idk how you are the only guy to combine hardware tests and VR. I mean considering how VR is the most demanding type of app, you would think more gamers would be making things like this, but apparently not. I guess it still doesn't have that many users yet.
Yeah, I'd love to see a bigger channel with far more resources than I have at my disposal get into doing some VR benchmarks, but it's still kind of a niche things compared to gaming in general. There's a good chunk of people who play VR but a lot of it is in standalone.
@@Omniwhatever hmm I guess there is standalone too that I forgot about. I still think PCVR is the way forward though. People say that making exclusive PCVR games just isn't profitable, but there just need to be longer length games to keep people using these headsets. Can't prove it, but I know that's what I think as a player. We will see.
I was surprised myself at the results! Especially given how everyone recommends the 5800x3D. It still is a good CPU don't get me wrong! Some of these conditions were quite heavy. The 5800x3D seems to do fairly well up to a certain point, then the Vcache seems to get overwhelmed and it starts to falter since there's so much going on it can't rely solely on it. That's some ancendotal comments from a few people I've talked with who had the CPU. Still not a bad value option, especially if on AM4 already since it's just a drop in upgrade.
@@Omniwhatever Dang maybe I’ll get some use out of my overkill DDR4 CL14 32gb ram kit. I bought that just before 5800X3D came out and went with that. Not knowing the vcache basically negates the need for ram in all games. Looks like vrchat might see slight improvements (but nothing near the generational uplift you’d get with ddr5 and zen 4). I’d be interested in a benchmark with 5800x3d comparing a low end 32gb ddr4 kit to a high end 32 gb ddr4 kit in vrchat.
While your take away may be that the 7950x3d is the best CPU for VRChat here, it actually performs the same as a 7800x3d, so if you are looking to upgrade you can save yourself some money by not jumping all the way up. the 7950x3d uses the same cores as the 7800x3d when gaming.
Fantastic information. This content is why I enjoy and keep returning to your channel. I've been rocking the 7950X with the best AMD DDR5 since launch day, but I've been curious if the 7000 series X3D CPU's would make any difference for the Aero/Crystal resolution in VR. Thanks again for the detailed video!
I’m actually friends with green, I didn’t know he did benchmarks with you, that’s actually really cool to know, thanks for showing off the differences of the cpus with him
was awaiting this vid for quite a while, glad you like your 7950x3D as much as i do my 7800x3D which at least doubled my fps in vrchat. have still a RAM issue which locks me to the default 4800mhz instead of the 6000mhz that it should run on my msi board, but i don't notice any fps difference anyway, so i will keep it that way for now.
My 7800x3d has same problem. I have corsair dominator 6000mhz. It works perferct at 5200mhz but I have it at 5600mhz also works perfect but takes longer to start up.
A few months ago, I went from a 5900x to a 7800X3D, 64gigs of DDR5 (6000 at cl-30) with a 4090, X670 chipset. For VR, especially sims (racing and flight) the difference was nutzoids. The 7800X3D just makes everything smoother and better FPS being a bonus (this is without my changing any in game settings). This is using a Reverb G2v2 at SteamVR (or OpenXR) 100% resolution of 3160 x 3092. Yes, I am impressed. As for VR Chat, very difficult to get anything steady in that title, so many poorly optimized worlds and avatars, you could be sitting in a world getting 90 fps, a couple bad avatars walk in and your frame rate take a major dive. I have a Crystal Light on order (late August delivery according to Pimax delivery chart), very anxious to try this thing out. :) As for my 7800X3D, very happy with my purchase. :)
You really gotta get used to constant frame spikes when playing something like VRChat with how poorly optimized a lot of things are, I tell you. The x3D CPUs help A LOT with it though. People really underestimate the power of a good CPU in that game!
I'm feeling so good with my 7800X3D (I don't play VRChat) having seen your video. 😃Looking forward to your Crystal review! I'm curious if you find the same pros and cons as I did. I even uploaded two very short videos about my issues, maybe you want to watch them. Thanks for the video and keep it up! 🙂👍
Thanks! The Crystal review is gonna be... Extremely, extremely long. Since there is so damn much to talk about. Easily will be the longest video I've ever done. I can at least confirm the tracking issues you saw in your videos seem much improved with the latest update that went out just a little over a week ago for beta testers.
