Support NeoGAF

[Thick Thighs Save Lives]

Don't expect Steam Deck 2.0 anytime soon, Valve has said, as the technology to launch a handheld with an adequately beefy jump in power just doesn't yet exist.

Speaking to Eurogamer ahead of today's Steam Deck OLED announcement, Valve engineers discussed the features it is adding to its shinier new handheld model that were not possible to provide back when the original Steam Deck debuted.

The company also said it was working on game projects - plural - right now which were still targeting current Steam Deck hardware performance levels (which remain unchanged in the OLED).

"Both the screen and the battery were fairly obvious things we'd have liked to do early on," Valve veteran and Steam Deck product designer Greg Coomer told me, when discussing what the company had most wanted to improve with the launch of Steam Deck OLED.

"But the screen, I think it's the biggest example of something we would have shipped in the first-generation model but we weren't able to do so because OLED screens with these characteristics in this size just did not exist.

"Back then, we really couldn't engage with a display manufacturer to do exactly what we were after because they didn't really understand the product category, or who would be buying the screen, or why it would matter. Now that picture has changed and we're able to get custom work done."

Other new features for Steam Deck OLED have been prompted by user feedback, such as the importance of having a dedicated Bluetooth antenna for situations when people were playing docked to a TV with a lot of Bluetooth controllers. Coomer noted that Valve had also seen a number of requests for a longer power cable, which the OLED model now includes.

But what couldn't Valve include? Or at least, not include yet? Valve hardware engineer Yazan Aldehayyat said that "even more performance" was the answer - and it was here we entered the territory of what Valve would consider a full Steam Deck 2.0.

"Obviously we'd love to get even more performance in the same power envelope, but that technology doesn't exist yet," Aldehayyat said. "That's what I think we'd call a Steam Deck 2.0.

"The first Steam Deck was the first moment in time where we felt like there was enough GPU performance in a portable form factor that lets you play all your Steam games. We would love for the trend of perf-per-watt to progress rapidly to do that, but it's not quite there yet."

On the upside, this means that Steam developers still have a single perfomance point to hit when ensuring their games run well on the Steam Deck platform, whether the original model or the new OLED.

But how about Valve's own games? Had it been tempted to create something new to show off its shinier handheld hardware?

"There is work like that going on at Valve but we're not debuting a piece of content that is tailored to the OLED screen," Coomer replied.

"I think even when we shipped the first version of Steam Deck it was surprising to a lot of people that we didn't really orient a lot of our own games - or create a new game - specifically targeted to our hardware or exclusive to Steam Deck. We're basically conducting ourselves the same way roughly where we don't have title that is imminent that we're going to unveil with the launch of OLED.

"But all of the titles - a lot of which exist but some are unannounced - are definitely targeting this device. So over time you'll see that from Valve but not in a big splash at the moment of launch."

Lastly, Valve discussed the issue of being able to self-repair devices - something that Steam Deck OLED leans into by being easier to open up and fiddle around with.

"When you open the Steam Deck the method of doing so is now a series of machine screws with metal threads, as opposed to screws that went directly into plastic," Coomer said. "It used to be a little bit of a destructive act, you'd degrade the plastic a bit every time you removed or put those screws back in. To some degree you'd be degrading the Steam Deck's ability to remain intact if you dropped it. Now that's no longer the case."

The importance of giving users the ability to tinker with their own devices is not lost on a company built on serving the PC market, Coomer continued.

"We basically believe in users having control over the things they purchase. As a company, we're definitely the beneficiary of operating in a space where these kind of things are open. We think having an open software stack is valuable, we think having users being able to control these things is valuable - including to put another operating system on it.

"But even having access to the hardware components, all of those things matter and we benefit when users have access to the full range of options, because users innovate with us when they have the ability to do so. Valve, as a business, we've always benefited from this openess from a PC ecosystem that treats devices and customers with open access. So we think going forward we will continue to benefit, customers will continue to benefit, and they can help each other benefit by making changes or modifications available to each other, so it lifts everybody."

