Status
Not open for further replies.

Tmespe

Member
Oct 27, 2017
2,569
Obviously, the Hauppauge works to a certain extent with Linux, so the test was partially successful, but I'm confused as to why the picture was all messed up. The laptop is a Toshiba Satellite Pro and is around 10 years old, maybe a bit more. The GPU is an AMD HD 6450M but I'm not sure that will have anything to do with it. The next test would be to rebuild the Alpha with Arch Linux to see if the hardware which works with Windows works with Linux - but I can't do that until much later in the year when the F1 season, which I'm recording, is over.
Creds to you for managing to install and manage Arch without any prior experience with Linux! If you want to use an Arch based distro that's more user friendly and easier to manage, you should try Manjaro. But as another poster said Ubuntu is a more sensible choice for a newcomer.
 

karnage10

Member
Oct 27, 2017
5,622
Portugal
Everyone is comparing steam deck to Nintendo and here I am just happy that I can play crusader kings 3 on the go.

Anno 1800 is probably the most AAA strategy around at the moment though 😂 It's a beautiful game and rather demanding to be run at highest settings.
I'd actually say that total war is the biggest budget strategy game around. Graphicalwise, IMO, it is by far the most impressive strategy game i ever seen.
It also eats CPUs for meals.
 

shodgson8

Shinra Employee
Member
Aug 22, 2018
4,379

th1nk

Member
Nov 6, 2017
6,568
Man, I hope an OLED version will be offered soon. The LCD screen is a dealbreaker for me, but everything else about it is awesome. I love that technology is finally at this stage where high-end games can be played in bed. 🤩
 

Mr.Deadshot

Member
Oct 27, 2017
20,285
Everyone is comparing steam deck to Nintendo and here I am just happy that I can play crusader kings 3 on the go.


I'd actually say that total war is the biggest budget strategy game around. Graphicalwise, IMO, it is by far the most impressive strategy game i ever seen.
It also eats CPUs for meals.
IMHO Anno 1800 looks better than Total War, but I agree. Total War is also AAA and pretty impressive.
 

SolarLune

Member
Jun 22, 2021
935
How long does that usually take ? I figured that would be something Valve already did before letting the influencers test the device.

It can take different amounts of time (ranging from a few seconds to a minute or two) depending on how many (and how complex) the shaders are. Shader compilation for Proton is usually done on first run. If you choose to skip compilation, the game would have to compile the shaders during gameplay, which would account for those microstutters.

I feel like I heard on a video that the testing consoles were "freshly installed", which would explain the need for shader compilation, though I might be misremembering. Looking over the footage now, I don't really see the very start of any games through Steam OS (which is where it would indicate it's compiling the shaders) except for Hades in GiantBomb's video. This could just be because they started it after a previous outlet used the same console and started up Hades for its first run previously.
 

bmfrosty

Member
Oct 27, 2017
1,896
SF Bay Area
I'd love something similar to the Switch Pro Controller personnally.

51dxc1F59zL._AC_SX679_.jpg
Love this controller and the Xbox ones. Wish that button labelling were uniform across all of it along with the ok/cancel button positions.

Should have been:
Code:
 Y
B X
 A
 

Fadewise

Member
Nov 5, 2017
3,210
Love this controller and the Xbox ones. Wish that button labelling were uniform across all of it along with the ok/cancel button positions.

Should have been:
Code:
 Y
B X
 A
Nintendo can go to hell with their nonsense backwards button labeling and confirm/cancel rules. I don't care if they "did it first", it's still inferior.
 

hanshen

Member
Jun 24, 2018
3,946
Chicago, IL
Might have already been posted but I hadn't seen this before regarding the M.2 slot and how user upgradable the memory actually is (sounds beyond my own comfort level to be honest, I know how packed these portable devices can be).




Plagman works at Valve.



It still depends. I removed the shielding in the Nintendo Switch to replace the joycon rail before, it's honestly not that hard to do. As long as they don't glue things down and the m.2 slot is on the back side not the screen side, I don't really see how much harder it can be.
 

shodgson8

Shinra Employee
Member
Aug 22, 2018
4,379
It still depends. I removed the shielding in the Nintendo Switch to replace the joycon rail before, it's honestly not that hard to do. As long as they don't glue things down and the m.2 slot is on the back side not the screen side, I don't really see how much harder it can be.

