Steam Deck vs. ASUS ROG Ally Arch Linux Gaming Performance

[IActuallyKnowItAll]

Whole bunch of people are missing that Valve is not a hardware company, nor a game dev studio. Valve is Steam. They are making SteamDeck, so Linux gaming gains traction, Valve keeps a card against M$, who wants to make Xbox Store the exclusive Windows gaming platform. MS tries their hand, Valve pulls from Windows, taking whole bunch of customers with them.

Steam Machines, Proton or Steam Deck was never about gaining market share. If whole bunch of brands made their own Linux handheld, Valve would be so goddamn happy.

[stiiixy]

Valve IS game dev studio. It ALSO does this Steam thingy. Sure, they made a reference hardware device to set a benchmark in order to show off their work on this Steam thingy on this linux crap I hear about, but that's not hardware company, sure.

They make money from content delivery, and the rest from silly hats and corny blats in marketplace trades from those games they dont develop.

[stiiixy]

I would love ti see how much Valve makes in marketplace commissions alone.

[IActuallyKnowItAll]

You don't know which of their products has been their profit center since 2008, but you talk like you are an expert in the matter? Ok.

[stiiixy]

Duh, it's not nuclear rocketry to know Valve makes a dozen games, and constantly updates them with patches and marketplace modifications. Not sure why common knowledge of a publicly listed company makes me some sort of expert, or nothwrwise, for simply stating what really is quite obvious. Shame on me for stating the obvious. (just checked Wikipaedia...it's all right their in the opening paragraph).

I wonder, because I really couldnt actually give two shits about trolling through umpteen pages (probably all of a couple dozen ) of public finacial data, and I shall remain forever-more ignorant, oh how it burns me....oh look something more interesting over there

[DrYak]

AMD seemingly canceled the line... Otherwise Asus would have used Dragon Crest (or its successor) instead of Phoenix silicon.

In the end, it's still a chip that is only found in the SteamDeck.
So the SteamDeck still has a peculiar power target that isn't found elsewhere (all the other have higher power consumption more typical of laptops).

Anyway, this is an important distinction, as its disingenuous to say Valve has the volume to economically justify a fully custom chip like the Xbox/PS5.

Okay, yes. Point taken.
(Specially since even Gabe himself seems impressed by the current success of SteamDeck. Apparently Vlave wasn't sure how much it would be a success and ~3mio units sold is good unexpected news. So back then, during negociation AMD would surely not been able to justify anything with expeted units, specially since at that point, their previous "linux gaming machine" venture - Steam bnoxes - was a flop).

not sure which USB PD it supports to be able to play longer connected to a external battery

Both devices support 65W, but the SteamDeck ships with a 45W adapter which is still plenty for the higher power setting, but doesn't leave headroom to fast-charge the battery, Ally ships with a 65W which allow simultaneous max powerlevel and fast charging.

The Steam Deck has no business strategy. {...} Basically, 1-2 years from now, Valve will be looking at SteamDeck-by-other-bigger-vendors taking over the market they essentially built.

Valve wasn't really planning to become a big player in the hand-held console field. One of the main goal of the SteamDeck was to be a demonstrator of SteamOS 3.0 and how good it is at playing PC games with a controller.

There's no lock-in, no exlcusiveness and no follow-up hw in the pipe.

The butter smooth console-like out-of-the-box experience is the exclusive.
Windows is a very ill suited OS for this form factor.
And specially hardware manufacturer can't customise it enough.

A possible long-term strategy is for Valve to partner with other manufacturer and provide them with SteamOS. (Which again would be perfectly fine for the target that Valve had).

Regarding follow-up:

In my opinion: The next Steam Deck seriously should have a better screen. I think it is not that hard to have a cheap "garbage" variant and a more expensive "Pro-Model" with better parts and more RAM.

Which is more or less what Valve has hinted. The next will NOT be a "SteamDeck 2", that is still a couple of years away.
But some "SteamDeck +" with similar CPU/GPU perf but tweaked specs (including screen and battery life) is possible. Do not expect a bump in resolution, though. The whole point is to have developper still have one single performance to target (its not going to by PS4 / PS4Pro).