Valve says "technology doesn't exist" yet for full Steam Deck 2.0

Don't expect Steam Deck 2.0 anytime soon, Valve has said, as the technology to launch a handheld with an adequately bee…

www.eurogamer.net

Ooops, this was posted while I was making mine. Anyway I added this Bloomberg article

Valve Announces New Steam Deck With OLED Screen, More Storage

The new handheld device will begin shipping on Nov. 16 and feature significantly improved visuals but no power boost

www.bloomberg.com

It's paywalled but here's an excerpt:

Steam Deck 2 is in the works, won't be available for at least 2-3 years. It'll feature a “next-generation” power upgrade "A lot of folks at the company are excited about this product," Yang said. "We are very invested in the Steam Deck."

 

[diffusionx]

You can see this with the "competitors", they really are at best a half-step above the Deck with problems like even worse battery life. I think they made the right decision by refining the existing one and waiting for a 2.0 until they can really blow it away.

They also have all the usage data and surely know that better battery life and screen is more useful than some increases to performance around the margins.

 

[Skifi28]

I don't know about steam deck 2, but I can tell you steam deck 3 is never happening. Ever.

 

[Magic Carpet]

How does this factor in with the Switch 2 rumors?

 

[Interfectum]

Apple has tech that could but they'll never do something like this.

 

[danklord]

Battery and Screen are obvious and welcome improvements. It would be cool if, by the time Steam Deck 2 is released, games had custom settings to take advantage of the hardware beyond the typical PC settings menu. I want games to release with an OPTIMIZED FOR STEAM DECK badge with a curated settings profile for the developer's "best experience."

 

[Sleepwalker]

Ooops, this was posted while I was making mine. Anyway I added this Bloomberg article

Valve Announces New Steam Deck With OLED Screen, More Storage

The new handheld device will begin shipping on Nov. 16 and feature significantly improved visuals but no power boost

www.bloomberg.com

It's paywalled but here's an excerpt

Steam Deck 2 is in the works, won't be available for at least 2-3 years. It'll feature a “next-generation” power upgrade "A lot of folks at the company are excited about this product," Yang said. "We are very invested in the Steam Deck."

 

[Banjo64]

Wait - there’s a deck Oled?! Ace man. Amazing.

 

[Holammer]

How does this factor in with the Switch 2 rumors?

Nintendo like everyone else needs to obey the laws of thermodynamics and find a balance between price/power/battery life. If it's a handheld device.

 

[Panajev2001a]

Ooops, this was posted while I was making mine. Anyway I added this Bloomberg article

Valve Announces New Steam Deck With OLED Screen, More Storage

The new handheld device will begin shipping on Nov. 16 and feature significantly improved visuals but no power boost

www.bloomberg.com

It's paywalled but here's an excerpt

 

[Darkkahn]

4D gamin, here we come!

 

[Klosshufvud]

The technology absolutely exists, but not at the intended budget limitations Valve has for the device.



The 4nm Phoenix Point APUs outperform Deck pretty hard at 15W and above. Those devices running at 20W can in more recent games outperform Deck by near 100%. And 20W is absolutely feasible on devices with good cooling and 50Wh+ batteries. But this would inflate SD price by several hundred dollars for sure.

 

[Zathalus]

The technology absolutely exists, but not at the intended budget limitations Valve has for the device.



The 4nm Phoenix Point APUs outperform Deck pretty hard at 15W and above. Those devices running at 20W can in more recent games outperform Deck by near 100%. And 20W is absolutely feasible on devices with good cooling and 50Wh+ batteries. But this would inflate SD price by several hundred dollars for sure.

A 10-30% jump in power at the same wattage is not a generational leap that Valve is looking for. Sure, pumping up the TDP and putting in a bigger battery is possible, the Steam OLED actually has a bigger battery but Valve has opted to use that capacity to improve battery life. Judging by the wording of everything Valve has said so far they want to have double the performance at the same power envelope at the minimum, likely with a increase to 32GB of memory as well. That is going to require 3nm and newer version of Zen and RDNA at least.

 

[Thick Thighs Save Lives]

I don't know about steam deck 2, but I can tell you steam deck 3 is never happening. Ever.

This meme reminds me of this Gabe/Valve song.



Also this one where he memes around the number 3 - 'you've killed more than 2 people and less than 4'

 

[Klosshufvud]

A 10-30% jump in power at the same wattage is not a generational leap that Valve is looking for. Sure, pumping up the TDP and putting in a bigger battery is possible, the Steam OLED actually has a bigger battery but Valve has opted to use that capacity to improve battery life. Judging by the wording of everything Valve has said so far they want to have double the performance at the same power envelope at the minimum, likely with a increase to 32GB of memory as well. That is going to require 3nm and newer version of Zen and RDNA at least.