Fair enough! It could be actually straightforward and definitely doable for some (though I agree we don't really know the extent to which the device could be glued or otherwise crammed in). I have done some laptop screen repair & replacements along with general PC building and I would be twitchy if I had gone for the 64GB with plans to upgrade the memory myself but I have never really messed with portables or smartphones.

I think it is just worth mentioning that installing more storage outside of the SD card slot has potential to be challenging.
 

Mivey

Member
Oct 25, 2017
18,169
It's an AMD chip so no.

Unless they add some daughter board for Thunderbolt (which I highly doubt) it will not be in the Steam Deck so no eGPU support.
There are Intel certified motherboards with Thunderbolt ports out there for AMD CPUs, so it's not something out of the realm of possibilities. It's an open standard that doesn't need the CPU to support it.
But, I agree it's about as likely to be supported on the Deck as the chance of people stopping the Switch comparisons.
 

bmfrosty

Member
Oct 27, 2017
1,896
SF Bay Area
Nintendo can go to hell with their nonsense backwards button labeling and confirm/cancel rules. I don't care if they "did it first", it's still inferior.
I agree they got it wrong, but so did Sony and Microsoft.

Nintendo went through many prototypes for the SNES controller. This a mock-up of one of the better ones:

Screen-Shot-2021-08-08-at-3-31-41-PM.png


Frankly, if you compare the wireline protocol of the NES controller and compare it to the SNES controller, you'll see that the button labels match up to the positions in the protocol. Also, on the NES Super Mario games A was jump and B was dash, but on the SNES, B was jump and X was dash.

NES Protocol:
wiki.nesdev.com

Standard controller

All NES units come with at least one standard controller - without it, you wouldn't be able to play any games!

SNES Protocol (see section 6):

Super Nintendo - Pinouts & Protocol FAQ - Super Nintendo - By JChristy - GameFAQs

For Super Nintendo on the Super Nintendo, Pinouts & Protocol FAQ by JChristy.

Another interesting bit of note is that in Japan, until the last couple of years, ok/cancel matched between Sony and Nintendo. It was SCEA that insisted on flipping the two. Since the western market has become more important to Sony than the Japanese market, when they finally decided to standardize, they followed the western layout.

Yes, I have thought way too much about this. Yes there is nothing I can do about it. If I had a time machine and any say in the matter, the US would be on the metric system, we'd all be using the World Calendar, and video game controllers would follow the standard that I laid out above. Additionally, Nintendo would have sold NES to Game Port adapters, SNES to RS232 adapters, and then the N64 Joystick (which would have been closer to the Playstation joystick) would have either been an early USB device, or they would have eventually sold N64 to USB adapters.

Yes, I'm way too far down the rabbit hole.

Also the Genesis controller would have been C B A.
 

Nzyme32

Member
Oct 28, 2017
5,325
kotaku.com

The Steam Deck Might Struggle To Play Some Games On Big TVs

Valve has confirmed the portable device won't get a power boost when docked and plugged into a external monitor

That's a clickbait headline.
The otherway of viewing it, is that Deck has unrestricted performance when undocked, becuase the user sets whatever limitations they want rather than the vendor capping performance "unless docked" (which is still a capped performance that can be overriden on Deck because it's a PC).
On a "4k TV" the Switch as is, outputs at 1080p, but internally is rendering per whatever profile the dev has decided upon for their "docked" mode up to the confines of the systems limitations. This is not how Deck operates, as a PC. Perhaps not the brightest idea to compare with Switch on this topic.

Regardless of what a user chooses to do when docked or on the go, they'll get to define the positives and negatives and how battery life is used up by setting performance as they want, even with whatever upscalers they may want, or going so far as to push the systems clocks and thermal limits at their risk.

I don't think I'd call either approach good or bad, though I certainly prefer the freedom with Deck
 

TetraGenesis

Member
Oct 27, 2017
3,169
Valve should send them an eGPU with their Steam Deck and a "here's your docked power boost for your 8K TV" note attached to it.

I know you're joking but since it's not Thunderbolt or USB4, Steam Deck won't work with eGPUs. Which — as excited as I am for steam deck (extremely!) — does feel like a bit of a missed opportunity. But I mean, if there is anything more niche then using a handheld Linux machine as your primary computer, it's plugging that machine into a several hundred dollar eGPU dock hahahaha (I say this as someone who used an eGPU with my laptop for years and loved it.)

