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

The first conformant M1 GPU driver

rosenzweig.io

Post reply on HN
Print view View original

551–560 of 699 posts

Re: The first conformant M1 GPU driver

#551

Post by riedel »

jmull wrote: post #156
bsaul wrote: post #12

"Of course, Asahi Lina and I are two individuals with minimal funding. It’s a little awkward that we beat the big corporation…" Love the euphemism. This puts Apple to shame, plain and simple. They obviously don't care about standards, or compliance, because they like people to be walled in their own little private garden (still waiting for the facetime standard, or any kind of cross-platform technology created in the…

I just don't see how it's incumbent on Apple to implement a standard only because it exists. It takes significant resources and impacts release schedules to, say, support an on-going vulkan implementation... I think there would need to be a business argument for it. Avoiding "shame" probably doesn't cut it. It just strikes me as strange when people expect Apple to spend money and focus on their own personal prioritie…

I personally do not believe in those 'the market will fix it' arguments, there is not only community goodwill but also some corporate responsibility to support open standards. Since apple is dominating a large part of the market and strongly profits from locked in effects, open standards are the few things that make competition over the best OS (aka a market) work. I understand however, that it is mostly commercial or regulatory incentives that makes current big tech moce. A bit of shame is still approriate IMHO: they could easily also support an open source driver without releasing something official.

Re: The first conformant M1 GPU driver

#552

Post by kalleboo »

TillE wrote: post #82
paulddraper wrote: post #59

Earlier quoted context omitted.

> yet it seems like people are fighting battery life issues all the time I miss the days when I charged my cell phone twice a week. It didn't take 100 mexapixel photos or display 4K cat GIFs, so I understand why. But I miss that battery life nonetheless.

You can still buy basic phones that last all week, they're very cheap. But nobody wants a phone , they want a tiny portable computer.

A modern smartphone will last for literally weeks if you turn off all data and only use it for calls, the batteries are massive to support streaming video and playing games.

Re: The first conformant M1 GPU driver

#553

Post by saagarjha »

ralferoo wrote: post #196
hedora wrote: post #177

Earlier quoted context omitted.

They can also implement better battery management technology (cooling, charge rate curves, keeping the charge between 10%-90% instead of 0-100%, but reporting 0-100% to the user via scaling), etc, etc. For a good counter-example, look at the early Nissan Leafs. They burned out their batteries in a matter of a few years, but battery replacements for other brands from that time are basically unheard of. (The inherent i…

Funny, I heard the total opposite about Nissan Leafs. The industry was guesstimating that batteries would last 8-10 years. The first Nissan Leafs (which was about the first commercially mass-available EV) had battery lives where something like 90% were still going strong and still 80% of original capacity left after 13 years. Rather than the Leaf being problematic, it was the car that showed the market that worrying…

I had one of those early Leafs, the battery degradation was real. Newer vehicles seem to have much better curves.

Re: The first conformant M1 GPU driver

#554

Post by saagarjha »

marzullo wrote: post #222
insanitybit wrote: post #138

Earlier quoted context omitted.

> When most people you know use iMessage then Android is a bad experience. Isn't this because Android uses open standards for its SMS and iOS refuses to do so?

This is correct. Google put a lot of effort into making carriers adopt RCS (which has most of the functionality of iMessage), but Apple will not adopt it to keep their "competitive advantage." https://www.android.com/get-the-message/ I use an iPhone now but these kinds of business tactics and the others mentioned here really make me wish there were more competitive products on the other end.

RCS is pretty garbage without Google’s extensions. Of course this ends up being very similar to iMessage.

Re: The first conformant M1 GPU driver

#555

Post by geokon »

zamadatix wrote: post #276
softfalcon wrote: post #252

Earlier quoted context omitted.

Yeah, I know it's just a file format for scene description. I still believe rendering would work better for the industry as a whole if they could agree to all support, say, Vulkan. I doubt it'll ever happen, but a consistent graphics driver/API standard across operating systems would be pretty rad.

