The Still New What Did You Last Do Thread

[Graf Zahl]

Wndows on ARM needs a radical change of direction if it ever wants to be successful. Instead of this locked down platform for low end systems it needs to play to Windows's strengths, but that includes better API support.

Apple has clearly shown that ARM is a viable platform, especially for laptops. But that means pushing the high end not the low end.
I think it's a telling sign that the survey ony showed a handful of users on this platform. It's really just something for tinkerers at this time.

There will be always ARM SBCs that aren't infested with Secure Boot bullshit, but there may not be one for x86 computers in the future.

And for the average user that will be beneficial - how?
If there are problems on the Linux side with Secure Boot, it's mainly a failing of the affected distros. The concept itself is absolutely fine for better system protection.

*) Note that the lower dynamic light performance compared to OpenGL on old macs have nothing to do with MoltenVK. That is happening because I didn't bother implement the uniform buffer path for Intels.

About this, is there any information available telling us which Intel chipsets are affected by reduced SSBO performance?
Is this something only relevant on older ones or does it still persist on current hardware?

[dpJudas]

With Graf proposing a middleware layer - if he's talking about a system like how Quake 2 did it with ref_soft and ref_gl - I don't think that is necessarily a bad thing - you could probably consider the ref_*.dll files wrappers, too, but they are form-fitted towards the game and are only functional for the things the game needs them to do - they are not trying to emulate a whole entire library.

A system like ref_soft and ref_gl is what the current hardware renderer is already doing. It isn't put into DLLs, but the basic principle is the same. I believe Graf was referring to using a library that abstracted away D3D/Metal, but that would mean it isn't tailored directly to GZDoom.

As for Windows ARM - I do disagree with you there, but not because I like locked down systems - I actually am hoping Microsoft embraces openness with the ARM platform and lets you install Windows on whatever-the-hell device has an unlocked boot loader.

What openness? ARM platforms are the most locked down things in existence. Just because the instruction set is "open" doesn't make the platform so.

Ultimately it doesn't matter though - there is no value proposition here for the end user. The ARM processors available (which doesn't include the M1, because Apple won't sell you one) are all pretty slow and crap compared to x64. And for embedded devices there's no logic in using Windows ARM either because why on earth would you voluntarily give away part of your profits to Microsoft? It all really boils down to a terrible deal and so far the market seems to agree because nobody is using this version of Windows.

@Graf: I don't know if Intel still got terrible SSBO support. At the time I figured it couldn't be so bad, but it certainly was on my 2014 Mac laptop.

[Graf Zahl]

As for Windows ARM - I do disagree with you there, but not because I like locked down systems - I actually am hoping Microsoft embraces openness with the ARM platform and lets you install Windows on whatever-the-hell device has an unlocked boot loader.

What openness? ARM platforms are the most locked down things in existence. Just because the instruction set is "open" doesn't make the platform so.

Things can change. Who knows what the future for ARM may bring.

Ultimately it doesn't matter though - there is no value proposition here for the end user. The ARM processors available (which doesn't include the M1, because Apple won't sell you one) are all pretty slow and crap compared to x64. And for embedded devices there's no logic in using Windows ARM either because why on earth would you voluntarily give away part of your profits to Microsoft? It all really boils down to a terrible deal and so far the market seems to agree because nobody is using this version of Windows.
[/quote]

If they continue on the current path, yes, I agree that there's no value proposition here. And I most certainly see no point investing resources into such a platform or compromising others to allow supporting it. Should the situation still be the same when the time comes to say goodbye to OpenGL I wouldn't care.

But as things are I doubt that ARM will forever remain in that pitiful niche. Apple has shown that powerful systems can be built with it but for Windows to be successful on them they definitely need to change how they support it

@Graf: I don't know if Intel still got terrible SSBO support. At the time I figured it couldn't be so bad, but it certainly was on my 2014 Mac laptop.

Well, in that case I'm going to assume it was typical first generation problems. Since I cannot find any info on the WWW, it really seems to be isolated to obsolete low end hardware.

[yum13241]

There will be always ARM SBCs that aren't infested with Secure Boot bullshit, but there may not be one for x86 computers in the future.

And for the average user that will be beneficial - how?
If there are problems on the Linux side with Secure Boot, it's mainly a failing of the affected distros. The concept itself is absolutely fine for better system protection.

