New Steam hardware inbound

[Dragon1-1]

25 minutes ago, kksnowbear said:

It's not personal, it's an informed and qualified professional observation.

Just because you fancy yourself a professional doesn't mean ever personal opinion of yours is a professional market analysis. In fact, Valve has actual people who are paid not to build PCs in a brick and mortar store, but to look at market trends, surveys and telemetry, then forecast what kind of products people the company can make that customers will want to buy. That is where you go for an actual professional opinion. Yours is merely a personal opinion of an industry professional.

My opinion is informed and qualified, too. I'm a biophysicist, so a professional scientist, which includes a good understanding of thermodynamics. A computer radiator is a fairly trivial system compared to what goes on in a living cell. In a personal opinion of this scientific professional, there's nothing preventing this little toy box from being a reliable means to play games on a TV. Not power draw, not TDP, not heat. If it has issues with any of these, it'll be because someone screwed the pooch somewhere, and Valve engineers have so far been good at not doing that. As for performance, we'll see, but the numbers Valve give are sufficient for most gaming an average person does.

You act like your experience means your opinions somehow hold more weight, despite all arguments to the contrary. In fact, you've shown pretty conclusively that you're ready to dismiss everything that doesn't fit into your experiences with an extremely limited customer base. Again, Valve doesn't care about you or any people that you support. It cares for the 75 million that bought a PS5, and might, for their next console, pick a Steam Machine instead.

It'll be easier for everyone if you just accept you're wrong and get on with your life. The universe doesn't revolve around you or your specific use cases. That you can't imagine all the other reasons people buy gaming laptops is a prime example of that.

Edited November 16 by Dragon1-1

[LucShep]

3 hours ago, kksnowbear said:

Anyhow, have a great night fellas.  I have better stuff to do.  Buy all the underpowered miniPCs ya want.  Tell others, too.

That way, when they all finally decide to step up to real gaming computers, there will be more clients for builders like me

Sorry, but couldn't help commenting this one - the whole “real gaming computer”  label carries a lot of baggage, and not the good kind.
It groups people, it gatekeeps, and it creates the exact kind of negativity that has always pushed -and still pushes- a lot of potential players away from the hobby.
This is the part that clearly - to me anyway - you clearly haven't understood.

A lot of people are excited for this Steam Machine precisely because they’re not interested in “hardcore gaming” or chasing the highest-end hardware.
Precisely because they can't relate and hate that "grouping" and "elistism" aspect of the hobby. It's more than just the money aspect (as relevant as that is).
They want something small, simple, quiet, affordable, supported, and capable of running the games they actually enjoy - without the ultra-high expectations that enthusiasts (like in these forums) aim for.

Whether it’s for gaming in the living room, in a bedroom, at a friend’s place, with family, or during a party… let’s be honest: your $4,500+ tower isn’t exactly ideal for that, is it?
And honestly, that’s what kept LAN parties alive back in the day (good times - anyone remember those?) : portability and convenience, not 2-kilowatt monster rigs.

Nobody needs a $1,500+ GPU to have fun. Plenty of people buy them, sure - but spending that kind of money doesn’t make anyone more “legit” than someone happily playing on a modest box.
It just makes them an enthusiast - the same way someone with a friggin $20,000 carbon/titanium push bike isn’t “more of a cyclist” than someone enjoying a basic $200 one.

Valve understands this broader audience, and they clearly saw a huge gap - a real opportunity - that nobody else was addressing.
That’s exactly why a small, efficient SteamOS box makes sense for so many people (more or less depending on the final unknown price).

Different kinds of players, different needs - and there’s room for all of them.
And who knows, some of those folks might eventually jump deeper into the hobby and end up flying DCS... 

Edited November 16 by LucShep

[Aapje]

9 hours ago, kksnowbear said:

I didn't say it was "automatically" anything.

You actually did say that "certain things inherently contain weaknesses." And "The relevant point is that, in any event, the rules of heat/size apply...always have, always will.  Pesky physics."

