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

AMD leaps after launching AI chip that could challenge Nvidia dominance

msn.com

Post reply on HN
Print view View original

81–89 of 89 posts

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#81

Post by fragmede »

lesuorac wrote: post #47
logicchains wrote: post #43

Earlier quoted context omitted.

>From what I've heard they aren't offering competitive salaries Seems like a common thing in hardware companies, they chronically underpay, which for some reason hardware/electrical engineers seem to accept, but that makes them a last-choice for competent software engineers, who have much better-paying options.

> which for some reason Didn't you just explicitly say the reason? A SWE can go off and make Google in their Gargage; a EE can't make a fab in their garage.

Even if they could make a fab, it would still be a logistical nightmare to scale from 1-1,000 users. Meanwhile, my SaaS company could have 100,000 users thanks to the cloud, and I wouldn't even have to get up from my desk.

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#82

Post by HDThoreaun »

joshstrange wrote: post #57
skyyler wrote: post #54

Earlier quoted context omitted.

Did you read that article you linked? >The important thing to remember is don’t use possessive apostrophes with any pronouns, either possessive pronouns or possessive adjectives. >If you see an apostrophe with a pronoun, it must be part of a contraction. >its—possessive adjective of it >it’s—contraction for “it is” "its" would be correct in the root comment.

No I didn't read it fully, I knew what the concept was called, found an article, skimmed it until I was sure it was talking about the thing I thought it was. I'm almost positive I was taught in school that "it's" is a valid possessive apostrophe case and honestly I think it's stupid that it's not. I find "its" more confusing personally.

You’re thinking of its’

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#83

Post by riffic »

HDThoreaun wrote: post #82
joshstrange wrote: post #57

Earlier quoted context omitted.

No I didn't read it fully, I knew what the concept was called, found an article, skimmed it until I was sure it was talking about the thing I thought it was. I'm almost positive I was taught in school that "it's" is a valid possessive apostrophe case and honestly I think it's stupid that it's not. I find "its" more confusing personally.

You’re thinking of its’

Please prove that is a valid construction.

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#84

Post by Symmetry »

echelon wrote: post #78
Symmetry wrote: post #64

Earlier quoted context omitted.

That's from 2022. AMDs move to start generally supporting consumer cards is very recent.

Where's the official show of support? I'll believe it when I see it.

They're listed as supported on their website and they work. I'm not sure what there is besides that.

https://rocm.docs.amd.com/en/latest/release/gpu_os_support.h...

You have to click on the "Radeon" tab for the commercial cards.

Yes, it's annoying that they only officially support Ubuntu 22.04 but it is official support and you can get other OSs and cards to work.

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#85

Post by metabagel »

joshstrange wrote: post #65
skyyler wrote: post #63

Earlier quoted context omitted.

>I knew what the concept was called, found an article, skimmed it until I was sure it was talking about the thing I thought it was You just managed to summarise why having conversations with strangers is so difficult on the internet these days. Instead of considering that you were incorrect, even for a moment, you sought an article that you thought would confirm the ideas that you already had. Without even actually r…

You do understand that's not really what happened here right? The concept IS called possessive apostrophe and literally until this minute I wasn't aware that "it's" isn't grammatically correct when the "it" in the sentence is being used possessively. I didn't just find an article I thought agreed with me and fire it off, I thought "it's" was valid and riffic didn't know about possessive apostrophes (which again, I ha…

> "I got a minor piece wrong"

You were entirely wrong.

I generally remember this bit of grammar as the apostrophe replacing the missing letters of the contraction (which is not needed for the possessive situation).

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#86

Post by 65a »

thrtythreeforty wrote: post #55
65a wrote: post #46

Earlier quoted context omitted.

You can run inference and training on consumer AMD cards today. It works fine, including llama.cpp, stable diffusion, hugging face transformers, etc. Way cheaper for a given performance/VRAM target as well.

Maybe so. But it isn't confidence inspiring when I go to see which cards are supported and I see this issue: https://github.com/ROCm/ROCm/issues/1714 With Nvidia cards, I know that if I buy any Nvidia card made in the last 10 years, CUDA code will run on it. Period. (Yes, different language levels require newer hardware, but Nvidia docs are quite clear about which CUDA versions require which silicon.) I have an AMD Z…

Pretty sure my Radeon R9-285 would work if I force gfx802 offload arch when building for ROCm, but...what are you going to do with decade-old VRAM support? 2gb is not enough for anybody.

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#87

Post by anuraaga »

thrtythreeforty wrote: post #55
65a wrote: post #46

Earlier quoted context omitted.

You can run inference and training on consumer AMD cards today. It works fine, including llama.cpp, stable diffusion, hugging face transformers, etc. Way cheaper for a given performance/VRAM target as well.

Maybe so. But it isn't confidence inspiring when I go to see which cards are supported and I see this issue: https://github.com/ROCm/ROCm/issues/1714 With Nvidia cards, I know that if I buy any Nvidia card made in the last 10 years, CUDA code will run on it. Period. (Yes, different language levels require newer hardware, but Nvidia docs are quite clear about which CUDA versions require which silicon.) I have an AMD Z…

The article specifically is about AI. Don't most useful LLM models require too much RAM for consumer Nvidia cards and also often need those newer features, making it irrelevant that a G80 could run some sort of cuda code?

I'm not particularly optimistic that ecosystem support will ever pan out for AMD to be viable but this seems to be giving a bit too much credit to Nvidia for democratizing AI development, which is a stretch.

Re: AMD leaps after launching AI chip that could challenge Nvidia dominance

#88

Post by nightski »

anuraaga wrote: post #87
thrtythreeforty wrote: post #55

Earlier quoted context omitted.

Maybe so. But it isn't confidence inspiring when I go to see which cards are supported and I see this issue: https://github.com/ROCm/ROCm/issues/1714 With Nvidia cards, I know that if I buy any Nvidia card made in the last 10 years, CUDA code will run on it. Period. (Yes, different language levels require newer hardware, but Nvidia docs are quite clear about which CUDA versions require which silicon.) I have an AMD Z…

The article specifically is about AI. Don't most useful LLM models require too much RAM for consumer Nvidia cards and also often need those newer features, making it irrelevant that a G80 could run some sort of cuda code? I'm not particularly optimistic that ecosystem support will ever pan out for AMD to be viable but this seems to be giving a bit too much credit to Nvidia for democratizing AI development, which is a…

First of all, LLMs are not the only AI in existence. A lot of ML, stats, and compute can be run on consumer grade GPUs. There are plenty of problems that aren't even applicable with an LLM.

Second, you absolutely can run and fine tune many open source LLMs on one or more 3090s at a time..

But being able just to tinker, learn to write code, etc.. on a consumer GPU is a gateway to the more compute focused cards.

Post reply on HN