About this, is there any information available telling us which Intel chipsets are affected by reduced SSBO performance?
Is this something only relevant on older ones or does it still persist on current hardware?

About Secure Boot

-----

As a concept, Secure Boot is fine. (It should be called Bootloader Signature Enforcement because that's all it does)


The problem is M$ decided to unilaterally deny anything GPLv3 to be signed ever (shim being a notable exception, GRUB isn't even signed), and signing the bootloader and adding it to the Secure Boot database isn't straight forward.

Essentially, out of the box M$ has FULL CONTROL over what will boot on your PC. Anything signed with their key will boot. Anything that isn't will just pop up a message designed to scare you. All they have to do is to push an update thru the universal backdoo- I mean Windows Update.


The first thing I do is turn off Secure Boot. Why? I know what I'm doing and I don't want to bother with signing the bootloader on every update (on Arch this can actually be quite frequent)

The distro I use (EndeavourOS) actually doesn't support secure boot OOTB anyway (and neither does arch )


I'm not stupid enough to leave my laptop lying around in my hotel room for a maid to try to open up my PC and replace the BIOS chip. If I caught her the first thing I'd do is sue. (read: maid attack)



Every backend available in GZD right now works fine for me, GLES performing best, and the others barely hitting 60 FPS.

---

Also, WHY IN HELL are 4GB laptops still sold in 2022 (soon to be 2023?)


Once I found a laptop in a store that proudly said that its graphics drivers are Microsoft drivers with all these absurd claims. "High performance" etc.


I dunno about you, but MS drivers can't even play Q13K of all things, w/o screen tearing.
---

To answer Cacodemon: If you build it yourself you automatically know what stuff is in there. So just disable it


To add one more thing: Manjaro (an Arch derivative) isn't exactly a supportable OS. Reason being that they do all sorts of shenanigans regarding their repos and in general can't be trusted to not have their certs expire again, among other things. Since GZDoom is on the AUR AND a Flatpak, a bug could arise from Manjaro effectively being always 2 weeks out of date. (read:https://manjaro.snorlax.sh) (watch:https://www.youtube.com/watch?v=5KNK3e9ScPo)

[Graf Zahl]

The problem is M$ decided to unilaterally deny anything GPLv3 to be signed ever (shim being a notable exception, GRUB isn't even signed), and signing the bootloader and adding it to the Secure Boot database isn't straight forward.

Essentially, out of the box M$ has FULL CONTROL over what will boot on your PC. Anything signed with their key will boot. Anything that isn't will just pop up a message designed to scare you. All they have to do is to push an update thru the universal backdoo- I mean Windows Update.

You should ask yourself how Microsoft was able to get this control. Wasn't it because the Linux community collectively said "NO!" when it really mattered to make themselves heard?
Ever since Secure Boot was introduced all the channels were flooded with conspiracy accusations and a general reluctance to accept the technology. Where were they when it was decided what keys to support?

[yum13241]

Microsoft played a large part in setting the UEFI spec.

[Rachael]

I don't know how true that is, but even assuming it is, that isn't really the reason.

The main thing is mainboard manufacturers have control over the digital signing keys that are manufactured out-of-the-box. There is likely a signing deal between Microsoft and those manufacturers for them to include Microsoft's keys, but even considering this it does still stink to high heavens of corruption. (And it is)

Welcome to late-stage capitalism, baby. Hope you like it. And all the corruption and exploitation and side deals and other ways to screw you that go along with it.

And before the torches and pitchforks come out about how I am a "socialist" or "communist" - just keep in mind that capitalism and selfishness go hand in hand, it's inherent and part and parcel to each other. Part of that selfishness is doing things to serve your own best interest - if you were a business magnate who had a mandate to serve their shareholders at any and all costs come hell or high water, you would engage in this fuckery too, to screw everyone else over. Or you would quit your cushy job (if you were sane). That's how corruption seeps into the system. That's how capitalism works. It rewards bad behavior, and punishes empathy.

[Graf Zahl]

Microsoft played a large part in setting the UEFI spec.

They are a big player in the business. Of course they had their word as they would be taking the brunt of implementing support for it into their OS.