Then when people address those statements, you come up with contradictory claims that frustrate the other people in the discussion and cause them to seek clarification or to make objections, but you just keep coming up with more inconsistent statements that means that the debate fails to move past the point. This kind of debating style is exactly why debates with you have a tendency to keep going in circles.

9 hours ago, kksnowbear said:

But it is, in my estimation, a design with unnecessary potential for reliability weaknesses.

From an engineering perspective, this is just total nonsense. A Raspberry Pi is very reliable despite being very small and having minimal cooling, because it doesn't use much power. Thermal reliability issues are about keeping components within safe thermal limits. This is not even just about cooling, but also keeping heat away from thermally sensitive components, even if that means that other components stay hotter.

If the thermal solution is adequate for heat that is being produced, then there is no reason why a relatively small device can't be reliable. Given that these devices can measure the heat, they can ramp up the fan or throttle down the chips when excessive heat is detected.

Note that full size desktops nearly always have a cooling solution that is very inefficient, because desktops also sacrifice optimal cooling. However, they sacrifice them for flexibility in components and room to do easy maintenance.

9 hours ago, kksnowbear said:

Millions of people buying underpowered "gaming" laptops doesn't prove they're a good idea.  It simply shows people will buy anything that is marketed properly, that's it.  They're good for traveling people who like to game, or similar situations.  Performance is sacrificed for the sake of portability. That's fine, just call it what it is.

It's not about what I don't like personally.  You consistently try to bring "personal" into a discussion.  It's not personal, it's an informed and qualified professional observation. 

So on the one hand you admit that there is a benefit to this trade-off, but then you argue that the only reason why people buy these things in large volume is due to deceptive marketing, rather than that many people simply prefer this trade-off.

This dismissal of choices that you don't seem to understand and/or like is exactly why people are saying that you judge things based on your personal preferences, and not professionally.

9 hours ago, kksnowbear said:

Is it possible for you to have a discussion without the personal attacks?

It is typical for discussions with you that you interpret even the mildest criticisms of your debating style as personal attacks, and this utter inability to take criticism is surely why you keep persisting in behavior that causes these out of control and unproductive debates. When those happen on this forum, you are typically the common factor. These bad debates appear to frustrate you greatly, so this inability to improve yourself seems to hurts you. So when I make these statements, I do so because I think that if you would take them to heart, it will be good for everyone, not in the least, you.

Edited November 16 by Aapje

[Dragon1-1]

5 hours ago, Aapje said:

Given that these devices can measure the heat, they can ramp up the fan or throttle down the chips when excessive heat is detected.

That is actually the whole problem at the heart of the issue. Nobody is worried that this thing will desolder its own chips from the boards, Apple III style. Valve would have to grab a massive idiot ball to do that. The concern is that we've got a console that's barely adequate to run games it's supposed to be running, and then it throttles. Steam Machine has fairly low end specs as far as gaming hardware goes, so any further reduction will not be good for the playing experience. It has to run reasonably cool when it's going full tilt specifically because it doesn't have much of a performance margin.

That's not to say it is certain to run into this limitation. Of course, it has to reject something like 150W of heat (30CPU+110GPU+I estimate about 10 for other stuff). It's not a tall order. 

5 hours ago, Aapje said:

Note that full size desktops nearly always have a cooling solution that is very inefficient, because desktops also sacrifice optimal cooling. However, they sacrifice them for flexibility in components and room to do easy maintenance.

This is not a good argument. A properly built gaming desktop will have good airflow, leading to a very efficient cooling of components. The room that's that gives you easy maintenance is also room for the air to flow through the case, and for top level stuff you'll likely have water cooling, which can shift a lot of heat with proper setup. This also spaces heat generating components out, meaning they transfer heat to the cooling medium instead of one another. Likewise, a desktop case can accept multiple 120mm or 140mm fans, which can achieve good airflow while being reasonably quiet and cheap. Trying to do the same with a single small fan is not possible, because it'd have to spin so fast the motor wouldn't be able to fit in the hub. For air cooling, you'd probably need some sort of high performance air pump, and those aren't exactly quiet. For water cooling, big case fits a big radiator, which has more surface area, making it more efficient at rejecting heat. Trying to replicate this with an SFF machine would probably result in the machine being 50% computer, 50% radiator, and/or being obscenely loud.

