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

Valve is testing ARM64 support for popular games

notebookcheck.net

Post reply on HN
Print view View original

161–170 of 202 posts

Re: Valve is testing ARM64 support for popular games

#161

Post by agosz »

sedatk wrote: post #155
agosz wrote: post #64

Earlier quoted context omitted.

ARM64EC requires source and it produces an ARM64EC executable, which runs close to ARM64 speeds. x86/x64 executables are emulated.

Just to clarify: yes, it requires the source AND re-building it. But, with a simple switch, you make your x86 code JITtable on ARM64. That's why it runs at native ARM64 speeds. Sometimes, you can't build natively for ARM64 due to some dependency, and this allows portions of your code to be faster at least. Otherwise, if you have the source AND budget to rebuild, just build it natively for ARM64 of course :)

ARM64EC doesn't really make your code JITable. It is ARM64 code with thunks to enable transitions to x64 code. That's why it runs at near native speeds; not because of the JIT. The x64 portions of the binary do get JITed though, but the ARM64EC portions are usually much faster.

All x64 code gets JITed but that's regardless of whether ARM64EC is used or not; ARM64EC allows ARM64 applications to interface with x64 binaries.

Re: Valve is testing ARM64 support for popular games

#162

Post by chithanh »

mmcnl wrote: post #151
chithanh wrote: post #65

Earlier quoted context omitted.

It seems pretty clear to me that the main target of this is ARM Chromebooks, at least for now. As these are the only products currently shipping in any relevant numbers where this makes sense.

Since you're talking volume: what about MacBooks?

MacBooks certainly have volume, but Wine/Proton on macOS is facing serious technical challenges which they only started to address (Wine 9.0 with WoW64 was only released in January). Chromebooks are much lower hanging fruit.

Re: Valve is testing ARM64 support for popular games

#163

Post by chithanh »

zamadatix wrote: post #134
chithanh wrote: post #115

Earlier quoted context omitted.

Intel certainly would not. Qualcomm as the new owner of Intel, trying to extract maximum profit out of their IP, possibly. Think of when Oracle bought Sun Microsystems.

Qualcomm would certainly not either, x86-64 is what they want to emulate with ARM... otherwise you're saying their only interest in purchasing Intel is to completely shut its main business down to reduce competition by producing neither x86 chips nor emulating them... which, apart from being extremely unlikely, would trigger even the most laid back of regulatory environments. Doubly so after already receiving over a…

In the scenario that I am envisioning, Qualcomm will be indifferent to Intel products, the only focus being to extract the maximum money out of their purchase. Oracle did the same, and nearly all Sun products are now dead or a shadow of their former selves. Plus Oracle narrowly missed the Jackpot in Oracle v. Google.

> Doubly so after already receiving over a billion in antitrust fines in 2009 for simply being "unfair" about the agreement, not even outright removing it.

Intel continued their unfair business practices, particularly in their compiler. If you wonder why Intel CPUs didn't move up from declaring CPUID Family 6 until now, that is why. More recently there was the MKL scandal. No antitrust authority was interested in either.

> Not to mention the cross-licensing deal between Intel and AMD also includes x86 patents from both sides anyways

The AMD-Intel cross-licensing agreement will need to be renewed when Qualcomm buys Intel, because it has a termination clause in case of ownership change. The companies can continue to sell existing products, but new products would need a new patent deal.

But Qualcomm won't be interested if they can just sell their Snapdragon CPUs to former x86 customers, who have no other choice because of Intel's patents on emulating the newer x86 instructions.

Re: Valve is testing ARM64 support for popular games

#164

Post by caseyy »

dudus wrote: post #146
caseyy wrote: post #10

Earlier quoted context omitted.

It’s not the end of the world to develop for macOS and iOS. These platforms are just idiosyncratic and don’t have many serious gamers. Neither smartphones nor Macs have been powerful or ergonomic enough to play current-generation games for decades. So the user base is almost entirely people who don’t play games. For a developer, the juice is not worth the squeeze (and by far!). The platforms are challenging to port f…

While this seems counterintuitive Mac makes it very convoluted to release on their platform. To the point that the low sales really don't justify. Here's a video from a game dev driving this home https://youtube.com/shorts/qRQX9fgrI4s?si=Wg2NkJlo8r0kd9NC

This is what I meant, in part, by saying how idiosyncratic the platform is. There are many more ways it is so.

