The first conformant M1 GPU driver • Hacker News Board

Re: The first conformant M1 GPU driver

#231

Post by semi-extrinsic » Tue, Aug 22, 2023, 6:29 PM UTC

paulddraper wrote: post #59

europeanNyan wrote: post #36

Earlier quoted context omitted.

I agree with you completely and yet it's practically impossible finding a worthwhile alternative. I've tried finding alternatives on my last upgrade cycle and it's like having to live with endless amounts of compromises just to get away from Apple. For the iPhone, I tried looking at the Pixel for the Vanilla Android experience and long support, yet it seems like people are fighting battery life issues all the time. N…

> 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.

I don't get these battery issues people have. I use a 3 year old Huawei P30 Pro, and I still get 2 days of battery life when charged to full. Apparently I average around 4 hours of screen time per day.

Re: The first conformant M1 GPU driver

#232

Post by mschuster91 » Tue, Aug 22, 2023, 6:32 PM UTC

bee_rider wrote: post #168

jasonjmcghee wrote: post #54

Earlier quoted context omitted.

Use to be 100% linux person. I've gone through many thinkpads. M1 Air is so far beyond them it's not even close. Battery life is (not exaggerating) 5x or more in my experience. It's honest to god, much faster. Completely silent and cool. (one of my thinkpads almost burned me it got so hot, and the fans get so loud) I've never had a single crash, random restart, failure to sleep, failure to charge, driver problem, tou…

A 5x battery life improvement seems like the baseline had some issues, unless your MacBook goes like a week without charging.

My M2 MBA routinely goes past 10-16 hours of life time with simple coding stuff (phpstorm and PHP in Docker), and it stays comfortably cold all time. Most Windows laptops struggle to get more than 3-4 hours and will fry off your balls if used as actual laptops.

Re: The first conformant M1 GPU driver

#233

Post by johnnyAghands » Tue, Aug 22, 2023, 6:32 PM UTC

Huge!!

Re: The first conformant M1 GPU driver

#234

Post by mschuster91 » Tue, Aug 22, 2023, 6:34 PM UTC

freedomben wrote: post #115

matthew-wegner wrote: post #75

Earlier quoted context omitted.

Right, but that clarification was pre-edit? I'm aware of people's sensitivity to putting demands on efforts like this, which is why I tried to frame it as the genuine curiosity it is. This sort of low-level work is out of my depth. Is there painstaking probing to find different addresses/states between a Max and an Ultra? The M2 Ultra has two GPU variants (60-core and 72-core), although the Pro and Max also have GPU…

Yes I would have put it in the original (pre-edit), but better late than never :-) Yeah I've never done GPU driver hacking so I can't answer your question with optimal specificity, but there's always something new. It's usually (but not always) backwards compatible, but Apple is a special case since they are their only supported customer. That means that they can iterate very fast with flagrant disregard for backward…

> That means that they can iterate very fast with flagrant disregard for backwards compatiblity

In practice though, Apple is famous for not replacing everything left and right. IIRC, the UART part of their SoCs dates back to the first generation of iPhones, if not even earlier.

Re: The first conformant M1 GPU driver

#235

Post by ok123456 » Tue, Aug 22, 2023, 6:36 PM UTC

nla wrote: post #64

Someone should ping Tim and let him know Alyssa is not working at Apple for some strange reason.

How do you know they're not?

Re: The first conformant M1 GPU driver

#236

Post by mschuster91 » Tue, Aug 22, 2023, 6:38 PM UTC

lamontcg wrote: post #74

how close are we now to being able to plug in an eGPU into an Apple Silicon device?

That's two completely different things. I'm not even sure if the required hardware elements for eGPU - namedly, a "back transport" of rendered frames via Thunderbolt and shuffling them directly to the screen - are there on M-series machines.

Re: The first conformant M1 GPU driver

#237

Post by ta8645 » Tue, Aug 22, 2023, 6:39 PM UTC

Have been eying a MacBook Air 15-in (M2), and wondering on the possibility of running Linux on it. Anyone had success, or failure, on doing this themselves? And, how hard will it be to port this M1 GPU driver to the M2?

Re: The first conformant M1 GPU driver

#238

Post by ferbivore » Tue, Aug 22, 2023, 6:42 PM UTC

Jasper_ wrote: post #73

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…

It's because OpenGL is a dying specification based on an outdated, horrible programming model, and Khronos's conformance tests are weird and ridiculous. The conformance tests are not open-source (there's a separate "conformance suite" available on GitHub which is based on Google's dEQP, not the Khronos internal test suite), and there can be some pretty major bugs and gaps in your implementation while still getting it…

For GLES all the relevant tests are open-source, under KhronosGroup/VK-GL-CTS on GitHub.

There's a small set of legacy "confidential" tests that you have to pass for GL conformance. They can't be open-sourced for legal reasons. The current CTS working group would like to get rid of them, but it's hard to justify spending time on GL these days...

You can definitely pass conformance with a driver that's horribly broken in practice. GL/ES/GLSL are huge and there are holes found in these specs all the time. And it's not like game developers read them anyway; whatever works on their test devices gets shipped.

Re: The first conformant M1 GPU driver

#239

Post by terhechte » Tue, Aug 22, 2023, 6:43 PM UTC

ta8645 wrote: post #237

Have been eying a MacBook Air 15-in (M2), and wondering on the possibility of running Linux on it. Anyone had success, or failure, on doing this themselves? And, how hard will it be to port this M1 GPU driver to the M2?

The article says: "Conformant OpenGL® ES 3.1 drivers are now available for M1- and M2-family GPUs"

Re: The first conformant M1 GPU driver

#240

Post by tomjakubowski » Tue, Aug 22, 2023, 6:43 PM UTC

i_am_jl wrote: post #226

gumby wrote: post #211

Earlier quoted context omitted.

> 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…

>The iphone was an HTML device, loudly repudiating the proprietary (and terrible) Flash much less the crappy, mostly stillborn "mobile HTML" attempts. Skipping Flash wasn't so much an ideological decision as a practical one. At the time Steve Jobs listed a ton of reasons that they didn't implement Flash. Listed among them were concerns about it not being an open standard, inferiority to H.264, security and performanc…

Apple did find a solution in mobile Safari for touchscreen hover states on the Web. However the Web platform generally offers more affordances for accessibility than Flash ever did, which I'm sure helps.