Nice summary mirv. I tend towards the view that the most important multi-threading skill is in analyzing what you are trying to do and creating separable tasks. Once you have the tasks, you can design work queues, thread allocation, and any synchronization needs with more confidence.

I think, whether OpenGL, Vulkan or D3D12, I would still be inclined to allocate a single thread/work queue to do nothing but execute GPU commands, simply to ensure the submission order and avoid subtle bugs, leaving other threads to prepare data and execute API commands that do not cause GPU submission.

One of the most difficult things to do is to determine what benefit you will actually achieve. Multi-threading is not cost-free. Even CPUs that support hardware threading generally just allow a lower-cost switch between 2 thread contexts per core. Designing something that works well on anything from 2-16 hardware threads is not always trivial.

One intriguing aspect of OpenGL 4 that offers a potential alternative method for reducing draw call overhead ( a primary driver for Mantle/Vulkan/D3D12 and their multi-threading enhancements ) is MultiDrawIndirect. This allows you to construct a collection of draw commands as a data array on the GPU, which can then be invoked as a single draw call. This could lead to a dramatic reduction in draw call overhead in certain circumstances, but I'm not sure whether it is flexible enough for all uses. If anyone has used this on Linux, I would be interested to hear their views.

View this comment - View article - View full comments

Quoting: Mountain ManSo, I'm, like:

Open source is good! Yah! :D
But Microsoft is bad! Boo! :S:

I really have no idea what to think of this.

I'm not sure it means anything much. It is a front-end compiler that translates from a Microsoft-defined programming language to a Microsoft-defined Virtual Machine instruction format.

It may be useful to someone wanting to define a programming language that compiles to the same Virtual Machine, or someone wanting to compile HLSL to a different Virtual Machine.

I don't think it either helps or harms them really, but there seems to be more interest in Vulkan/SPIR-V.

View this comment - View article - View full comments

Quoting: GuestI literally said 100 euro for Windows 10 and Office 2016, not for Windows 10 alone. And I think that you are misinformed.

Yes, sorry, I completely misread your first post, I thought you were just expecting the retail cost of Windows to be deducted if the same computer ships with Linux.

Even re-reading your first post, I'm not actually sure what you were saying, unless it was just that you REALLY hate Windows and don't see why anyone would use it :)

View this comment - View article - View full comments

Quoting: Nyamiou

Quoting: etonbearsThe problem with large OEMs offering Linux pre-installs in the way you suggest is that the financials don't work that way.

It may cost you or me 100 €/$/£ for Windows and 0 €/$/£ for Linux, but for a large OEM the costs are worked out on a per-unit basis.

Then how do you explain that?:

http://www.omgubuntu.co.uk/2017/01/dell-precision-3520-ubuntu-laptop [External Link]

Ooh, I stand corrected.

Things are obviously changing, and I admit that I haven't looked at prices in quite a while. What I was attempting to point out is that OEM contractual relationships with their supply chains mean you can't necessarily expect OEM pricing to reflect retail pricing.

The particular offer that you link to is comparing Ubuntu to the Pro Windows, which is a bit more expensive. The Dell UK site seems to have a £30 difference between Ubuntu and the Home Windows price.

It is good to see this happening, as it probably indicates either a weakening of Microsoft's position, or that Dell have much more confidence in the costs of supporting Linux. Either would be welcome.

View this comment - View article - View full comments

Quoting: GuestI was searching for "best PC games" for 2012 "13 "14 "15 "16 in google yesterday.

+- 40% or more have a native linux port :)

for example: Firewatch, Life is strange, Oxenfree, Alien Isolation, XCOM 2, Deus Ex, Pillars of Eternity, SOMA, Sunless Sea, The Talos Principle, Tomb Raider, Shadow Tactics, ..

+- 25% of the remaining "best pc games" (without linux support) have a platinum or gold rating on winehq, so they perform great on linux via the wine app :)

for example: Dishonored, The Walking Dead, Skyrim, Metro: Last Light, Path of Exile, The Sims 3, Mass Effect 3, ..

If HP, Lenovo, Acer and Asus would decide tomorrow to offer the customer two options 1) Windows 10 and Office 365 for 100 Euro or 2) GNU/Linux for 0.00 Euro

How many people would buy Windows 10 if computer devices would be sold without operating system?

There is not one good reason for Windows 10 to exist anno 2017.