I am very happy someone put the sales figures out there in the public. Not everyone can afford to if they must keep a professional relationship with Apple.

<100 sales for Mac like this guy claims are realistic numbers.

Re: Valve is testing ARM64 support for popular games

#165

Post by chithanh »

zamadatix wrote: post #149
chithanh wrote: post #130

Earlier quoted context omitted.

You would need the FTC to intervene to stop the acquisition. And currently there is a massive lobbying push by US technology companies to replace Lina Khan with someone who is more merger friendly. https://www.bloomberg.com/news/articles/2024-09-06/kamala-ha... (non-paywall archive link: https://archive.md/PhHhU )

The FTC isn't the only one that can intervene:. Qualcomm (US based) and NXP (Dutch based) blocked by China: https://www.reuters.com/article/technology/qualcomm-ends-44-... Microsoft (US based) and Activision Blizzard (US Based) blocked by UK: https://techcrunch.com/2023/08/22/microsoft-activision-uk-cm... These are far from the only examples but they come to mind immediately.

China is busy replacing imported CPUs with domestically produced ones on a massive scale. Qualcomm's China business is at risk of collapse anyway. I don't think the Chinese will be interested nor able to buy Intel and AMD x86 chips for much longer either. My guess is that Qualcomm can placate them by promising to leave Zhaoxin alone (legally binding of course).

UK CMA folded extremely quickly after FTC lost their lawsuit. Microsoft had already communicated that they would withdraw Activision Blizzard products from the UK in case the deal were continued to be blocked.

The EU is a bit of a question mark, but blocking the FTC approved merger of two US companies could be seen as too costly politically.

Re: Valve is testing ARM64 support for popular games

#166

Post by zamadatix »

chithanh wrote: post #163
zamadatix wrote: post #134

Earlier quoted context omitted.

Qualcomm would certainly not either, x86-64 is what they want to emulate with ARM... otherwise you're saying their only interest in purchasing Intel is to completely shut its main business down to reduce competition by producing neither x86 chips nor emulating them... which, apart from being extremely unlikely, would trigger even the most laid back of regulatory environments. Doubly so after already receiving over a…

In the scenario that I am envisioning, Qualcomm will be indifferent to Intel products, the only focus being to extract the maximum money out of their purchase. Oracle did the same, and nearly all Sun products are now dead or a shadow of their former selves. Plus Oracle narrowly missed the Jackpot in Oracle v. Google. > Doubly so after already receiving over a billion in antitrust fines in 2009 for simply being "unfai…

Merging the 2 comment threads for simplicity:

Taking all of your assumptions together as givens:

- Intel's unacted claims around patents 7 years ago were valid in the first place.

- No major government acts against the action because of simultaneous individual tentative politics in each and every single one. Including the EU, which has indeed blocked deals between 2 US companies even after the FTC has approved the merger (e.g. GE+Honeywell).

- Qualcomm concludes the best way to squeeze the most money out of Intel is to shitcan x86 completely by cancelling the cross-licensing deal (despite INTC having higher revenues than Qualcomm and selling competitive CPUs in markets it doesn't even target at all).

- The result of shitcanning x86 by cancelling the cross-licensing deal is people go exclusively to ARM

It's still not clear why Qualcomm would have free reign on the x86 PC market it just ate the cost of destroying. MediaTek, Samsung, and Rockchip already make consumer ARM laptops - the first two often beating Qualcomm in performance anyways. Apple also has their own walled garden offerings in the consumer ARM device space. Qualcomm's biggest advantage in the PC space was their exclusivity agreement with Microsoft for Windows on ARM but that is already dying at the end of the year. So say all the above wasn't a concern, Qualcomm really did dump billions into Intel just to stop making their processors and it was allowed, how is Qualcomm going to get their money back at all let alone better than squeezing Intel normally without first killing its main source of revenue?