I have a good PC now and don't need it, but eGPU docks specially made for the Steam Deck would realize the internet's long running fantasy of a Switch dock with a discrete GPU hahahaha
 
Oct 25, 2017
6,086
Nintendo's button layout on the Pro Controller is the main reason it's not my primary PC controller. Will never get used to having to swap AX and BY. Fortunately the Xbox Wireless controller is the next best thing. I prefer symmetrical sticks like the DS4/5 but those have their own headaches and don't feel as good in the hand as the former two controllers.
 

Aiii

何これ
Member
Oct 24, 2017
8,486
I know you're joking but since it's not Thunderbolt or USB4, Steam Deck won't work with eGPUs. Which — as excited as I am for steam deck (extremely!) — does feel like a bit of a missed opportunity. But I mean, if there is anything more niche then using a handheld Linux machine as your primary computer, it's plugging that machine into a several hundred dollar eGPU dock hahahaha (I say this as someone who used an eGPU with my laptop for years and loved it.)

I have a good PC now and don't need it, but eGPU docks specially made for the Steam Deck would realize the internet's long running fantasy of a Switch dock with a discrete GPU hahahaha
it is still (mostly) Intel tech, after all. I am sure there is a way to do it (I think ASUS has announced an AMD MB with Thunderbolt 4), but I am not sure how feasable it would be in such a small form factor, as I imagine it would mean an extra controller chip somewhere at the very least.
 

Genio88

Banned
Jun 4, 2018
964
Might have already been posted but I hadn't seen this before regarding the M.2 slot and how user upgradable the memory actually is (sounds beyond my own comfort level to be honest, I know how packed these portable devices can be).




Plagman works at Valve.


I ordered the 64GB version for Q1 22, if it'll be really that hard to access the SSD slot i'll just stick with my GPD Win 3, which come with a 1TB m2 ssd and is easy to replace supporting even 2TB gen 4 ssd. And i'll wait a new, hopefully smaller, Deck revision in the future, i know there is the micro sd option, but seriously micro sd loading times are atrocious in 2021/22 and even console now have fast ssd with PC getting direct storage support soon
 

Alexandros

Member
Oct 26, 2017
18,111
I've been thinking that if we use the Doom Eternal comparison that Linus did as a baseline, maybe we could get a rough approximation of how Steam Deck will perform by looking at Aya Neo tests? I watched this video:



For example, since Borderlands 3 averages about 40something FPS on 800p Low, wouldn't it make (some) sense to assume that on Steam Deck you could get to 60 fps at 800p Low or 30 fps at 800p medium/high? Obviously I'm talking about very rough estimates.
 

Rikimaru

Member
Nov 2, 2017
853
I tried to warn people in the Linux thread: just use Ubuntu 20.04. That's what Valve recommends to developers, not Arch. Whatever Arch config you manage to get working is probably still pretty different from what Valve ships as SteamOS.

Ubuntu is dramatically more usable, with a fully GUI installer from a live USB stick.
*Kubuntu. Well you can install KDE on Ubuntu, but it is easier just to install Kubuntu.
 

GhostTrick

Member
Oct 25, 2017
11,559
I've been thinking that if we use the Doom Eternal comparison that Linus did as a baseline, maybe we could get a rough approximation of how Steam Deck will perform by looking at Aya Neo tests? I watched this video:



For example, since Borderlands 3 averages about 40something FPS on 800p Low, wouldn't it make (some) sense to assume that on Steam Deck you could get to 60 fps at 800p Low or 30 fps at 800p medium/high? Obviously I'm talking about very rough estimates.



Yes. But it all depends of the TDP and the limiting factor. On a GPU limited scenario, based on Doom, Deck performs between 50 to 60% better than the Neo at a lower TDP (15W vs 17W).
 

Dreamwriter

Member
Oct 27, 2017
7,461
They had a choice and went for the 'glass half empty' headline for the clicks.
Worse, they went for the "misleading" headline for the clicks. The size of the tv won't ever make any game struggle on the device. Choosing to run the game at a higher resolution could, even if you play on the handheld's 7 inch screen.
 

Dolce

Member
Oct 25, 2017
14,355
Nintendo can go to hell with their nonsense backwards button labeling and confirm/cancel rules. I don't care if they "did it first", it's still inferior.

It's less a Nintendo thing and more a regional thing. It was only recently that Sony stopped using Circle to confirm in places like Japan and there was quite a bit of annoyance that they changed it. Before that change, Sony had been using Circle to confirm since 1994 in Asia.