Yes, we have about 1/4 of all Windows games on Steam, and 2/3 those available for MacOSX. So we are much better served than a few years ago. ( By the way, Metro:LL actually does have Linux port ).

The problem with large OEMs offering Linux pre-installs in the way you suggest is that the financials don't work that way.

It may cost you or me 100 €/$/£ for Windows and 0 €/$/£ for Linux, but for a large OEM the costs are worked out on a per-unit basis.

They actually pay very little for a copy of Windows, but assume the first-line support costs, which are amortised over a very large number of sales and so are quite low. Many also get revenue through deals to include extra software trials that you almost certainly did not want. So they basically lie if they suggest that you are paying 100 €/$/£ for Windows but force you to take an OS.

Conversely, for Linux, although the OS is free, they may actually have to pay Canonical, or some other entity for second/third line support ( i.e. to get fixes ), while their per-unit costs for first-line support will be higher than Windows due to smaller sales volumes.

This means that when companies like Dell have offered Linux pre-installs, it has been at a similar price point to Windows, which is not much of a pull factor. Sad but true.

View this comment - View article - View full comments

Quoting: Purple Library GuyWell, the big engines probably make money hand over fist. Maintaining and improving Linux support I would expect is cheaper than adding it was in the first place. And the thing about Linux is, even if there were no non-techie Linux desktops at all, Linux dominates in most other spaces and shows no sign of that diminishing. Between that and the attractiveness of the FOSS idea, Linux has a strong allegiance among programmers. Once they've jumped the hurdle of putting in support in the first place, even if keeping it isn't worth it financially it's probably worth it just for morale within the company and goodwill outside.
There's also the checklist effect. If you offer a product it's kind of hard to measure which features are making you money, which make the difference in people's choices to use your product or a competitor's. But if that kind of product has a checklist of features that most of them support, I suspect most companies would be leery of changing their product so it gets a red X beside something on the list while their competitors still have a green check mark.

So I suspect chances are pretty good Linux support in the engines will stay for a while.

I certainly hope you are right. I think we can assume a few years of speculative Linux support, but beyond that there probably has to be a commercial argument, and it is currently unclear how that will pan out. At one point the DICE team seemed quite keen to make their engine run on Linux ( enough to advertise for the expertise ), but then I think the money men overruled them.

I'm still hopeful that other AAA studios will think more about cross-platform support when they develop new engines, but they may remain risk averse without a clear financial argument.

View this comment - View article - View full comments

Yes, it is amazing how far Linux gaming has progressed since Valve started trying to marshal support for the platform. Before that, there had only really been patchy support from the likes of Loki, LGP and a few other small porting houses; now we have over 3000 Linux games on Steam.

You are also right to point out the potential for stretching the Linux purchases too thin. Arguably, this may already be the case, as even the most ardent Linux supporter would be unlikely to buy 3000 games; I have about 100, have played only a fraction of those, and I'm sure that I will never do justice to many titles.

My real concern for the future of Linux gaming as a whole ( as opposed to any particular game/genre ) is in understanding the reasons why we have the current level of support, and how the motivations of the players involved might change.

For example, a major reason for Linux game availability is the rise of high quality game engines, many of which have added Linux support. As these engines generally have many Indie developer customers, I'm guessing that as long as the Indies make enough from their Linux sales, then it will be worth the engine companies continuing to support Linux. But I don't know of any concrete figures that would indicate that continued Linux support is worth it for these companies.

A second reason for increased Linux support is the general improvement in tools/techniques/understanding of cross-platform development over the last decade. The failure of Microsoft to gain the same stranglehold on mobile devices as on the PC ( and the rise of iOS and android ) along with greater exposure of developers to Unix/Posix at University are important there. This is an improvement that is likely to continue.

But for me, the main reason for improved Linux game support remains as a hedge against the fear of the walled garden ( Microsoft, Apple, Google or whoever else ). It is certainly the reason Gabe Newell gave for developing SteamOS and the Steam Client for Linux. It is also probably why several of the hardware and game software companies are happy to support Valve in balancing power between the different industry players. Again, I have seen no figures as to the costs for these Linux supporting companies, and how they justify them against monopoly threat.

Obviously the ideal safeguard would be for Linux to become a significant games platform ( in the region of 20% market share ), but it is unclear how this might occur. The vast majority of people ( gamers included ) simply don't care enough, and/or are not sufficiently able to change their OS from pre-install, and indeed many would actively not want to. I honestly doubt that organic growth would ever result in Linux getting above 5% market share, so real growth depends on pre-installs, which have failed to gain traction so far.