Re: Valve is testing ARM64 support for popular games

#167

Post by zamadatix »

chithanh wrote: post #165
zamadatix wrote: post #149

Earlier quoted context omitted.

The FTC isn't the only one that can intervene:. Qualcomm (US based) and NXP (Dutch based) blocked by China: https://www.reuters.com/article/technology/qualcomm-ends-44-... Microsoft (US based) and Activision Blizzard (US Based) blocked by UK: https://techcrunch.com/2023/08/22/microsoft-activision-uk-cm... These are far from the only examples but they come to mind immediately.

China is busy replacing imported CPUs with domestically produced ones on a massive scale. Qualcomm's China business is at risk of collapse anyway. I don't think the Chinese will be interested nor able to buy Intel and AMD x86 chips for much longer either. My guess is that Qualcomm can placate them by promising to leave Zhaoxin alone (legally binding of course). UK CMA folded extremely quickly after FTC lost their law…

Merging the two threads for simplicity - https://news.ycombinator.com/item?id=41631935

Re: Valve is testing ARM64 support for popular games

#168

Post by hot_gril »

shmerl wrote: post #159
dwaite wrote: post #76

Earlier quoted context omitted.

Microsoft's reluctance to break things comes from commercial entities being a big source of licenses both for desktops and backend servers. If some poorly maintained mission-critical app doesn't run on Windows 11 because it hasn't had a recompile in a decade, that means a Global 2000 company isn't able to update their workforce. Apple considers it a partnership with developers, even if that partnership in reality is…

Gamers don't need to worry about "how games are supported" as along as they work. You can run many old games in Wine even though original developers stopped supporting them decades ago. Putting hurdles in the way of that is the opposite of supporting though and that's where Apple is causing a problem. This applies to refusal to support Vulkan as well as stuff like gutting 32-bit and OpenGL.

Sometimes it works, but the performance is worse. Like back when I had an Intel Mac, csgo was fine on Boot Camp but had noticeable input lag in macOS, and iirc under Wine it had less input lag but stuttered more. It'd be unplayable if I cared about winning, but I didn't care.

Re: Valve is testing ARM64 support for popular games

#169

Post by sedatk »

agosz wrote: post #161
sedatk wrote: post #155

Earlier quoted context omitted.

Just to clarify: yes, it requires the source AND re-building it. But, with a simple switch, you make your x86 code JITtable on ARM64. That's why it runs at native ARM64 speeds. Sometimes, you can't build natively for ARM64 due to some dependency, and this allows portions of your code to be faster at least. Otherwise, if you have the source AND budget to rebuild, just build it natively for ARM64 of course :)

ARM64EC doesn't really make your code JITable. It is ARM64 code with thunks to enable transitions to x64 code. That's why it runs at near native speeds; not because of the JIT. The x64 portions of the binary do get JITed though, but the ARM64EC portions are usually much faster. All x64 code gets JITed but that's regardless of whether ARM64EC is used or not; ARM64EC allows ARM64 applications to interface with x64 bina…

Okay, I had assumed JIT for x64 wasn't possible (yet) and ARM64EC enabled that partially. So, ARM64EC images are actually native ARM code that can interact with emulated (+JITted) components. For some reason, I thought native ARM64 binaries already had this capability, but they don't. Thanks, that's actually a better state of affairs in Windows on ARM land.

Re: Valve is testing ARM64 support for popular games

#170

Post by dagmx »

gpderetta wrote: post #158
dagmx wrote: post #144

Earlier quoted context omitted.

Proton didn’t exist in a vacuum though. It existed for Valves previously failed vision of the Steam Machine. Those were actual product lines that existed (briefly) to go along with Proton. The Steam Deck was Valve finally getting sick of nobody doing what they wanted, and technology getting to where they could do something more interesting than just another desktop.

Proton came well after the steam machine failed. It was plan B.

I could have worded it better but that’s my point. It was sort of their third attempt to make Linux gaming a thing (steamOS 2 + proton) before they finally came up with a compelling product of their own.
Post reply on HN