I’ll also add on that there are a LOT of blog posts and youtube shorts about “Game X is 20% faster on Linux than Windows!!!” that everyone loves to regurgitate. And the reality is that it was a single outlier or it all boils down to Steam distributing “good enough” shaders to Linux but not Windows (and let’s not get into the weeds of why).
Whereas GN, especially since “All New Data” a few years back, have very heavily focused on reproducible and “good” data. That is why Steve basically apologized for not having error bars or having what looks like messy data for a few of those runs. And they’ve done entire videos on their testing methodology that often includes MANY runs to normalize out the noise.
So without being able to explain exactly why? I doubt they will EVER put Windows and Linux data on even the same page of their website. But… someone who cares will be able to see trends.
GamersNexus’ normal GPU benchmark videos are to help gamers compare GPU performance on various games to determine what they should purchase for their needs. With this new video, they are now providing the same service for Linux Gamers going forward.
The goal of this video was not to compare Windows performance to Linux performance. There are videos that exist which do that, if that is what you’re were hoping for.
If we want to convince the windows crowd that linux is a viable alternative, we really need the comparison to show the difference isn’t so bad.
I’ve seen the ancient gameplays video you linked, but there is very little out there for linux vs windows benchmarks that are of high quality. Most videos tend to be incredibly amateur. I really hoped they would throw in a couple of composite charts with windows vs linux results since GN has all this data already and for a simple summary it shouldn’t have been much work at all. Instead I need to look at multiple videos.
What it needs is for the rig you have now to be basically playable. If it’s playable on Windows, it should be playable on Linux. Losing a couple of fps is frankly not a big deal if it’s still overall playable.
Setting aside some anti-cheat issues, it mostly is. Nvidia has some abysmal inconsistencies in a few titles.
I know first hand that if you have a great rig, things are fine. 9800x3d + 9070xt can play anything, anticheat willing :P
It’s still important to see the difference though imo. The more it becomes a meaningful metric the more software developers will consider it. Especially if the western gaming publications start publishing it. It’s the metric of “Is this game shit because the devs haven’t bothered to look at how it performs in wine?” ProtonDB is a start, but still lacks non-steam titles and it isn’t prominently shown in places where it ought to be, like steam store pages. The closest thing to a linux compatibility check on a steam store page is if it’s supported by the deck. Tons of games I play aren’t really suitable to the deck but run wonderfully on linux.
Plus the 5060 example is a good illustration as to why it’s important to see this stuff. On at least one game it’s practically unplayable on linux, but runs basically fine on windows at the same settings. These scenarios are exceedingly few, but the cheaper, weaker GPUs are the ones that sell the most.
If linux market share continues to increase we will see more and more linux native builds and the situation will improve substantially. It’s already a wonder that wine works as well as it does, way better than just a few years ago anyway.
I’ve seen the ancient gameplays video you linked, but there is very little out there for linux vs windows benchmarks that are of high quality.
The video I linked is 2 months old, only focuses on graph data (no gameplay videos) across both AMD and NVidia cards and multiple distros. It’s quite high quality IMHO. Are you confusing it with something else?
EDIT: I just realized you’re referring to the name of his youtube channel. D’oh!
I do want linux to gain market share, but this stuff is important. Eagle 5060 can’t even run Dragons Dogma 2 really on linux at 1440p max settings, but can handle it on windows at 1440p max settings. Doesn’t apply to me… but this is why we need to see the parity.
FWIW, Gamers Nexus specifically said during the video multiple times not to compare the results between Linux and Windows as they aren’t apples to apples in data.
Sure, and you can never have something be apples to apples when the entire architecture simply works differently between the two, unless you have some kind of external hardware monitor.
It doesn’t matter though. Imperfect comparisons are ten thousand times better than no comparisons at all, especially when entertainment is on the line!
Is there really zero windows benchmarks for comparison in this entire video?
Yes. They were very concerned about head to head comparisons because the tools for measuring FPS and stuff works differently.