@@Omniwhatever That's good to hear! The last regular firmware-update for the controllers a few days ago made no difference - just like the others before. Fingers crossed that the current beta-stuff will be released to public soon, also because I'm a little frustrated by the fact that I see DFR with so many RUclipsrs for many weeks now...
Yeah, I feel you there. I'm hoping it'll be out rather soon! Because it's honestly pretty incredible some of the stuff you can do with it. Gonna be making a dedicated video on DFR too at some point.
Im on a 9700f. Looking at a 13600kf to upgrade to pair with 4060 ti 16gb. I was torn between a 5800x3d and i5 13600kf. I sometimes do 2d animation and video editing. I play mostly VR (soon on a Quest 3 via airlink) games and sometime go on to vr chat with slime trackers. After watching your video, im leaning towards i5 13600fk now. Thank you for in depth testing instead of just calling a product trash like a lot of people do.
Interesting, and thanks for sharing all this mate. I think it may have been better to compare with a newer 13th gen i9 13900k or i9 13900ks. Comparing a new cpu with a much older one isn't a fair comparison imho. In any case, I'm getting fantastic results with my 10mo QPro (mainly wireless pcvr with air link) and my i913900k/rtx4090/32Gb 5200ram/z790 mb. It will be interesting to see how much better the soon to be released 14th gen 14900k cpu is going to perform. If it's significant I may look at upgrading it, plus my 5200ram. Thanks again for your vids mate and best cheers from Australia.
did you used the --affinity=FFFF parameter for the 7950x3d? Tupper actually points that one out as necessary to use that processor at it's full potential
when you benchmark performance or something you must use identical specs and test parameters(test conditions). i think comparing in VRChat is never equal in any 10 minutes past as users move, avatars change, and people stand still or just log out or log in with that 300MB avatar sometimes and that's a huge frame drop.
That's exactly why we did short tests of only around a minute or so and in almost the exact same spot in VRChat, without moving around in the public lobbies. It's about like for like as you could get with the circumstances because two different systems were used for the public lobbies and we only compared the 5800x3D vs one other CPU. It's why I'd never test on a single system in REAL WORLD circumstances in VRChat since the very nature of it means you kind of need to have two systems benchmarking at the same time if you're in lobbies with people. I noted there were things we still couldn't perfectly control but comparison is close enough I think it's still valid to get the general point across. We could only do a proper cross comparison with all 3 CPUs in the completely private lobbies where they'll be little to no changes because of that.
@@Omniwhatever 12:00 there's no way a 130 on the 7950x3D which is absurd as there are not many FPS unlocked VR headsets. You must isolate the variables or you get nonsensical results like these. Benchmarking apex legends is hard for this reason, but channels like Hardware Unboxed(?) and Gamers Nexus were able to by making a private lobby with placed bots/characters/other computers in specified locations that move at specified intervals specified distances and take pre-written action. They should have more details in the videos where they mention the methodology. The difference actually exists, you got your point across. But titles for VR are generally not CPU heavy but GPU heavy. One more suggestion I have is lowering the steamVR resolution for future tests so it does not reach 90% or above to avoid being GPU bound.
@@harurosech.4848 No, that is correct. Sorry, you are misunderstanding these benchmarks and the metrics they take. You can absolutely have frametime which is below your headset's hz limit, because it means the part is capable of rendering faster than the game demands to run at that hz, be it on the GPU or CPU as those frametimes are measured independently by FPSVR. But unlike flatscreen you can't pre-render more frames than your screen can handle for things like latency reasons with how VR is. If my GPU could hypothetically keep up to deliver 130~ FPS, then that's what it would be able to deliver BEFORE the CPU became the limiting factor. These results are not at all nonsensical. Broadly speaking for what the user can see without ripping apart the entire render pipeline but you'd have to be the game dev for that. If the CPU frame time is higher than the GPU you're generally looking at the CPU being the limiting factor where the GPU is the limiting one if the frametimes are higher on that. In that scenario the GPU was the limiting factor. The entire point of this was to show a real world situation and compare how a CPU upgrade could or could not improve your actual frames. In that situation the 7950x3D was SO much better than the 5800x3D running at identical settings and practically the same res that it removed the bind from the CPU to the GPU, with significant headroom at that. Kind of the point to show. The comparison is totally valid and those frametimes are accurate. We even did some quick private room tests with nobody else in them too. The tests in the lobbies were meant to illustrate something that's actually what people are gonna run into since not many just run in a purely empty world. Hell, even in GPU bound situations you can have frame times below 8.3ms in FPSVR on a 120hz headset. Because the GPU only takes that amount of time to render the frame, but then it's just kind of sitting there resting since it can't deliver more than 120hz. In a situation like that, if you got a higher hz HMD it would then be able to deliver those frames, all else identical. The point of using the "100%" res is because VR has far less standardized resolutions than flatscreen so it's be stupid to lower the resolution significantly just to try and force a CPU bind with how variable it can be from headset to headset. The entire point was trying to show real world differences in as reasonable identical conditions. It's not like flatscreen benchmarks where you only have 1080p, 1440p, and 4k as the main used ones with the rare(but still standardized) outlier.