Again: Where was the Linux community when it mattered? Just like with Secure Boot they were bickering and throwing around accusations of conspiracies and at least in parts were hoping that the whole thing would fail. It's easy to get pushed to the wayside then.

Here's the thing: If you refuse to play a constructive role in such a business your concerns won't be heard and most likely be ignored as a result.

[yum13241]

I don't know how true that is, but even assuming it is, that isn't really the reason.

The main thing is mainboard manufacturers have control over the digital signing keys that are manufactured out-of-the-box. There is likely a signing deal between Microsoft and those manufacturers for them to include Microsoft's keys, but even considering this it does still stink to high heavens of corruption. (And it is)

Welcome to late-stage capitalism, baby. Hope you like it. And all the corruption and exploitation and side deals and other ways to screw you that go along with it.

And before the torches and pitchforks come out about how I am a "socialist" or "communist" - just keep in mind that capitalism and selfishness go hand in hand, it's inherent and part and parcel to each other. Part of that selfishness is doing things to serve your own best interest - if you were a business magnate who had a mandate to serve their shareholders at any and all costs come hell or high water, you would engage in this fuckery too, to screw everyone else over. Or you would quit your cushy job (if you were sane). That's how corruption seeps into the system. That's how capitalism works. It rewards bad behavior, and punishes empathy.

First thing I do if and when I get a new PC? Wipe it clean, turn off Secure Boot, and replace it with my stuff.

There was this person I was "debating" with on Reddit of all places who made a poll saying "Is capitalism good?" and I voted no. Then when I announced that in the comments, OP (not sure) was like "CaPiTaLiSm PuLlS pEoPlE oUt Of PoVeRtY. sOcIaLiSm AnD cOmMuNiSm CoNtRoL wItH gUnS" (read: i don't remember this all well but smth along these lines happened.


What I last did: Install a new copy of Office 2021 for my cousin.

[Chris]

Again: Where was the Linux community when it mattered?

They were there. "The Linux community" isn't a single monolithic hive-mind, each person has their own thoughts on it, and many of them expressed their thoughts. But thoughts aren't money, and that's something Microsoft has a lot of to get their way.

In either case, as was said, Secure Boot itself isn't the problem, it's the fact that Microsoft can and does make deals with individual computer manufacturers to have it on by default with their keys in it. Linux users will have a hard time getting influence over these manufacturers to avoid these problems because it has such a small user base, and it has such a small user base in part because of all the difficulties Microsoft has ensured it would have (between FUD campaigns trying to claim Linux users were criminals, and the unwillingness to work with open standards while making their own propriety, which increases cost to port/support other systems). Yes, Linux has its own share of problems, but you need more exposure/users to really address them, and when that growth is stunted by profit-motivated companies, it becomes a self-reenforcing problem. Linux needs to get better to get more users, but Linux needs more users to get better. And rather than letting Linux grow and improve naturally, various external pressures keep its growth and improvements down because it would be a threat to their profits.

Maybe with Valve putting their weight behind Linux, we can start seeing more growth and improvement (which there has been significant improvements already because of them, given how much they've been working with and funding various projects for Proton/Steam Deck), but it's rather telling that a competitor would need a juggernaut like Valve and their deep pockets behind it to even have a chance of one day being noticed as a viable option next to Windows.

[Graf Zahl]

Again: Where was the Linux community when it mattered?

They were there. "The Linux community" isn't a single monolithic hive-mind, each person has their own thoughts on it, and many of them expressed their thoughts. But thoughts aren't money, and that's something Microsoft has a lot of to get their way.

True, but many of these "thoughts" were outright confrontation and dismissal and led to them being ignored.

In either case, as was said, Secure Boot itself isn't the problem, it's the fact that Microsoft can and does make deals with individual computer manufacturers to have it on by default with their keys in it. Linux users will have a hard time getting influence over these manufacturers to avoid these problems because it has such a small user base, and it has such a small user base in part because of all the difficulties Microsoft has ensured it would have (between FUD campaigns trying to claim Linux users were criminals, and the unwillingness to work with open standards while making their own propriety, which increases cost to port/support other systems). Yes, Linux has its own share of problems, but you need more exposure/users to really address them, and when that growth is stunted by profit-motivated companies, it becomes a self-reenforcing problem. Linux needs to get better to get more users, but Linux needs more users to get better. And rather than letting Linux grow and improve naturally, various external pressures keep its growth and improvements down because it would be a threat to their profits.