Sure double the perf at the same TDP is a few years off. But since the 7840U does so well at 20W, one may circumvent this long wait by achieving that performance target via a larger battery instead. IMO it is an inevitability as these APUs come with more CPU cores and CUs that power requirements increase aswell. Just the nature of things. We've seen that development in desktop components aswell. Power requirements gradually increase over time.

 

[Lumpyoatmeal]

“Half-Life maker has several games targeting current hardware in pipeline”​

 

[YOU PC BRO?!]

Valve are probably right. If we consider a true next-gen APU; it would likely be a 2nm/3nm part featuring further architectural improvements. The key requirement being a meaningful performance improvement at the same power envelope as the Deck.

It’s funny as ARM would be a much better fit for a portable device but just wouldn’t have the compatibility to play existing games.

 

[FingerBang]

The technology absolutely exists, but not at the intended budget limitations Valve has for the device.



The 4nm Phoenix Point APUs outperform Deck pretty hard at 15W and above. Those devices running at 20W can in more recent games outperform Deck by near 100%. And 20W is absolutely feasible on devices with good cooling and 50Wh+ batteries. But this would inflate SD price by several hundred dollars for sure.

I have both. The jump in performance is absolutely not worth justifying the release of a new product. You get a device that is, yes 30%-ish more powerful, at twice the power consumption. We should expect twice the power for around the same TDP to consider it a generational improvement.

AMD seems to have some insane APU coming in the next couple of years. I can see the Steam Deck 2 being released at around the same time. Probably 2025. The power race is slowing down after all

 

[WitchHunter]

You can see this with the "competitors", they really are at best a half-step above the Deck with problems like even worse battery life. I think they made the right decision by refining the existing one and waiting for a 2.0 until they can really blow it away.

They also have all the usage data and surely know that better battery life and screen is more useful than some increases to performance around the margins.

They let everyone copy the thing, then introduce deck 2.0 which will be a full fledged pc with a sticker on the side saying: Steam Deck 2.0.

 

[Silver Wattle]

I have both. The jump in performance is absolutely not worth justifying the release of a new product. You get a device that is, yes 30%-ish more powerful, at twice the power consumption. We should expect twice the power for around the same TDP to consider it a generational improvement.

AMD seems to have some insane APU coming in the next couple of years. I can see the Steam Deck 2 being released at around the same time. Probably 2025. The power race is slowing down after all

I can see a 6 core Zen5c+14-16CU RDNA4 GPU @3nm with faster memory being the target specs.


Personally I think as long as they double performance it will be enough and just invest any extra budget on the screen.

 

[Klosshufvud]

I have both. The jump in performance is absolutely not worth justifying the release of a new product. You get a device that is, yes 30%-ish more powerful, at twice the power consumption. We should expect twice the power for around the same TDP to consider it a generational improvement.

AMD seems to have some insane APU coming in the next couple of years. I can see the Steam Deck 2 being released at around the same time. Probably 2025. The power race is slowing down after all

I don't see how going from 15W to 20W is twice the power consumption. And some of the benchmarks in Phawx's video had perf gains in the range of 70-100%. Which is pretty damn important if that difference is between 20 fps or 40 fps. Just compare Baldur's Gate 3 or Returnal from SD to the Phoenix Point handhelds.

Regardless, my theory is that Van Gogh will remain the low watt king even in the future as future APUs come loaded with more features and cores and CUs and whatnot. The general trend for x64 components is that power consumption increases, despite the shrinking of nodes. I absolutely do believe 15-20W will be the standard power envelope even for handhelds in the APUs to come. Unless SD2 also opts for a stripped down CPU, which I doubt since that would hurt high-end performance.

 

[64bitmodels]

Does anybody just not care about the second subtitle?

"Half Life maker has several games targeting current hardware in pipeline"

Now i'm not saying we're getting Half-Life 3, but more IP from Valve would be super cool. Specially ones designed around the Steam Deck.

 

[Kenpachii]

It needs a massive jump not a small one for the steam deck 2 to be interesting in my opinion or else just stick with the steam deck 1 for now.

