Nvidia to Acquire Arm for $40B • Hacker News Board

Re: Nvidia to Acquire Arm for $40B

#781

Post by dannyw » Thu, Sep 17, 2020, 9:36 AM UTC

teruakohatu wrote: post #184

dannyw wrote: post #162

Earlier quoted context omitted.

Really? You get 2x perf per dollar with Ampere. That’s not good enough on the pricing front?

Yes really. Performance that is cleverly hampered by RAM (and driver licensing) on the low end from an ML perspective. The only reason they can do this is because of lack of competition. The performance of 3000 series cards could be dramatically improved for large models at a modest increase in price if RAM was doubled. It really is possible to be critical of a monopoly without disparaging the product itself. It is w…

The GeForce series is targeted at gamers. Video games are not making use of more than ~8GB of VRAM.

There is literally no benefit to 90% of the audience if they doubled the RAM. Of course, they also want to do market segmentation too, but you can’t blame GeForce for not being designed for ML training.

Re: Nvidia to Acquire Arm for $40B

#782

Post by bitL » Fri, Sep 18, 2020, 2:23 PM UTC

read_if_gay_ wrote: post #718

bitL wrote: post #704

Earlier quoted context omitted.

One could argue that's what happened to NOKIA - the only non-US company controlling the largest growing market at the time. For anyone saying that it was dead already, you don't understand that despite minimal presence on the US market it controlled over 60% of the world's market and that could have lasted much longer due to loyalty of customers and upcoming MeeGo (that even US reviewers liked) if it wasn't Eloped.

Everyone I know here in Europe thought Nokia was dead when MS bought it.

Yes, when they bought it it was already dead, but I was talking about the time they were #1 and just got their new ex-MS CEO with a $20M bonus in case he sold the company. Then he proceeded to dismantle all that worked.

Re: Nvidia to Acquire Arm for $40B

#783

Post by loeg » Sat, Sep 19, 2020, 5:51 PM UTC

magic_quotes wrote: post #269

loeg wrote: post #258

Earlier quoted context omitted.

You say that, but: Nvidia's "blob" driver doesn't have CUDA, NVENC/NVDEC, nor DDC I2C support on FreeBSD — despite all of this functionality being specific to Nvidia's own hardware, and the support being present in the Linux blob, which runs on a relatively similar platform. If the only differing bits were the portability framework, this would just be a matter of adding missing support. But it isn't — the FreeBSD obj…

> If the only differing bits were the portability framework, this would just be a matter of adding missing support. But it isn't — the FreeBSD object file Nvidia publishes lacks the internal symbols used by the Linux driver. In fact it is. For Linux and FreeBSD Nvidia distributes exactly the same blob for compilation into nvidia.ko; the blobs for nvidia-modeset.ko are slightly different. (Don't take my word for it, d…

The userspace CUDA/NVENC libraries are distributed as pre-compiled artifacts in the Linux driver, and no such FreeBSD library is distributed in the FreeBSD driver package.

Re: Nvidia to Acquire Arm for $40B

#784

Post by mzs » Thu, Sep 24, 2020, 6:01 PM UTC

hastradamus wrote: post #556

Why would Nvidia spend $40B to ruin Arm? I can't see them making a return on this investment. No one wants to work with Nvidia they are notoriously roothless. I'm sure everyone is making plans to move to something else ASAP. Maybe RISC-V

Some decision makers there dislike Apple that much.