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

Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

phoronix.com

Post reply on HN
Print view View original

321–330 of 722 posts

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#321

Post by bayindirh »

WanderPanda wrote: post #317
bayindirh wrote: post #307

Earlier quoted context omitted.

You use at least half of this stack for desktop setups. You need copying daemons, the ecosystem support (docker-nvidia, etc.), some of the libraries, etc. even when you're on a single system. If you're doing inference on a server; MIG comes into play. If you're doing inference on a larger cloud, GPU-direct storage comes into play. It's all modular.

No you don‘t need much bandwidth between cards for inference

Copying daemons (gdrcopy) is about pumping data in and out of a single card. docker-nvidia and rest of the stack is enablement for using cards.

GPU-Direct is about pumping data from storage devices to cards, esp. from high speed storage systems across networks.

MIG actually shares a single card to multiple instances, so many processes or VMs can use a single card for smaller tasks.

Nothing I have written in my previous comment is related to inter-card, inter-server communication, but all are related to disk-GPU, CPU-GPU or RAM-CPU communication.

Edit: I mean, it's not OK to talk about downvoting, and downvote as you like but, I install and enable these cards for researchers. I know what I'm installing and what it does. C'mon now. :D

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#322

Post by Ekaros »

gs17 wrote: post #31
vesrah wrote: post #22

Earlier quoted context omitted.

They're going in alphabetical order: A - Alchemist B - Battlemage C - Celestial (Future gen) D - Druid (Future gen)

Yes, I understand that. I'm saying it doesn't read as easily IMO as (modern) NVIDIA/AMD model numbers. Most numbers I deal with are base-10, not base-36.

On other hand considering Geforce is 3rd loop of base 10 maybe it is not so bad... Radeon is on other hand a pure absolute mess... Going back same 20 years.

I kinda like the idea of Intel.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#323

Post by rini17 »

keyboard_slap wrote: post #227
amelius wrote: post #209

Earlier quoted context omitted.

What if they put 8 identical GPUs in the package, each with 1/8 the memory? Would that be a useful configuration for a modern LLM?

It could work, but would it be cost-competitive?

Also, cooling.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#324

Post by jms55 »

jvanderbot wrote: post #205
1propionyl wrote: post #129

Earlier quoted context omitted.

> upscaling is absolutely vital for a reasonable experience on some games This strikes me as a bit of a sad state of affairs. We've moved beyond a Parkinson's law of computational resources –usage by games expands to fill the available resources– to resource usage expanding to fill the available resources on the highest end machines unavailable for less than a few thousand dollars... and then using that to train a mo…

Isn't it insane to think that rendering triangles for the visuals in games has gotten so demanding that we need an artificially intelligent system embedded in our graphics cards to paint pixels that look like high definition geometry? What a time to be alive. Our most advanced technology is used to cheat on homework and play video games.

> Isn't it insane to think that rendering triangles for the visuals in games has gotten so demanding that we need an artificially intelligent system embedded in our graphics cards to paint pixels that look like high definition geometry?

That's not _quite_ how temporal upscaling work in practice. It's more of a blend between existing pixels, not generating entire pixels from scratch.

The technique has existed since before ML upscalers became common. It's just turned out that ML is really good at determining how much to blend by each frame, compared to hand written and tweaked per-game heuristics.

---

For some history, DLSS 1 _did_ try and generate pixels entirely from scratch each frame. Needless to say, the quality was crap, and that was after a very expensive and time consuming process to train the model for each individual game (and forget about using it as you develop the game; imagine having to retrain the AI model as you implement the graphics).

DLSS 2 moved to having the model predict blend weights fed into an existing TAAU pipeline, which is much more generalizable and has way better quality.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#325

Post by mkl »

vundercind wrote: post #126
seritools wrote: post #112

Earlier quoted context omitted.

1440p is colloquially referred to as 2.5K, not 2K.

It'd be pretty weird if it were called 2k. 1080p is in an absolute sense or as a relative "distance" to the next-lowest thousand closer to 2k pixels of width than 4k is to 4k (both are under, of course, but one's under by 80 pixels, one by 160). It's got a much better claim to the label 2k than 1440p does, and arguably a somewhat better claim to 2k than 4k has to 4k. [EDIT] I mean, of course, 1080p's also not typical…

You are misunderstanding. 1080p, 1440p, 2160p refer to the number of rows of pixels, and those terms come from broadcast television and computing (the p is progressive, vs i for interlaced). 4k, 2k refer to the number of columns of pixels, and those terms come from cinema and visual effects (and originally means 4096 and 2048 pixels wide). That means 1920×1080 is both 2k and 1080p, 2560×1440 is both 2.5k and 1440p, and 3840×2160 is both 4k and 2160p.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#326

Post by dragontamer »

jsheard wrote: post #289
dragontamer wrote: post #285

Earlier quoted context omitted.