I would like to see in the next steam deck ~3x performance increase + dlss + framegen. That would give the steam deck 2 a massive boost in performance. the 50-100% more performance chips aren't interesting.

 

[Bernoulli]

The technology exists but it's expensive, valve didn't want to say technology really cheap that gives them good margins doesn't exists
And software upscaling plays a big part in it now

 

[64bitmodels]

It needs a massive jump not a small one for the steam deck 2 to be interesting in my opinion or else just stick with the steam deck 1 for now.

I would like to see in the next steam deck ~3x performance increase + dlss + framegen. That would give the steam deck 2 a massive boost in performance. the 50-100% more performance chips aren't interesting.

Nvidia helping on the next model would be great, but that'd require them to work with AMD or Intel since Nvidia has no proper x86 experience. Maybe in the future when AMD gets their upscaling together with FSR3.1 or whatever it might be a good substitute.

 

[Bernoulli]

It needs a massive jump not a small one for the steam deck 2 to be interesting in my opinion or else just stick with the steam deck 1 for now.

I would like to see in the next steam deck ~3x performance increase + dlss + framegen. That would give the steam deck 2 a massive boost in performance. the 50-100% more performance chips aren't interesting.

You are never getting that jump with the same price, the tech is there but costs too much
Unless you wait an entire console cycle

 

[hinch7]

Probably have to wait until 2nm and a lot of advancements in IPC in both CPU and GPU. Plus bigger jumps in rasterization and RT performance that works in low TDP chips (sub 25W). And better/more mature Ai, framegen and upscaling technologies.

Maybe by the time Zen 6 is out, plus RDNA 5 and we have way faster LPDDR RAM (LPDDR6) it should be a viable upgrade over what we have right now. The CPU upgrade even moving to Zen 2 to 4 would be a huge upgrade but if we're several years out they might as well wait for a good package from AMD, APU wise.

 

[REDRZA MWS]

Nintendo like everyone else needs to obey the laws of thermodynamics and find a balance between price/power/battery life. If it's a handheld device.

Nintendo is the one company that hasn’t considered “power” since GameCube.

 

[Minsc]

I don't see how going from 15W to 20W is twice the power consumption. And some of the benchmarks in Phawx's video had perf gains in the range of 70-100%. Which is pretty damn important if that difference is between 20 fps or 40 fps. Just compare Baldur's Gate 3 or Returnal from SD to the Phoenix Point handhelds.

Regardless, my theory is that Van Gogh will remain the low watt king even in the future as future APUs come loaded with more features and cores and CUs and whatnot. The general trend for x64 components is that power consumption increases, despite the shrinking of nodes. I absolutely do believe 15-20W will be the standard power envelope even for handhelds in the APUs to come. Unless SD2 also opts for a stripped down CPU, which I doubt since that would hurt high-end performance.

I think a lot of people who regularly use the Deck as a portable - even the site SteamDeckHQ - use the device at 5-10W, particularly closer to 5W if possible to get that upper end 8-12 hour battery life. That's something that just is not possible on the other handhelds, and I think why people prefer the Deck as at the moment there's no better choice for that 5W performance / battery life.

That the OLED uses less power and can run at a lower brightness than the LCD (2nit OLED vs 5nit LCD), all makes that even more in the favor of getting more out of that low TPD profile.

 

[Minsc]

To expand a little further - looking at that SteamDeckHQ site, and a very recent game that has a strong appeal for playing portable - Star Ocean The Second Story R you can see from that review that the game runs great - 50fps at 8 TDP. And that's on the LCD model. The OLED should be able to hit 6 or 7 TPD I'd think, with the slight performance boost and ability to drop to 45fps doubled to the native 90. And it'll look even better on OLED of course.

And I feel like there's 4-5 games like that a week sometimes. Brand new, incredibly fun games, that run perfectly on the Deck. And at low (under 10) TDP too.

 

[BootsLoader]

The technology absolutely exists but the problem will be the price. They will have to release a handheld with base price of 700$ or 800$ which will be very risky.
If you look at Apple, they have the M Series processors which their iPads have, very capable machines but very expensive also.

 

[Buggy Loop]

How does this factor in with the Switch 2 rumors?

Not much as Deck is stuck on x86 principles to keep PC compatibility and everyone but AMD has moved away from that for mobile chips.