Do you still recommend the 7950x3d with a 4090 a year later? I am shopping for this rig right now and keen hear your thoughts, noting I do lots of VR and some content creation (early days). Thanks in advance!
7950x3D is still the best you can buy right now if you do a lot of gaming and productivity or run a metric ton of programs on the side. I know some people really like to run a lot of stuff with VRChat, for instance. The 9000 series seems kind of like a letdown so might not expect much of a change with the 9000x3D chips, but if you got some patience may be worth waiting to see how they pan out just in case. But could also be a good time to snag the top end of last gen for cheap without much of a loss. 4090, I'd wait a month to see if Nvidia announces the 5000 series. They announced their next gen in September, give or take a month, usually.
Pretty cool video! thnx for the info. i do a ton of Sim racing in VR. i specifically play Iracing and im super cpu bound right now. current setup, ryzen 5 5600 and a rtx 4080. looking to get the Ryzen 7 7800X3D. do you recommend just spending the extra cash on the 7950x3d? i appreciate everyones opinion. thnx in advance
They are both very good cpus I just got the i9 12900k for my pc and still getting other parts for this pc and I can’t wait to get it up and running I play Vr chat all the time on a gtx 1650 laptop rn😅 and I get 70 fps with ppl in the world
Thanks you so much for this test. If it possible I need you to test between 12900k with 7950x3d with high polygons workload avatars. because I still thinking about I should upgrade from intel to Amd with x3d series or not. Thanks you!
Hey, I've already got the 12900k packed up as i was selling it to a friend to afford the x3D CPU. And I can't really test the 12900k vs 7950x3D in a real world situation with full lobbies becuase of how much the content can vary. I only have 1 PC to use at a time, least with identical specs, sorry.
It's FPSVR on SteamVR. Don't think there's really a lot worth covering in a dedicated video since I think it's rather self-explanatory with some of the stuff I've already covered.
Do you still have the possiblity to do a test like this in dcs? from what ive heard it is very intensive on the cpu and loves 3dcache so itd guess the difference would be very big
Unfortunately, I don't sorry. I sold the 12900k set up as part of the reason to afford the 7950x3D for this test. Not quite a big enough channel to keep a bunch of expensive gear around solely for benchmarking.
Does your conclusion for non-VRChart games extend to the 5800x3d as well, that is not noticeable difference? I have a 4090 and Crystal on the way. Don't play VRChat, main focus is sim racing and will want to use DFR to reach 120fps for 120hz mode. Many thanks in advance!
The 5800x3D couldn't be tested outside of VRChat, so I can't really comment, sorry. VRChat's a pretty unique game so I'd be hesitant to draw conclusions for other games concerning it. If it's even within 10% of the 12900k though like 90% of games won't struggle on it. Not sure on sims, but I know some sims, like ACC, REALLY love Vcache(At least with the 7000x3D series), seeing even notable benefits at 4k in flatscreen even when most stuff is long GPU bound.
@@Omniwhatever thank you! I'm new to VR (only tested a Pico 4 recently which has lead me down the VR rabbit hole) so a noob question to ask but correct me if I'm wrong, I get the sense from your video that DFR can cause extra CPU load which can then lead to overall worse performance than the benefit gained from reducing GPU load with the number of pixels needed rendering? And that can cause performance to be worse in some cases? Also, is if fair to look at 4k 4090 benchmarks of the race sims I'm interested in between the 7800x3d and 5800x3d and any differences in performance (if any) should translate across with the Crystal as well?