I would certainly be interested to hear any concrete figures you may have heard.

View this comment - View article - View full comments

Quoting: slaapliedje...
If gamers can just refuse to buy stuff like that, then maybe the companies will stop doing it. Paradox with it's grand strategies does the same thing... Pretty sure I've spent WAY too much on Crusader Kings II.,,

Yes, one of the worst trends of recent years. Even worse is buying the "Season Pass" up front for completely unknown content. I personally preferred the more major "expansion pack", though I'm sure there are others who prefer the ability to pick and choose bite-sized pieces and not wait 6 months for an expansion.

I know that some developers have said that they lose traction for a title if there is a significant gap between game release and an expansion, and that expansions in the past have had disappointing sales. So, the logic for developers is clearly to offer bits and pieces as soon as possible, and stop when sales drop off. This doesn't feel right for a lot of people, and tends to exploit the lack of judgment of younger gamers.

I'm not sure there is a perfect answer to this, but I tend to wait for first impressions of a game. If the game seems good in its own right, I buy the base game but ignore all additional content until it is complete, and buy it all when released as a pack ( which always happens in the end ). If the game is considered incomplete or poor, I will only buy later as a "gold" edition ( or whatever ) that includes everything.

I think the worst excesses will go away eventually.

View this comment - View article - View full comments

@tuubi : Oh dear, I sincerely apologize that my comment seems to have upset or angered you. I do not comment very often, and generally try to keep my tone as neutral as possible to avoid giving offence, but I have clearly failed on this occasion.

On re-reading, I should probably have avoided the word "junk" to describe the games explosion. Some truly are that bad, but most are just uninspired. That may be in part because the games engines powering many titles lack flexibility, but is also influenced by a lot of copying what is out there already.

The rest of my comments are generalizations ( i.e. don't apply to every developer ) relating to the effect of PC gaming going mainstream. These are examples of the trends in gaming that people complain about that resonate with me, as I can think of titles that reflect each of these, especially in the big-budget portion of the market. These particular trends may not concern you, but I would be surprised if you never found a title to be unexpectedly disappointing.

I don't think old PC games are better, but the move from niche to mainstream has changed the dynamic. In the niche market developers had less competition for gamers money, and with less diverse gamer opinions could focus on fewer "hot" areas and arguably afford to be more radical in design evolution.

The larger market provides a greater opportunity to make money, but is more diverse, and more difficult to address, since gamer demands and volume of criticism reflect the added diversity. This is particularly difficult if you are a small or independent developer with low budgets and low visibility, as EgoSoft - one of my favourite indy developers and an early Linux adopter - found when they released X Rebirth.

I would also say that I agree with your earlier comment that you would never have expected this many Linux titles to even be available. I have a hundred or so, most of which I have not played yet; good or bad, I hope they keep coming.

View this comment - View article - View full comments

If you are comparing Linux gaming with Windows gaming, there is no question that, in the words of Harold Macmillan, we have "never had it so good". There has simply not been such a wide variety of games available for Linux before.

But this is not unequivocally a good sign. Along with all the genuine improvements that have been possible as games hardware and software technology have progressed, it has also become markedly easier to produce titles, leading to the huge number of games now being produced. The normal rules of supply and demand operate, meaning that most of these titles are low budget, low quality, and few will make any money.

Unfortunately, having to wade through so much junk to find the gems isn't the only problem I have these days. I also do not find many of the games trends of recent years to be particularly appealing.

I'm pretty ancient now, and started gaming in the physical world, progressing from "family" board games as a child, to complex map-and-counter games, tabletop wargames and early role-playing games in the 1970s. Computer gaming to me was a natural extension of these more mature/serious games markets, with niche appeal.

Games now have become mass-market more than specialist market; as revenue rather than craft or passion have taken front seat, many games seem to try to appeal to too many tastes in order to maximize revenue. The evidence of most titles seems to be, you can't add on-line multi-player without compromising single-player, you can't make heavy use of cinematic techniques without impacting good game-play, you can't make titles that have juvenile appeal without compromising adult appeal, and you can't use in-title payments without compromising any integrity the title might otherwise have had!

It seems to me that the mass-market causes most genres to stagnate pretty quickly. Just as with the film industry, most games become formulaic and risk-averse. I don't really think pure numbers is something Linux gamers should worry too much about, but rather that there are enough games of the type you want to play that meet a high standard of quality.

View this comment - View article - View full comments