What is their criteria for a big leap, who knows. The likes of Snapdragon XR2 gen 2 are in the ~2.4TFlops range with ML and RT, but not some massive leap above Deck’s 1.6TFLops

What Nvidia can do with the likely 2~2.5 TFlops Switch 2 is to leverage their ML to make it punch way above its weight.

Also Switch 2 is rumoured to have an I/O manager for streaming, Deck could really have benefited from an update like this, it would make Ratchet and Clank rift apart way more close to the original vision of rapid loading worlds.

 

[Deerock71]

I'm amazed at how much an OLED screen transformed my 70/30 mix of handheld/docked on my Switch to 99/1. I imagine the same should hold true for an OLED Steam Deck.

 

[Klosshufvud]

To expand a little further - looking at that SteamDeckHQ site, and a very recent game that has a strong appeal for playing portable - Star Ocean The Second Story R you can see from that review that the game runs great - 50fps at 8 TDP. And that's on the LCD model. The OLED should be able to hit 6 or 7 TPD I'd think, with the slight performance boost and ability to drop to 45fps doubled to the native 90. And it'll look even better on OLED of course.

And I feel like there's 4-5 games like that a week sometimes. Brand new, incredibly fun games, that run perfectly on the Deck. And at low (under 10) TDP too.

Star Ocean is an extremely light weight game. I don't see any reason why it can't run at that kind of performance on a Phoenix Point APU. I can't find any videos on it though. My personal experience is that emulation including 6th gen on these newer APUs is around 5-10W. Never above that. I don't understand this notion that those devices are somehow unable to perform at lower watts. There are games that SD indeed runs better at the low end power consumption, but most of the time, these games work just fine even on the mentioned APUs. Especially with drivers being way better than they were when benchmarks initially dropped. But some of it can't by bypassed by the nature of the Phoenix Point high-end CPU.

 

[SolidQ]

I can see a 6 core Zen5c+14-16CU RDNA4 GPU @3nm with faster memory being the target specs.

That currently line up from AMD

 

[Minsc]

Star Ocean is an extremely light weight game. I don't see any reason why it can't run at that kind of performance on a Phoenix Point APU. I can't find any videos on it though. My personal experience is that emulation including 6th gen on these newer APUs is around 5-10W. Never above that. I don't understand this notion that those devices are somehow unable to perform at lower watts. There are games that SD indeed runs better at the low end power consumption, but most of the time, these games work just fine even on the mentioned APUs. Especially with drivers being way better than they were when benchmarks initially dropped. But some of it can't by bypassed by the nature of the Phoenix Point high-end CPU.

That's true - I guess maybe the point is that the Deck is not surpassed in that lower end of power consumption. It's not like the Phoneix Point APU is able to run the newer, demanding games at the same lower power point.

I think that a Deck 2 would need to see an improvement there, not at the upper end. Having BG3 run at 60fps at the lower power point would certainly be a feat. Maybe that's asking too much, but that's the area we need to have a concrete improvement in for a Deck 2 to be realized.

 

[64bitmodels]

Sure double the perf at the same TDP is a few years off. But since the 7840U does so well at 20W, one may circumvent this long wait by achieving that performance target via a larger battery instead. IMO it is an inevitability as these APUs come with more CPU cores and CUs that power requirements increase aswell. Just the nature of things. We've seen that development in desktop components aswell. Power requirements gradually increase over time.

We can circumvent it or we can just wait for AMD to eventually get to insane performance at lower wattage... It's their specialty and what makes them superior to intel in the X86 CPU space.

15 watts I think is a good limit because it shows the actual performance increase better than anything on the phoenix powered systems- computers will naturally perform better when they have more power to work with. Not to mention that it's sort of the maximum wattage for these lithum ion batteries to get decent performance.

 

[Klosshufvud]

Well well well. Not gonna say I told ya so but...



Great Phawx video. Apparently Steam Deck LCD reports faulty TDP. It often breaks the power budget. At 5W TDP, OLED is performing worse than LCD. Which of course makes no sense since the former is on more efficient process node. But Phawx points out that 7W OLED equals 5W LCD in SoC TDP. And he also says the Phoenix APU is less prone to break TDP budget as Van Gogh is.

So there you have it. There is no secret sauce as to why SD is overperforming at extremely low TDPs. It's simply not respecting the power budget limitation.