While not at the kernel level I think WebGPU is probably better than something like Vulkan for that "I can target this low level access to the GPU resources but expect it to work everywhere" use case. Ignoring the name hint that it was designed for the web (it works fine outside the browser), it's a lot more portable than Vulkan across its various implementations precisely because it needed to work on various devices…

Out of idle curiosity I thought I'd give it a try. Maybe run a demo on the REPL in Clojure. I try to Google "Java WebGPU" and I get no real results

I could be wrong but It seems it's technically possible to use it outside the web, but the ecosystem looks halfbaked

Re: The first conformant M1 GPU driver

#556

Post by xeonmc »

delta_p_delta_x wrote: post #531

> Instead of one row after the next, pixels in memory follow a spiral-like curve. This is called a z -order curve[1, 2]. It makes texture mapping and mipmapping significantly more cache-friendly than the usual x-y ordering, because these operations are often performed in tiles of 2^n size. I'd say it's more a fractal pattern than spiral-like, to be precise. [1]: https://demonstrations.wolfram.com/ComparingXYCurvesAnd…

Interesting. Wouldn't it be more strictly 2^(n*(2^m)) size where m is the number of dimensions?

Re: The first conformant M1 GPU driver

#557

Post by londons_explore »

hot_gril wrote: post #536
sheepscreek wrote: post #417

Earlier quoted context omitted.

Perhaps it was fast due to lack of tight security policies etc. These days, everything has to be containerized or be in a sandbox. That adds layers of overhead. Probably also why Google killed NaCl (Native Clients for Google Chrome). It was loosely in the same space as Flash and ActiveX - anyone remember that?

I've thought about that too. Can only guess about it myself. Old software tends to be faster just because it has to be, and there isn't always a compromise other than time-to-implement. Not sure how ActiveX's sandboxing worked, but I'll bet it was even less than Flash, since it was running actual x86 code.

ActiveX required that the ActiveX code you wanted to run be already installed on your computer. The security mechanism was that users obviously wouldn't install insecure stuff...

Turns out nearly every extension had gaping security holes.

Re: The first conformant M1 GPU driver

#558

Post by astrange »

Izikiel43 wrote: post #466
astrange wrote: post #269

Earlier quoted context omitted.

I don't think being able to reverse engineer your project is actually a good qualification for hiring you, considering your employees don't need to do that. Also, isn't she a high schooler/freshman? There's always internships but it's best to do other things before grinding those out.

> I don't think being able to reverse engineer your project is actually a good qualification for hiring you, Ask Mark Russinovich, now Azure CTO

I don't think any technical skills are a good way to pick someone for a leadership position.

Re: The first conformant M1 GPU driver

#559

Post by simonh »

gumby wrote: post #211
softfalcon wrote: post #197

I wonder if support for OpenGL, Vulkan, etc will improve now that Apple is partnering with nVidia, Adobe, Autodesk, Microsoft, etc around the OpenUSD rendering/animation/CAD/3D-scene format? Considering the whole schtick of OpenUSD is "one file format that renders consistently everywhere" (paraphrasing), I would be surprised if Apple doesn't use it as a means to cement more 3D software vendors into macOS land. It's r…

> In my experience, Apple doesn't like any standards it doesn't control with an iron fist I would add some nuance to this statement: "Apple likes open standards when it is weak." The iMac and early OS X went big on open standards, and Jobs made a point of pointing this out: USB for the iMac, JPEG, MPG, mp3, Postscript etc for OSX. IP/TCP built in. They even paid the danegeld for .rtf. Then as they clawed their way ba…

Or even “Apple doesn’t like open standards that are weak”. OpenGL was just no longer fit for purpose. They were competing against DirectX, needed to do ML acceleration, and at the time Vulcan didn’t exist. They had to do something, and especially given their chip strategy that must have been in full swing at the same time, Metal must have been a no brainier.

Re: The first conformant M1 GPU driver

#560

Post by saagarjha »

andrewmcwatters wrote: post #31
Buttons840 wrote: post #18

Earlier quoted context omitted.

Sometimes I think the biggest secret in this industry is that 2 people can outproduce even the largest companies. The only thing keeping everyone's job secure is that nobody can accurately identify which 2 people.

Yeah, that's definitely not true. You'll know when you interview them and they have extensive contributions to open source, an active GitHub or GitLab account, can show you things they're proud of having made in their spare time, can talk at length about technologies and implementation details none of your other candidates can, can show you their participation in mailing lists, etc. There are a billion obvious signal…

Yeah, that describes me! I’m still insulted by this, because I have worked with plenty of people who basically do not exist on the internet but are far smarter than I will ever be. You’re wrong, plain and simple.
Post reply on HN