Yawn. That old claptrap again?
The problems Linux has aren't inflicted by Microsoft but by the Linux community itself mostly.
Just have a look at the tech issues forum here. Have you ever noticed that Linux seems to be the platform with the most reports there? I'd say the reason behind this is why it won't do better in the market.

Maybe with Valve putting their weight behind Linux, we can start seeing more growth and improvement (which there has been significant improvements already because of them, given how much they've been working with and funding various projects for Proton/Steam Deck), but it's rather telling that a competitor would need a juggernaut like Valve and their deep pockets behind it to even have a chance of one day being noticed as a viable option next to Windows.

Of course you need a strong unifying force to make progress. If you got 1000 people pulling the project in 100 directions, each one with their own ideology that is at odds with others it won't go anywhere at all. How will you ever win a battle if you don't have a strong leader that gives the directions to move in? You won't have an army but a scattered group of people not knowing what to do.

[Chris]

Just have a look at the tech issues forum here. Have you ever noticed that Linux seems to be the platform with the most reports there? I'd say the reason behind this is why it won't do better in the market.

Often because Linux versions of software tends to have less development resources/testing put into it, due to its small share not warranting a lot of resources. But there's another angle that it's not always Linux as the issue, rather that Linux builds happen to expose a preexisting issue due to simply being not-Windows. Like that recent issue with the change to CVars causing builds with GCC to crash. It just happened to not crash with MSVC/Windows (and Clang), but there was a legitimate bug with the way certain CVars were set up that was only apparent when built for Linux/GCC. Are cases like that Linux's fault? I'd say no, since even though it was technically working under Windows and not Linux, there were still bugs that can leap out and bite at any time. John Carmack would talk about porting their engines to Linux as a way to help test them, where even if id didn't release those Linux builds, they'd still make them as they'd often help find problems that existed in Windows builds too.

It's also the case that Linux users tend to be more proactive in reporting problems, so you have a higher percentage of users ready to report problems with builds that have less testing put into them on a system that doesn't have all the same quirks as Windows that is the primary development platform.

Not to say there aren't issues that are uniquely and legitimately because of Linux itself, but currently given the amount of data on the subject, I don't see reason to think it's significantly disproportionate to other OSs when accounting for other factors.

Of course you need a strong unifying force to make progress. If you got 1000 people pulling the project in 100 directions, each one with their own ideology that is at odds with others it won't go anywhere at all. How will you ever win a battle if you don't have a strong leader that gives the directions to move in? You won't have an army but a scattered group of people not knowing what to do.

Canonical did a pretty good job of rallying people behind Ubuntu. There were detractors, as there always will be with anything (Linux or not), but they maintained a steady run of popularity to the point where most people would say "make it for Ubuntu where most users are" (and many other distros didn't need to do anything to make it work for their users too). That's even what Valve did when initially bringing Steam to Linux, made it for Ubuntu and left it at Caveat Emptor for anything else; and they'd use Ubuntu for the base libraries for Linux apps on Steam (branded under "SteamOS"). But for how long they've been at it, they hardly made much of a dent in the large scheme of things.

[Graf Zahl]

Just have a look at the tech issues forum here. Have you ever noticed that Linux seems to be the platform with the most reports there? I'd say the reason behind this is why it won't do better in the market.

Often because Linux versions of software tends to have less development resources/testing put into it, due to its small share not warranting a lot of resources. But there's another angle that it's not always Linux as the issue, rather that Linux builds happen to expose a preexisting issue due to simply being not-Windows. Like that recent issue with the change to CVars causing builds with GCC to crash. It just happened to not crash with MSVC/Windows (and Clang), but there was a legitimate bug with the way certain CVars were set up that was only apparent when built for Linux/GCC. Are cases like that Linux's fault? I'd say no, since even though it was technically working under Windows and not Linux, there were still hidden bugs like that can leap out at any time. John Carmack would talk about porting their engines to Linux as a way to help test them, where even if id didn't release those Linux builds, they'd still make them as they'd often help find problems that existed in Windows builds too.
It's also the case that Linux users tend to be more proactive in reporting problems, so you have a higher percentage of users ready to report problems with builds that have less testing put into them on a system that doesn't have all the same quirks as Windows.