@@DrR1pper So, I plan on making a dedicated video concerning DFR as there's a lot to talk about there and discuss, along with some benchmarks, but that's not quite how it works. So, it depends on the kind of game you play. Most of the games I tested had a negligible CPU load impact. There were just a few anomolies(Like Skyrim) which saw a MASSIVE increase. In those edge cases, yes you can end up CPU bound, but it should be rare. As for your 4k question... Kind or? But also not really. It'll probably give you a general idea of what to expect in terms of a bottleneck. VR is much, MUUUUCH higher resolution and the vast majority of the time you won't be CPU bound because of that, as long as you've already got something that's upper end. There's just a handful of games which're gonna hit your CPU hard. Sims do appear to be one of those from what I have heard though. So you may want to look at the 7800x3D. It'll last you longer too since you might be able to upgrade to a better CPU in the future since the socket will be good for longer. While AM4 is EOL.
I'm really not into flight sims at all, especially one as hyper realistic as DCS, so I really don't think I'd be able to do a good benchmark for it, sorry.
@Wasmachineman Actually, it is power limited dumby, funny enough that's not how I was using the word anyways. I meant you can bring out the full potential of the cpu, "Power!" Also, PBO2 increased my 5800x3d performance by nearly x1.25 in certain games and applications. When setup correctly.
@@Purified1k I find it incredibly hard to believe PBO2, which is literally increasing your power limits and nothing more, increases your performance by 25%.
@Wasmachineman It did when setup correctly. Also, I didn't increase the power. Matter affect I used the negative option. Which keeps the CPU cooler allowing the entire CPU to perform better. I also decreased the power limits a little. (PPT, TDC, and EDC). Also that 25% was game and application dependent. So overall, it was a 5-25% increase.
You need to compare with a 4090 against the 13900K or KS to get even a reasonably fair comparison. 8000MTs 48GB RAM as well. 14900K incoming very shortly and I believe a 14900KS in January which I will get to upgrade my 13900K.
If I had the money, I would. But I'm not a big enough channel to buy a bunch of high end gear like that for the sole purposes of benchmarking and sort of use what I have around. I disagree that it's a completely unfair comparison though. Intel benefits from memory scaling but not to a massive degree. It's much less memory sensitive than Zen 4 is, though the x3D chips are not very memory sensitive either due to how they work. There's not any game benchmark I've seen where the 13900k's been so far ahead of the 12900k to bridge the gap where the 7950x3D pulled noticeably ahead here. But those could also just be Vcache sensitive scenarios.
Easily. That's pretty much the best gaming CPU you can get without getting REAL high budget. If you're doing a bunch of multi-tasking stuff it may get strained depending on the workload, but you'll be running fine.
@bhumelti636 Please don't discuss trying to sell anything in my comments, take it somewhere else. Any further discussion of details here will be deleted.
Lol that's ridiculous. I'm not even big on VRChat outside the Vket events. It's THE game which always gets brought up as being a big deal with Vcache, above all else, to where lots of people even(wrongly) claim the 5800x3D is much better than even higher end Intel stuff because of it and everyone constantly brings it up. Or massively exaggerate the benefits, I've seen some people flatout say the 7950x3D is several times better than a 13900k because of it when that's also not true. I even tested flat2VR mods like the Resident Evil VR mods, not using the upscaler beta branch which reduced CPU workload, and that's also really tough on the CPU and there was no real difference so I didn't want to waste time when it was the same story as virtually every game tested. And I'd be willing to bet you it's the same story there because numerous people who play MSFS, some which I know, got huge benefits from a 4090 even if they weren't on the best CPU in the world, which wouldn't happen if CPU bound. This isn't like comparing a CPU from 2-3 generations ago with a current top of the line and you're more liable to find meaningful differences in more than like 2-3 games, these are already two very high end CPUs and just like with 4k resolution your GPU tends to be overwhelmingly more important. But those few, very rare, instances were part of what inspired the video.
@@WorldHayes-VR That's a tough one, haven't thought about it too much. If I were to discount Alyx and Skyrim simply because almost everyone choses those and that's 2/5 already, then of what I've personally played(At least mostly to completion for single player games) I've really enjoyed Project Wingman, Contractors VR(For the mods), Saints and Sinners, the RE4 Quest Port, and Until You Fall. I don't have any experience with the PSVR line of things to comment much there.
idk how you are the only guy to combine hardware tests and VR. I mean considering how VR is the most demanding type of app, you would think more gamers would be making things like this, but apparently not. I guess it still doesn't have that many users yet.
