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

Nvidia is proposing a beast of a CPU system for Windows PCs

twitter.com

Post reply on HN
Print view View original

211–220 of 581 posts

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#211

Post by b112 »

vlovich123 wrote: post #205
stego-tech wrote: post #197

The Unified Memory pool is what will continue to be the “game changer” in systems architecture, especially outside of data centers. The reality is even cutting edge games and consumer workloads don’t actually take full use of the PCIe bandwidth of the GPU or the bandwidth of its GDDR memory. Even local AI use cases don’t substantially or meaningfully benefit from faster memory, at least to average consumers. A unifie…

> (which is good for Rust adherents, I figure). As a Rust adherent, please do not put words in our mouths or set up unrealistic expectations for other people by linking together concepts at a very shallow level. Language level memory safety has no answer for hardware security flaws which is what side channel attacks are. No programming language can provide memory privacy if another chip in your machine can read your…

But... it's rust.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#212

Post by alt227 »

modeless wrote: post #119

The Qualcomm Snapdragon X2 Elite Extreme trounces Nvidia's chip in single core CPU performance. It beats Intel and AMD's best, too. It has unified memory. It's the only CPU in the same league as Apple's M-series in both CPU performance and power efficiency. And it's available in laptops today, not later this year. People are sleeping on Qualcomm.

Why do people care so much about single core performance? We are all professionals here and I bet most of our workloads are multi core. I get that these new arm chips from Apple and Qualcomm are great at one thing at a time, but for professional workloads high end x64 chips still cannot be beaten on the desktop.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#213

Post by jmyeet »

stego-tech wrote: post #197

The Unified Memory pool is what will continue to be the “game changer” in systems architecture, especially outside of data centers. The reality is even cutting edge games and consumer workloads don’t actually take full use of the PCIe bandwidth of the GPU or the bandwidth of its GDDR memory. Even local AI use cases don’t substantially or meaningfully benefit from faster memory, at least to average consumers. A unifie…

Unified memory is only a feature because NVidia so aggressively uses VRAM for market segmentation.

The 5090 ($2k MSRP but realistically $3-3.5k) is almost the same as the RTX 6000 Pro (~$10k). Same memory bandwidth (1800GB/s). Slightly different CUDA cores (21k vs 24k). Big difference? VRAM (32GB vs 96GB).

NVidia ultimately doesn't want to upset this segmentation so the RTX Spark will never undermine their other offerings. This is why I think Apple has a real market opportunity if they choose to embrace it.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#214

Post by reactordev »

adabyron wrote: post #185
reactordev wrote: post #133

Earlier quoted context omitted.

10000000x this. They have been sleeping on Arm since windows phone. I just don’t see them ever having an original thought again. They could have had a 128core arm chip by now.

They have original thoughts! It's just that those employees get squashed by other divisions or having to meet short term quarterly profits it seems. There's also the whole giant trillion dollar company doesn't want to invest and let small ideas grow. They only focus on things that move the needle, which isn't much at the size. Had Microsoft executed and invested, they could have made a come back imo in both search, m…

No doubt the individuals have thoughts. It’s just corporate never actually does ANY of them. It’s a MBA mill.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#215

Post by stego-tech »

Asmod4n wrote: post #204
stego-tech wrote: post #197

The Unified Memory pool is what will continue to be the “game changer” in systems architecture, especially outside of data centers. The reality is even cutting edge games and consumer workloads don’t actually take full use of the PCIe bandwidth of the GPU or the bandwidth of its GDDR memory. Even local AI use cases don’t substantially or meaningfully benefit from faster memory, at least to average consumers. A unifie…

yeah, you only see double digits in performance degradation from going from pcie 5 to 3 with a 5090 (at x16 speed), with everything else its like in the single digits area.

And the thing we gamers forget is that we’re the outlier. We’re the edge case.

Most consumers will never really care about, let alone see, the difference in PCIe or memory bandwidth impacts from such a shift to unified memory pools. We might (being, at least in my case, a huge nerd), but I’m increasingly of the opinion that if modern blockbuster games are built for upscaling/reconstruction anyhow, then suddenly such sacrifices to performance seem acceptable relative to the gains in efficiency.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#217

Post by alt227 »

emsign wrote: post #121

They are useless if RAM prices are this high. $800 laptops with maximum 8GB are currently the norm, Windows 11 can't run on them decently. No matter how fast the SoC is with overpriced RAM they are slow. Systems that can make good use of them with 64-128GB are not affordable anymore thanks to Nvidia and co. This is a smokescreen. They'll probably sell them packaged as compute modules anyway.

> Windows 11 can't run on them decently

Windows 11 can run just fine on 8Gb of memory, what cant is Google Chrome.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#218

Post by wmf »

2OEH8eoCRo0 wrote: post #7

Are their enterprise orders slowing down? Why use precious maxed out fab capacity on consumer stuff when it could be an enterprise chip?

This chip was designed before the shortages. I think they'll order just enough units to say they released it but not enough to put a dent in Rubin.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#219

Post by bradfa »

re-thc wrote: post #208
bradfa wrote: post #125

Earlier quoted context omitted.

Qualcomm is a “fool me once, shame on you, fool me twice, you don’t fool me twice” kind of situation. So many horrible experiences in the past that people are going to be hesitant. Qualcomm are trying harder now it seems. But it will take time to repair their reputation in the PC market.

> Qualcomm are trying harder now it seems. Not really, the 1st. iteration got stuck in legal land and other delays.

They hired a good number of smart people who know how to do open source. So they’re trying. We shall see if it works.

Re: Nvidia is proposing a beast of a CPU system for Windows PCs

#220

Post by zozbot234 »

jmyeet wrote: post #213
stego-tech wrote: post #197

The Unified Memory pool is what will continue to be the “game changer” in systems architecture, especially outside of data centers. The reality is even cutting edge games and consumer workloads don’t actually take full use of the PCIe bandwidth of the GPU or the bandwidth of its GDDR memory. Even local AI use cases don’t substantially or meaningfully benefit from faster memory, at least to average consumers. A unifie…

Unified memory is only a feature because NVidia so aggressively uses VRAM for market segmentation. The 5090 ($2k MSRP but realistically $3-3.5k) is almost the same as the RTX 6000 Pro (~$10k). Same memory bandwidth (1800GB/s). Slightly different CUDA cores (21k vs 24k). Big difference? VRAM (32GB vs 96GB). NVidia ultimately doesn't want to upset this segmentation so the RTX Spark will never undermine their other offe…

Even low-VRAM cards are actually very useful for running the comparatively smaller dense layers in large local MoE models. This only requires transfering very small amounts of data across the PCIe bus (similar to pipeline parallelism) so it fits nicely around the existing bottlenecks on that hardware.
Post reply on HN