Yeah, sure. The problem is always elsewhere. Has it ever occured to you that this attitide of non-problem-solving may be one of the reason why Linux won't go mainstream?

Of course you need a strong unifying force to make progress. If you got 1000 people pulling the project in 100 directions, each one with their own ideology that is at odds with others it won't go anywhere at all. How will you ever win a battle if you don't have a strong leader that gives the directions to move in? You won't have an army but a scattered group of people not knowing what to do.

Canonical did a pretty good job of rallying people behind Ubuntu. There were detractors, as there always will be with anything (Linux or not), but they maintained a steady run of popularity to the point where most people would say "make it for Ubuntu where most users are" (and many other distros didn't need to do anything to make it work for their users too). That's even what Valve did when initially bringing Steam to Linux, made it for Ubuntu and left it at Caveat Emptor for anything else; and they'd use Ubuntu for the base libraries for Linux apps on Steam (branded under "SteamOS"). But for how long they've been at it, they hardly made much of a dent in the large scheme of things.

Maybe it's because they were satisfied to just play by the established rules the Linux community has set for themselves. Have they ever done a serious push into "foreign" territory? I cannot remember ever seeing such a thing.

Valve is different because they want to turn Linux into a reliable gaming platform - and that actually gives the chance that they break through the community's crust because they inevitably have to change some rules to pull this off.

[Chris]

Yeah, sure. The problem is always elsewhere. Has it ever occured to you that this attitide of non-problem-solving may be one of the reason why Linux won't go mainstream?

I specifically said there are problems Linux has, and I never once said they shouldn't be fixed or otherwise dealt with. I was talking about your comment regarding the number of bugs reported by Linux users, and why the relative numbers can be as they are, I wasn't denying their existence or whether there's anything to fix. Given the number of times I've worked to fix or improve OpenAL Soft specifically for problems GZDoom and other apps were running into (resampler options, better speaker configuration handling, output configuration options, come immediately to mind; and coming in the next release, support for different playback sample rates on Windows, which will help with performance for users running their devices at unnecessarily high rates like 96khz or 192khz), to say nothing of other fixes and improvements to the user experience (even when such issues are because of the user), I have no idea where you get "this attitide of non-problem-solving" I supposedly have. I may throw blame at who I believe to be the cause of a problem, but if there's a way I can help improve it regardless, I try to. Most developers I know are pretty much the same way (except Gnome, those guys have weird ideas).

Maybe it's because they were satisfied to just play by the established rules the Linux community has set for themselves. Have they ever done a serious push into "foreign" territory? I cannot remember ever seeing such a thing.

They tried to push into mobile, which a lot of people didn't like them focusing on it and it ultimately failed (the PinePhone seems to be working out much better, though). They also tried to replace X11 by working on Mir, which most everyone else rejected as Wayland was shaping up to be a better alternative.

[Cacodemon345]

The "attitude of non-problem solving" here refers to the reluctance to fix up the things that Apple fixed up in macOS in the Linux distributions.

Remember that their Wayland push is bad enough it led to even FreeBSD users complaining about it getting force-fed into everyone's mouths. And they still have an app format war ongoing; they still refuse GNUstep/macOS's app packaging format because static dependencies are generally considered a no-go there. They are still pushing Wayland even as unneeded or undesirable as it can be, much like systemd.

On Windows and macOS, you have multiple UI toolkits that will look and feel nearly the same in a single OS, whereas on Linux the usage of a different toolkit can be enough to make your app stick out like a sore thumb. And because libraries are far more often than not released without any synchronization with LTS releases, the moment you step outside Windows and macOS into Linux will be the moment when you have to degrade your software to appease the LTS users because the people who still think a centralized package repository of a distribution is still a far better idea than the rest still overwhelms everyone else who consider AppImages and Flatpaks the way to go forward.

Only like Google, Valve (only recently) and Apple like understands what makes anything Linux/UNIX a sell. And Valve doesn't use Wayland for the Plasma Desktop, only for Gamescope. Google intentionally kept their Steam gaming parts inside an Arch VM. And Apple did a very good job at keeping the Terminal largely unneeded.