Because LPDDR5x is soldered on RAM. Everyone else wants configurable RAM that scales both down (to 16GB) and up (to 2TB), to cover smaller laptops and bigger servers. GPUs with soldered on RAM has 500GB/sec bandwidths, far in excess of Apples chips. So the 8GB or 16GB offered by NVidia or AMD is just far superior at vid o game graphics (where textures are the priority)

> GPUs with soldered on RAM has 500GB/sec bandwidths, far in excess of Apples chips. Apple is doing 800GB/sec on the M2 Ultra and should reach about 1TB/sec with the M4 Ultra, but that's still lagging behind GPUs. The 4090 was already at the 1TB/sec mark two years ago, the 5090 is supposedly aiming for 1.5TB/sec, and the H200 is doing 5TB/sec.

HBM is kind of not fair lol. But 4096-line bus is gonna have more bandwidth than any competitor.

It's pretty expensive though.

The 500GB/sec number is for a more ordinary GPU like the B580 Battlemage in the $250ish price range. Obviously the $2000ish 4090 will be better, but I don't expect the typical consumer to be using those.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#327

Post by timschmidt »

pixelpoet wrote: post #318
timschmidt wrote: post #314

Earlier quoted context omitted.

That 128gb is hanging off a dual channel memory bus with only 128 total bits of bandwidth. Which is why you need the GPU. The Epyc and Xeon CPUs I'm discussing have 6x the memory bandwidth, and will trade blows with that GPU.

At a mere 20x the cost or something, to say nothing about the motherboard etc :( 500 eur for 16GB of 1TB/s with tons of fp32 (and even fp64! The main reason I bought it) back in 2019 is no joke. Believe me, as a lifelong hobbyist-HPC kind of person, I am absolutely dying for such a HBM/fp64 deal again.

$1,961.19: H13SSL-N Motherboard And EPYC 9334 QS CPU + DDR5 4*128GB 2666MHZ REG ECC RAM Server motherboard kit

https://www.aliexpress.us/item/3256807766813460.html

Doesn't seem like 20x to me. I'm sure spending more than 30 seconds searching could find even better deals.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#328

Post by jmward01 »

rs_rs_rs_rs_rs wrote: post #160
enragedcacti wrote: post #133

Earlier quoted context omitted.

> Am I missing something here? Video games

It's insane how out of touch people can be here, lol

How big is NVIDIA now? You don't think breaking into that market is a good strategy? And, yes, I understand that this is targeted at gamers and not ML. That was the point of the comment I made. Maybe if they did target ML they would make money and open a path to the massive server market out there.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#329

Post by throwaway48476 »

tokioyoyo wrote: post #144
elorant wrote: post #131

Earlier quoted context omitted.

Would it though? How many people are running inference at home? Outside of enthusiasts I don't know anyone. Even companies don't self-host models and prefer to use APIs. Not that I wouldn't like a consumer GPU with tons of VRAM, but I think that the market for it is quite small for companies to invest building it. If you bother to look at Steam's hardware stats you'll notice that only a small percentage is using high…

This is the weird part, I saw the same comments in other threads. People keep saying how everyone yearns for local LLMs… but other than hardcore enthusiasts it just sounds like a bad investment? Like it’s a smaller market than gaming GPUs. And by the time anyone runs them locally, you’ll have bigger/better models and GPUs coming out, so you won’t even be able to make use of them. Maybe the whole “indoctrinate users t…

Intel sold their GPUs at negative margin which is part of why the stock fell off a cliff. If they could double the vram they could raise the price into the green even selling thousands, likely closer to 100k, would be far better than what they're doing now. The problem is Intel is run by incompetent people who guard their market segments as tribal fiefs instead of solving for the customer.

Re: Intel announces Arc B-series "Battlemage" discrete graphics with Linux support

#330

Post by kimixa »

ben_w wrote: post #206
kimixa wrote: post #184

Earlier quoted context omitted.

I think there'll be a "financial" winter - or another way a bubble burst - the investment right now is simply unsustainable, how are these products going to be monetized? Nvidia had a revenue of $27billion in 2023 - that's about $160 per person per year [0] for every working age person in the USA. And it's predicted to more than double in 2024. If you reduce that to office workers (you know, the people who might actu…

While I'd agree monetisation seems to be a challenge in the long term (analogy: spreadsheets are used everywhere, but are so easy to make they're not themselves a revenue stream, only as part of a bigger package)… > Nvidia had a revenue of $27billion in 2023 - that's about $160 per person per year [0] for every working age person in the USA As a non-American, I'd like to point out we also earn money. > as no AI is go…

Sure, robotics help many jobs, and some level of the current deep learning boom seems to have crossover in improving that - but how many of them are running LLMs that affect Nvidia's bottom line right now? There's some interesting research in that area, but it's certainly not the primary driving force. And then is the control system the limiting factor for many systems - it's probably relatively easy to get a machine today that makes a Starbucks coffee "as good as" a decently trained human. But the market doesn't seem to want that.

And I know restricting it to the US is a simplification, but so is restricting it to Nvidia, it's just to give a ballpark back-of-the-envelope "does this even make sense?" level calculation. And that's what I'm failing to see.

Post reply on HN