Raspberry Pi's V3DV Vulkan Driver Can Now Run The Zink OpenGL Translation Layer

[ed31337]

I can't very easily compile Chromium under the 4GB limit of my current RPi4. I'm sure not going to buy the Pi-400, even though it's kind of cute. Really hoping for a 16GB Pi someday, then all my RAM issues would be gone. The 8GB is very tempting, but I just know that as soon as I buy it, they'll release the next better model.

[oiaohm]

But for the average consumer without technical skills, and people with low incomes, perhaps with children, the PI 400 is a great start. For many people even $70 is a lot of money, and $170 would be out of the question. For example, imagine struggling parents with two children. For $140 they could provide them with systems well suited for Internet access, but for $340 they could not.

The lower grade keyboard means its really not as great of start as one would think. Repetitive Strain injury (RSI) is really simple to forget. Cherry MX or clones in the blue class are in fact designed to teach good typing pressure. They were in fact designed to reduce RSI risk. Yes 100 dollars more is a lot more but to upgrade the keyboard to a good mechanical keys is like the change between improper fitting shoes and proper fitting ones. Long term there is going to be a very big difference. Its one fo the things that people who start getting early RSI from typing find out that their choice of keyboard has been partly to blame. The cost in treating RSI is way more than prevention.

Next is life span of keyboard. I learnt this when my school bought a stack of typing tutor items with membrane keyboards those did not last 12 months but the old apple II and microbees with mechanical keyboards kept on going. There is a very big difference in the number of keystrokes each key can take between proper full stroke keys and the short keys as well. Yes short stroke mechanical have less than 1/4 the lifespan and a membrane has 1/20 the lifespan at best 1/1000000 at worst. Yes proper mechanical I am classing as baseline here they are just close to the highest durability keyboard you can get. .

Lot of people fail to notice how many laptops are scrap because their keyboards have failed. Or how often they are replacing their cheap 15 dollar keyboards. Yes a 100 dollar in decent mechanical keyboard over the long term is cheaper.

There is always a point in cost where you get the best quality. Going up by 100 dollars to improve the keyboard means the keyboard will not be the reason the device fails. Going up by 100 dollars gives you the ability to extend the keyboard out to a full standard layout with full standard keycaps this is another thing that allows extending the life of the device due to be standard keycaps they are replaceable same with being a cherry MX design key. Yes also means this device will be a functional hand me down for quite some time.

Also something I would have considered is making the USB keyboard controller on the board designed so it can be broken away. This is so main-board in device can be replaced in future while keeping the keyboard controller.

I lived in a house hold where 70 dollars was a lot. Yet I was still wearing 150 dollar shoes because the cheaper shoes back then did not last as long. So that $140 vs $340 may not be the numbers. If the 70 dollar device for lets say the 5 years of high-school is only doing 1 year before breaking that 70*5 that is 350 dollars for 1 child that makes the 140 dollars not look expensive at all. Its really easy with lower device durability to make up 100 dollars worth fo cost particularly when the item costs 70 dollars.

There is a reason why I would say go for a design where its usable as a good mechanical keyboard as well has having a PI. Because a good mechanical keyboard can give a person a good 20 years of service even 10 years takes 170 dollars back to 17 dollars a year for how long you can keep on using it for. Mechanical keyboard are fairly cheap long term up front cost is a little ouch.

Lets try to get decent keyboards into everyone hands or at least give them the option to choose it. PI 400 is a goodish idea having computer integrated into keyboard its just a pity its not a great keyboard. Yes not being a great keyboard is going to come its failure point,

Yes to market the version with the better keyboard having the 70 dollar one still could be useful. As in for 70 dollars you can by this one with a compact keyboard with less durability or you can buy the 170 dollar one here that is a full keyboard that can be used as either a computer on its own or as a keyboard for a computer that should serve you well into the future.

[pal666]

For this reason I predict that a performance similar to the DXVK will never reach

dxvk is not that fast btw, it's fast compared to wine's dx on opengl, but native dx would be fastest

[pal666]

In future Zink will be the only way to run old OpenGL software on newer hardware. It isn't going to be worth writing both a Vulkan and an OpenGL driver.

opengl driver is already written in mesa. what you need to write is handling of new hardware generation, shader compiler etc. all of it can be shared between vulkan and opengl drivers without extra zink overhead

[pal666]

I think it will be the future for many future drivers, specially for SoC-GPUs. Not because it is technically the best solution. But you do not develop 2 drivers.

you do not develop 2 drivers now. you share all hardware-specific code

[pal666]

1. On macOS it's deprecated and will be removed in a future version (like we had with 32bit recently). Zink + MoltenVK will be the free opensource way to run OpenGL (next to the commerical closed source MoltenGL).

if you target macos, you need moltenzink without extra hoops of vulkan

[pal666]

V3DV and Zink are the final reason for me to get again a Raspberry Pi.

i'm not sure i follow. 104 fps wasn't enough for you and you waited for 40 fps?

[pal666]

I can't very easily compile Chromium under the 4GB limit of my current RPi4

use binary package or cross-compile from amd64

[oiaohm]

dxvk is not that fast btw, it's fast compared to wine's dx on opengl, but native dx would be fastest

Sorry it not that straight forwards. There are some games where dxvk is faster than Windows DX11 and at the same time using less CPU(and this is the clue to the problem). Native dx is not always faster. You have to remember that dx 11 and dx12 are abstraction layers themselves this means if cards are missing some features that games uses you can end up using a software emulation in Microsoft DX. This leads to some interesting cases where dxvk is mapping feature to vulkan so using the GPU and but those using windows 10 are using software rendering on CPU for that feature. Yes there is kind of a signature if Microsoft DX11 is running higher CPU than dxvk majority of the time dxvk is the faster one.

Dxvk generally between 15% worse to 15% better than Windows DX 11. Yes the cases of 15% better making the game us dxvk on windows will also see you get the 15%. Microsoft DX is not without the areas where it suxs in performance.

Lot of people think DX11 and DX12 under windows would be perfectly tuned the reality they are not and that is why dxvk can win in some cases. Cases where dxvk win have historically caused people to trip network based anti-cheat. Winning is not always a good thing.

[pal666]

Sorry it not that straight forwards. There are some games where dxvk is faster than Windows DX11 and at the same time using less CPU(and this is the clue to the problem). Native dx is not always faster.

native dx is always faster when it has same quality of implementation. windows dx can have worse implementation and be slower, i'm talking about (potential)linux native dx