI do chunks of the writing. Once you see it, it's much harder to take the post seriously.

(TL;DR Can we just judge written works by their actual content?)

I’m really in the “who gives a shit” camp on something like this. A lot of people probably have an LLM punch up a blog post. It is good at turning bullet points and notes into prose, fixing run-ons, etc. Maybe I’m naive but I trust that the kind of person who posts a clearly noncommercial post like this on HN gives a crap enough that they read the final draft and confirmed it isn’t inaccurate.

This pearl-clutching about the mere use of AI regardless of how responsible or appropriate the use is, seems like a professor in 1985 throwing an essay back in a student’s face as “this was obviously printed from a computer and not typewritten like a PROPER essay! I can tell just by looking at it!”

Re: I put a datacenter GPU in my gaming PC

#78

Post by unshavedyak »

iugtmkbdfil834 wrote: post #66
20wenty wrote: post #56

Earlier quoted context omitted.

There are many tells aren't there? There was clearly hard human work and experimentation here, but it's a shame the OP let AI do chunks of the writing. Once you see it, it's much harder to take the post seriously.

I disagree. Not everyone has a good writing style. In those instances I think it is fair to default to llm recommendation. We may be allergic to it, but we saw one formulaic response too many ( though admittedly it does raise a question of whether HN was the intended audience for it ). In any event, not all of us have a unique writing style worth preserving just like not all of us can write clear and clean code. Just…

I really wish it was more common to use AI for augmenting than authoring. Eg i find coding with LLMs neat when you primarily "talk" to it through code, by filling out structs, funcs, fields, etc - where it would use your changes as the template and then to work to effectively autocomplete the gaps. The more you iteratively write the less it fills in, but also the less it deviates from your intent, design, etc.

I feel like writing could use a similar harness, where it attempts to minimally reword the authors sentences, perhaps just tweaking grammar, spelling, etc. In the coding example i think the human code would be near unchangeable, the LLM would pivot around it - but in the writing example i think the human writing would have to be more mutable. I imagine it would be a configurable setting.

I've not really seen a system which focuses on this humanLLM look, but it feels interesting to me.

Re: I put a datacenter GPU in my gaming PC

#79

Post by qingcharles »

rafram wrote: post #35
qingcharles wrote: post #33

Earlier quoted context omitted.

Every single HN post has the same comment now.

Only because so many of the articles posted on HN now are AI-written, and badly, too. A lot of tech people are so impressed with LLMs’ capabilities in code that they fail to recognize how bad they are at writing enjoyable prose. And it feels like a chore to write out a whole blog post by hand when the machine could do it for you! But the result we get is so, so much worse and more annoying.

I dislike AI prose too, the cadence of it really rubs me the wrong way, but, that said we've had a lot of great, informative articles lately, written with AI help, where you just have to grit your teeth and get through them to get the underlying knowledge.

I don't think that commenting on every article is going to make the posters suddenly decide to go back and rewrite it by hand. Some of them probably don't even speak English natively. The comments are getting more tiresome than the AI prose at this point.

Hopefully in a year or so the LLM output won't be so janky and obvious, so this might just be a phase everyone has to pull through.

Post reply on HN