As someone who has always gone between both, they're essentially a preference thing. But on a personal level I can get used to either one and never actually notice when I'm playing a game.
 

BreakAtmo

Member
Nov 12, 2017
13,183
Australia
Yes. But it all depends of the TDP and the limiting factor. On a GPU limited scenario, based on Doom, Deck performs between 50 to 60% better than the Neo at a lower TDP (15W vs 17W).

Am I the only one who finds this to be a bit disappointing? I mean, the Deck has 60% more CUs, a more efficient architecture that should be hitting higher clocks and more RAM bandwidth to avoid bottlenecks. I was actually expecting more, mostly because I was assuming devices like the Aya and the GPDs had bandwidth bottlenecks.
 

cw_sasuke

Member
Oct 27, 2017
26,811
Am I the only one who finds this to be a bit disappointing? I mean, the Deck has 60% more CUs, a more efficient architecture that should be hitting higher clocks and more RAM bandwidth to avoid bottlenecks. I was actually expecting more, mostly because I was assuming devices like the Aya and the GPDs had bandwidth bottlenecks.
Devices like Aya Neo also cost quite a bit more. There is only so much better hardware you can deliver in that fork factor.
 

bytesized

Banned
Oct 27, 2017
5,882
Amsterdam
I've been thinking that if we use the Doom Eternal comparison that Linus did as a baseline, maybe we could get a rough approximation of how Steam Deck will perform by looking at Aya Neo tests? I watched this video:



For example, since Borderlands 3 averages about 40something FPS on 800p Low, wouldn't it make (some) sense to assume that on Steam Deck you could get to 60 fps at 800p Low or 30 fps at 800p medium/high? Obviously I'm talking about very rough estimates.

I wonder if flight simulator would run better nowadays after the performance improvements it got.
 

Bede-x

Member
Oct 25, 2017
10,133
In this video AYA Neo's CPU package power is at 15W so hopefully the results will be roughly comparable.

It's seems fair to expect, it'll be an upgraded Neo. I'm mostly interested in knowing how many games will have stutter/hitching? I noticed PC Gamer mentioning it in Death Stranding, in the Tested video Jedi Fallen Order clearly has it in the video near the beginning, while Prey exhibits the same in the Games Radar preview. It was also mentioned being in Dead Cells in one of the videos, but don't remember which one.

Would also like to know it runs Earth Defense Force 4.1/5, as that'll be one of my main use cases. Those could push PS4 to 0fps with the right combination of enemies and weapons, but I'd still be very happy and impressed if the Deck could approach similar performance as PS4 in general play.

Hope we get a lot more testing soon. They only just released more previews and I'm already hungering for more, especially a longer DF testing video.
 
Last edited:

eonden

Member
Oct 25, 2017
17,226
Am I the only one who finds this to be a bit disappointing? I mean, the Deck has 60% more CUs, a more efficient architecture that should be hitting higher clocks and more RAM bandwidth to avoid bottlenecks. I was actually expecting more, mostly because I was assuming devices like the Aya and the GPDs had bandwidth bottlenecks.
You cannot hit higher clocks without increasing the power consuption. The Aya is at 18W max while the Deck is at 15W max, and even then, Deck seems to be 40-50% stronger performacne.
Bandwidth bottlenecks is very dependant on games and I dont think Doom Eternal really has them. Perhaps the performance in open world games will be different.

Devices like Aya Neo also cost quite a bit more. There is only so much better hardware you can deliver in that fork factor.
The Aya Neo is also 40-50% weaker than the Deck!
 

GhostTrick

Member
Oct 25, 2017
11,559
Am I the only one who finds this to be a bit disappointing? I mean, the Deck has 60% more CUs, a more efficient architecture that should be hitting higher clocks and more RAM bandwidth to avoid bottlenecks. I was actually expecting more, mostly because I was assuming devices like the Aya and the GPDs had bandwidth bottlenecks.


That's a very least scenario, at a lower TDP. Also, Deck has 33% more CUs, not 60% (6 vs 8). It also has a 7% uplift in clockspeed. (1500mhz vs 1600mhz). 50-60% is a minimum. You can expect better results in some cases.
 

Alexandros

Member
Oct 26, 2017
18,111
Am I the only one who finds this to be a bit disappointing? I mean, the Deck has 60% more CUs, a more efficient architecture that should be hitting higher clocks and more RAM bandwidth to avoid bottlenecks. I was actually expecting more, mostly because I was assuming devices like the Aya and the GPDs had bandwidth bottlenecks.