My bet is that the more expensive variant is probably going to have OLED screen (still at 800p) and perhaps hall-effect analog sensors.
Any change of CPU/GPU are going to be impact stictly the battery life only (so same Zen2 and RDNA2 arch, same clock speed, though lower power consumption at the same performance profiles) - it won't change from the point of view of game dev, but might squeeze a bit more gameplay from the battery.
Possibly also increase in SSD storage depending on what's available at the same price point by then.

there is nothing preventing Steam on the Ally on Linux to also implement it.

Nothing beside the fact that the Ally wouldn't have all the specific tweaks that Valve did to better support all the peculiarities of the SteamDeck - i.e. Sleep might be broken on Ally due to lack of appropriate drivers.
BUT given that Ally is running a derivative of a laptop CPU, chance are high that at least some laptop manufacturer is going to implement whatever is needed to support flawless sleep on those chipsets, and through the power of open-source some enthousiaste could slap it on a special variant of SteamOS (or Chimera or whatver) for Ally.

[Almindor]

Valve wasn't really planning to become a big player in the hand-held console field. One of the main goal of the SteamDeck was to be a demonstrator of SteamOS 3.0 and how good it is at playing PC games with a controller.

The butter smooth console-like out-of-the-box experience is the exclusive.
Windows is a very ill suited OS for this form factor.
And specially hardware manufacturer can't customise it enough.

A possible long-term strategy is for Valve to partner with other manufacturer and provide them with SteamOS. (Which again would be perfectly fine for the target that Valve had).

Jolla/SailfishOS tried this approach and failed miserably. Doing the amount of work required to come up with your "own" hardware and then depending on others is a failed model. I realize mobile phones/tablets are a different beast, but not by that much. SteamOS AFAIK is also fully open source so taking it, rebranding it and releasing a new HW is fairly trivial for an established player.

The main problem anyone else but Valve has at this point if they wanted to compete in this space is how to get rid of Steam itself, which however doesn't have that much to do with the Deck. It's a bigger problem (for them) present on all the platforms where Steam dominates.

[octra]

Michael

"All versions are powered by the same AMD custom "Van Gogh" APU with four Zen 2 cores (8 threads) and integrated RDNA3 graphics with 8 CUs"
I think this should be RDNA2 and not 3.

"The ASUS ROG Ally meanwhile ships with AMD's brand new Ryzen Z1 Extreme SoC that offers 8 Zen 4 cores (15 threads) and complete with RDNA3 graphics with 12 CUs."
Should be 16 threads.

[tarasis]

The comparison was a little unbalanced because SteamOS is running on an older version of the Linux kernel and is also using Mesa 23.1 compared to the latest Mesa 23.1.3, Michael it would've been better to compare them when SteamOS 3.5 is released, i think things feel a little rushed.

At this point it definitely wasn't rushed, given SteamOS 3.5 still isn't released (although at least is finally in preview properly)

[tarasis]

Which is more or less what Valve has hinted. The next will NOT be a "SteamDeck 2", that is still a couple of years away.
But some "SteamDeck +" with similar CPU/GPU perf but tweaked specs (including screen and battery life) is possible. Do not expect a bump in resolution, though. The whole point is to have developper still have one single performance to target (its not going to by PS4 / PS4Pro).

My bet is that the more expensive variant is probably going to have OLED screen (still at 800p) and perhaps hall-effect analog sensors.
Any change of CPU/GPU are going to be impact stictly the battery life only (so same Zen2 and RDNA2 arch, same clock speed, though lower power consumption at the same performance profiles) - it won't change from the point of view of game dev, but might squeeze a bit more gameplay from the battery.
Possibly also increase in SSD storage depending on what's available at the same price point by then.

I like your thinking, particularly with it basically being basically a refactor of the existing device. Its too soon to add the chaos of "this is SD+ verified, but not SD verified"

With the references to Galileo being found, one wonders if it will be a SteamDeck+ or an actual console based device. Having only just got a refurb SD a month ago I'd be a little disappointed if a SD+ does come out, but with my experience of it so far I will buy a SD2 down the line.

One improvement I'd like to see for a SD+ is making it possible to swap the SSD without all the steps needed.