Desktops are great if you want peak performance for photorealistic VR, wall sized screens and/or absurdly high FPS. To pump 1kW into your electronics, as you will be doing with a top end desktop these days, you simply need a big case and a good cooling setup. That said, most people will never have a need to go that far. For me, VR in DCS is the only worthwhile reason for pushing performance. My father decided a little more than Full HD at 60FPS satisfies him, and thus he's happy with a 1080 and a CPU from its time (don't remember the model out of hand). He can play any game he wants with that, his rig is far from new, but it's sufficient for him. There will be people who'll be happy with Steam Machine, too.

[Aapje]

16 hours ago, Dragon1-1 said:

The concern is that we've got a console that's barely adequate to run games it's supposed to be running, and then it throttles.

Well, Dell sells 4050 and 5050 gaming laptops that are most likely slower than this. Valve claims 2.45GHz max sustained clock with 28 RDNA3 CU's, which would put it a bit below an RX 7600, so it might be around a 3060.

I don't know whether that max sustained clock requires you to keep it in a freezer, or whether that is achievable for normal temperatures with a decent airflow to the machine. Whether it throttles significantly for a somewhat reasonable setup is not something I can judge based on the specs, but it is not a given that it will.

16 hours ago, Dragon1-1 said:

This is not a good argument. A properly built gaming desktop will have good airflow, leading to a very efficient cooling of components.

Gaming desktops typically have air bypassing the radiators, pull exhaust air from one radiator into another radiator, pull exhaust air back into the case, have turbulence, have fans run too slow for optimal cooling because people don't like the noise, etc.

Can it be made good enough for high-wattage components? Sure. Is it efficient? No. Of course, making it more efficient has downsides, which is why I said that desktops sacrifice cooling efficiency to achieve other goals. The point of what I said is that the Steam machine can achieve better cooling efficiency by choosing a different trade-off.

Note that Corsair is about to release a triple-chamber case that solves one of the issues mentioned above (warmed up GPU air going into the CPU or vice versa). See this recent video by Tintin:

Edited November 17 by Aapje

[Dragon1-1]

9 hours ago, Aapje said:

Well, Dell sells 4050 and 5050 gaming laptops that are most likely slower than this. Valve claims 2.45GHz max sustained clock with 28 RDNA3 CU's, which would put it a bit below an RX 7600, so it might be around a 3060.

The "gaming" laptops using 50 series GPUs are low end and likely inadequate for most modern games. RX7600 and 3060 are right on the edge, which is why Steam Machine throttling would be a very bad thing. It's not a given that it will, its TDP is rather modest, but it is something to watch.

9 hours ago, Aapje said:

Gaming desktops typically have air bypassing the radiators, pull exhaust air from one radiator into another radiator, pull exhaust air back into the case, have turbulence, have fans run too slow for optimal cooling because people don't like the noise, etc.

They might do these things, depending on design, but they don't have to. My rig, for instance, has GPU airflow largely (not completely, my case lacking a place for a bottom exhaust fan) separated from the top-mounted CPU radiator. Serious watercooled setups run a custom loop over CPU and GPU (and for the craziest ones, the mobo as well), routing all heat to the radiator, keeping the whole setup from ever seriously heating up. On high end of air cooling, scimitar bladed fans are used to maximize airflow at the intake and exhaust. As for fan speeds, they only run slow when the computer doesn't need cooling. Fire up something demanding, and virtually every properly configured gaming rig will throttle them up to 100%. Noise cancelling headphones recommended.

What you describe are rookie mistakes sometimes seen on prebuilt rigs, but not on anything competently designed. Yes, there's a lot of desktops with sloppy airflow, but it's not inherent in the desktop form factor.