It might be more or less depending on the game. The difference in performance between Aya Neo and Steam Deck was estimated in the 1.6 to 1.8 range before the reveal and Steam Deck outperformed it by 60% so the difference fell within that range. It's also possible that Proton has a small performance penalty or the OS isn't yet fully optimized. I expect more demanding games to fall closer to the 1.8 side.

I wonder if flight simulator would run better nowadays after the performance improvements it got.

It should, hopefully someone will eventually test it since it's one of the most demanding games out there.

Hope we get a lot more testing soon. They only just released more previews and I'm already hungering for more, especially a longer DF testing video.

Tell me about it. My policy regarding new hardware is always "wait for benchmarks" and I really, really want to see how Steam Deck fares in real-world tests. I was already very interested in a PC portable but I didn't pull the trigger on previous models because they couldn't really achieve solid performance on medium settings in PS4-era games. I have a big Steam backlog and I don't want to resort to ultra low settings or low resolutions. In theory Steam Deck seems up to the task.
 

Bede-x

Member
Oct 25, 2017
10,133
Tell me about it. My policy regarding new hardware is always "wait for benchmarks" and I really, really want to see how Steam Deck fares in real-world tests. I was already very interested in a PC portable but I didn't pull the trigger on previous models because they couldn't really achieve solid performance on medium settings in PS4-era games. I have a big Steam backlog and I don't want to resort to ultra low settings or low resolutions. In theory Steam Deck seems up to the task.

Previous PC handhelds have felt too expensive for the performance they offer. Steam Deck seems to nail the price-to-performance ratio in a way others haven't. Personally, I'm fine with low settings and don't care much for resolution, especially on a small screen, but I do want to see how it performs. Stutters/hitching would annoy me to no end. I'm getting it regardless - there's just too much appeal in having handheld access to all those Steam games I already own or would buy anyways - but could we please have DF deep dive soon, Valve :)
 
Last edited:

Issen

Member
Nov 12, 2017
6,933
Previous PC handhelds have felt too expensive for the performance they offer. Steam Deck seems to nail the price-to-performance ratio in a way others haven't. Personally, I'm fine with low settings and don't care much for resolution, especially on a small screen, but I do want to see how it performs. Stutters/hitching would annoy me to no end. I'm getting it regardless - there's just too much appeal in having handheld access to all those Steam games I already own or would buy anyways - but could we please have DF deep dive soon, Valve :)
In the absolute worst scenario, it's going to be an emulation handheld that also plays all those emulators that aren't on Android, and then plays all of your old PC games and new indies on top of that. Honestly, I'd say that's a pretty darn good worst case. I don't think a scenario exists where the machine is not worth buying at all.
 

Xadra

Prophet of Truth - One Winged Slayer
Member
Oct 26, 2018
2,030
Not exactly new, the source is the PC Gamer interview published 3 days ago.



(Sorry if was already discussed)
 

BreakAtmo

Member
Nov 12, 2017
13,183
Australia
Devices like Aya Neo also cost quite a bit more. There is only so much better hardware you can deliver in that fork factor.

The Aya is more expensive because it's made in smaller quantities and has to be sold at a substantial profit vs the Deck that can be sold roughly at cost because it drives Steam purchases.

You cannot hit higher clocks without increasing the power consuption. The Aya is at 18W max while the Deck is at 15W max, and even then, Deck seems to be 40-50% stronger performacne.
Bandwidth bottlenecks is very dependant on games and I dont think Doom Eternal really has them. Perhaps the performance in open world games will be different.

I was actually under the impression that RDNA2 can in fact hit higher clocks than Vega at the same power consumption due to more efficient design, but we'll see. I'm also interested to see how the bandwidth differences show in different games.

That's a very least scenario, at a lower TDP. Also, Deck has 33% more CUs, not 60% (6 vs 8). It also has a 7% uplift in clockspeed. (1500mhz vs 1600mhz). 50-60% is a minimum. You can expect better results in some cases.

Thanks for this - I thought that the 4500U had 5 Vega CUs, not 6. That pretty much clears up any issues I had. And we might even be able to undervolt the Deck to hit the max clocks more regularly.

One other question in regards to the two systems - which Zen2 CPU configuration do you think is better for games? 6/6 or 4/8?
 
Status
Not open for further replies.