I’ll also add on that there are a LOT of blog posts and youtube shorts about “Game X is 20% faster on Linux than Windows!!!” that everyone loves to regurgitate. And the reality is that it was a single outlier or it all boils down to Steam distributing “good enough” shaders to Linux but not Windows (and let’s not get into the weeds of why).
Whereas GN, especially since “All New Data” a few years back, have very heavily focused on reproducible and “good” data. That is why Steve basically apologized for not having error bars or having what looks like messy data for a few of those runs. And they’ve done entire videos on their testing methodology that often includes MANY runs to normalize out the noise.
So without being able to explain exactly why? I doubt they will EVER put Windows and Linux data on even the same page of their website. But… someone who cares will be able to see trends.
And yet there’s still LTT forum idiots posting on how “Steve is just running a drama channel”.
GamersNexus’ normal GPU benchmark videos are to help gamers compare GPU performance on various games to determine what they should purchase for their needs. With this new video, they are now providing the same service for Linux Gamers going forward.
The goal of this video was not to compare Windows performance to Linux performance. There are videos that exist which do that, if that is what you’re were hoping for.
If we want to convince the windows crowd that linux is a viable alternative, we really need the comparison to show the difference isn’t so bad.
I’ve seen the ancient gameplays video you linked, but there is very little out there for linux vs windows benchmarks that are of high quality. Most videos tend to be incredibly amateur. I really hoped they would throw in a couple of composite charts with windows vs linux results since GN has all this data already and for a simple summary it shouldn’t have been much work at all. Instead I need to look at multiple videos.
What it needs is for the rig you have now to be basically playable. If it’s playable on Windows, it should be playable on Linux. Losing a couple of fps is frankly not a big deal if it’s still overall playable.
Setting aside some anti-cheat issues, it mostly is. Nvidia has some abysmal inconsistencies in a few titles.
I know first hand that if you have a great rig, things are fine. 9800x3d + 9070xt can play anything, anticheat willing :P
It’s still important to see the difference though imo. The more it becomes a meaningful metric the more software developers will consider it. Especially if the western gaming publications start publishing it. It’s the metric of “Is this game shit because the devs haven’t bothered to look at how it performs in wine?” ProtonDB is a start, but still lacks non-steam titles and it isn’t prominently shown in places where it ought to be, like steam store pages. The closest thing to a linux compatibility check on a steam store page is if it’s supported by the deck. Tons of games I play aren’t really suitable to the deck but run wonderfully on linux.
Plus the 5060 example is a good illustration as to why it’s important to see this stuff. On at least one game it’s practically unplayable on linux, but runs basically fine on windows at the same settings. These scenarios are exceedingly few, but the cheaper, weaker GPUs are the ones that sell the most.
If linux market share continues to increase we will see more and more linux native builds and the situation will improve substantially. It’s already a wonder that wine works as well as it does, way better than just a few years ago anyway.
The video I linked is 2 months old, only focuses on graph data (no gameplay videos) across both AMD and NVidia cards and multiple distros. It’s quite high quality IMHO. Are you confusing it with something else?
EDIT: I just realized you’re referring to the name of his youtube channel. D’oh!
big bars are from the video. small bars are from https://gamersnexus.net/gpus/amd-needs-just-shut-amd-radeon-rx-9060-xt-16gb-gpu-review in july
I do want linux to gain market share, but this stuff is important. Eagle 5060 can’t even run Dragons Dogma 2 really on linux at 1440p max settings, but can handle it on windows at 1440p max settings. Doesn’t apply to me… but this is why we need to see the parity.
9070xt seems fairly close, compared to nvidia
FWIW, Gamers Nexus specifically said during the video multiple times not to compare the results between Linux and Windows as they aren’t apples to apples in data.
Sure, and you can never have something be apples to apples when the entire architecture simply works differently between the two, unless you have some kind of external hardware monitor.
It doesn’t matter though. Imperfect comparisons are ten thousand times better than no comparisons at all, especially when entertainment is on the line!
Only a dead man gets in between me and my entertainment.
Sir, this is a funeral.