Tuesday, April 7th 2020
Steam Hardware Survey March 2020: Intel CPUs, NVIDIA Graphics Cards Rising
Steam has just released its hardware survey for March 2020, and it paints an interesting picture on the overall Steam user market. First of all, a disclaimer: Steam's survey's don't refer to the entirety of the systems that are running its digital distribution platform; it hails from only a part of the connected machines which are randomly selected - ideally, in such as way so as for its hardware data to be representative of the overall Steam ecosystem.
The March 2020 hardware survey shows Intel CPUs gaining traction over AMD's, albeit in a very limited fashion (a 0.24% increase in total number of machines running with GenuineIntel CPUs). Cross referencing this increase with the average clockspeeds of Intel CPUs, though, paints an interesting picture: the higher clockspeeds in the 3.3 to 3.8 GHz range (which can relatively cleanly be associated with the company's higher-performance products), saw a decline in usage (contrary to AMD, which saw the reverse happening).
The increase in Intel's numbers comes, instead, from lower-clocked parts, with the most significant increase being in the 2.7 GHz to 2.99 GHz range. Add to this the fact that Intel's increase in the Steam survey occurs in the month of March, when teleworking and the "bubble quarantine" began, and this actually seems like it's mostly motivated by users either turning on their older laptops, or as a result of increased sale of said laptops (where Intel still has a dominating product and consumer position over AMD's).
NVIDIA too saw increases in its market share for the month, which seems to support the theory that this run is powered by sales of new laptops or powering up of older hardware - especially considering the only DX-capable graphics cards to have shown increases are either in the DX12 or... DX8 or below support levels. Interesting times, to say the least.
Source:
Steam Hardware Survey
The March 2020 hardware survey shows Intel CPUs gaining traction over AMD's, albeit in a very limited fashion (a 0.24% increase in total number of machines running with GenuineIntel CPUs). Cross referencing this increase with the average clockspeeds of Intel CPUs, though, paints an interesting picture: the higher clockspeeds in the 3.3 to 3.8 GHz range (which can relatively cleanly be associated with the company's higher-performance products), saw a decline in usage (contrary to AMD, which saw the reverse happening).
53 Comments on Steam Hardware Survey March 2020: Intel CPUs, NVIDIA Graphics Cards Rising
That said, most of the time I have to reinstall when I change hardware so I wonder if it just sends it off when you run Steam for the first time...?
I hate stagnancy, and I am so happy that AMD comes back and challenge Intel once again.
I don't think we really need Intel leads over AMD to scrap every market.
I believe you have seen every news on this site on how every laptop/pc/desktop builder prefer to use Intel in their products.
I praise Intel marketing team and how they push every mobile/compact/desktop/server to use their CPU.
Count me among the less than 1% of 2080ti owners rocking a 5.1ghz cpuIt's my like 5th - 10th? I can't remember but I been there since day 1 like yourself over 16 years now I think my profile says. But I'm one of those 0.1 percenters with a 2080ti and 5.1ghz cpuJust bad luck I guess I've had many and been here since day 1 for half life 2.They can track this through a number of other ways besides the hardware survey.
. The moment you connect a vr enabled headset that works with steam it ask you to install a bunch of software to make your steam vr ready.
People mostly stay at home right now. Not necessarily gaming, but just using their desktops more often. And if you own a VR headset, it's probably connected to that desktop.
That said, it still surprises me that 1% Steam users have VR headsets at all.Of course. And people use Passmark or Mindfactory.de data in the exact same way - I don't remember you complaining.
Suddenly Intel gets a market share increase and "enthusiasts" go berserk. Because it's not allowed to show something like that.But why do you think it isn't? Any proof items are collected differently?
Sure, Steam doesn't give the exact methodology. So why not look into it? Ask them?Now you're talking about a motive. So what is it? How would Valve benefit?
[edit]
Oh, fun fact:
Also, its funny you connect the dots like this. I never said I don't believe Intel could have gained some % or anything like that. I'm saying, in general about this survey that we're not looking at unbiased, uncolored data. It is what it is... What each company does when is irrelevant in that story. But the sharp rise of VR headsets right now is simply a confirmation that we're not looking at a single dataset. It merely confirms that this data is colored and that is what I wanted to draw attention to.
The motive... do I need to spell it out? I already kinda did, didn't I. Valve's data will be explained along the lines of Valve's business objectives, and it will serve Valve's objectives, too. This news article is a prime example. They have that survey, so they can now show us how VR Is booming, and gosh, didn't they just release Alyx? In addition, their survey is exposure which is great marketing for Steam as a dominant platform for gaming. The fact they share it so openly however, should not be mistaken for transparency, and that is the problem here. That is also why I care a bit less about Mindfactory.de numbers - those are simply sales and those sales, relative to sales of same quarter in earlier years, can be extrapolated to some sort of accurate/plausible conclusion. Steam's numbers... you've given it a shot with VR, but do you really know? The only aspect that is great about the Survey is the sheer size and duration of it.
In the United States there are 280 million cars of which the top three most owned are pickup trucks. To get an accurate sample (within +/- 3%) of this fact I only need to randomly ask just 1,900 car owners what car they own. Not 10,000, not 1,000,000 just 1,900. To argue this sample size is equivalent to arguing 2+2 does not equal 4.
What will this survey prove? That the most popular car in the USA is a pickup truck. What this survey will not prove? That the best car is a pick up truck. That you should buy a pick up truck. That your friend should buy a pick up truck. Every question about cars that isn't "what's the most popular car in the USA?"
So lets get back to the steam hardware survey. What does it prove? It proves what the most popular hardware is across PC/laptops with the steam client and does so accurately. Does it represent most gaming hardware? It does a better job then most surveys because of its huge popularity.
What it does not prove? What hardware you should own, what hardware you should buy, what hardware is recommended. Every question about PC hardware that isn't "What's the most popular PC/laptop hardware for steam users?"
People need to stop being butt hurt because the survey fails to reflect their favorite brand, amount of cpu cores, system ram, monitor resolution, video ram, etc., etc.,
Also, are you familiar with the real vs simulated coin toss experiment? It's an excellent illustration of how poor the human brain fares when trying to grasp randomness.
First time in many years.