Yeah, I'd love to see a bigger channel with far more resources than I have at my disposal get into doing some VR benchmarks, but it's still kind of a niche things compared to gaming in general. There's a good chunk of people who play VR but a lot of it is in standalone.
@@Omniwhatever hmm I guess there is standalone too that I forgot about. I still think PCVR is the way forward though. People say that making exclusive PCVR games just isn't profitable, but there just need to be longer length games to keep people using these headsets. Can't prove it, but I know that's what I think as a player. We will see.
Crazy testing. Thanks for using the 5800x3d. I am pairing it with a 4070. Did not expect that huge a difference from the 7950x3d
I was surprised myself at the results! Especially given how everyone recommends the 5800x3D. It still is a good CPU don't get me wrong! Some of these conditions were quite heavy. The 5800x3D seems to do fairly well up to a certain point, then the Vcache seems to get overwhelmed and it starts to falter since there's so much going on it can't rely solely on it. That's some ancendotal comments from a few people I've talked with who had the CPU.
Still not a bad value option, especially if on AM4 already since it's just a drop in upgrade.
I really wanted to know for myself the comparison. I am glad I asked to join the benchmarking XD
@@Omniwhatever Dang maybe I’ll get some use out of my overkill DDR4 CL14 32gb ram kit. I bought that just before 5800X3D came out and went with that. Not knowing the vcache basically negates the need for ram in all games. Looks like vrchat might see slight improvements (but nothing near the generational uplift you’d get with ddr5 and zen 4).
I’d be interested in a benchmark with 5800x3d comparing a low end 32gb ddr4 kit to a high end 32 gb ddr4 kit in vrchat.
While your take away may be that the 7950x3d is the best CPU for VRChat here, it actually performs the same as a 7800x3d, so if you are looking to upgrade you can save yourself some money by not jumping all the way up.
the 7950x3d uses the same cores as the 7800x3d when gaming.
You got the same as me. What headset do you have? And what setting in msfs2020 if you use?
Interesting results. Makes feel even better about my 7950x3d build. Thanks for the video.
I’ve been looking for these kind of videos for months, so thankful of this channel!
Fantastic information. This content is why I enjoy and keep returning to your channel.
I've been rocking the 7950X with the best AMD DDR5 since launch day, but I've been curious if the 7000 series X3D CPU's would make any difference for the Aero/Crystal resolution in VR. Thanks again for the detailed video!
I’m actually friends with green, I didn’t know he did benchmarks with you, that’s actually really cool to know, thanks for showing off the differences of the cpus with him
was awaiting this vid for quite a while, glad you like your 7950x3D as much as i do my 7800x3D which at least doubled my fps in vrchat.
have still a RAM issue which locks me to the default 4800mhz instead of the 6000mhz that it should run on my msi board, but i don't notice any fps difference anyway, so i will keep it that way for now.
My 7800x3d has same problem. I have corsair dominator 6000mhz. It works perferct at 5200mhz but I have it at 5600mhz also works perfect but takes longer to start up.
Was looking through CPUs for VR gaming. Your video helped, thanks!
A few months ago, I went from a 5900x to a 7800X3D, 64gigs of DDR5 (6000 at cl-30) with a 4090, X670 chipset. For VR, especially sims (racing and flight) the difference was nutzoids. The 7800X3D just makes everything smoother and better FPS being a bonus (this is without my changing any in game settings). This is using a Reverb G2v2 at SteamVR (or OpenXR) 100% resolution of 3160 x 3092. Yes, I am impressed.
As for VR Chat, very difficult to get anything steady in that title, so many poorly optimized worlds and avatars, you could be sitting in a world getting 90 fps, a couple bad avatars walk in and your frame rate take a major dive.
I have a Crystal Light on order (late August delivery according to Pimax delivery chart), very anxious to try this thing out. :)
As for my 7800X3D, very happy with my purchase. :)
You really gotta get used to constant frame spikes when playing something like VRChat with how poorly optimized a lot of things are, I tell you. The x3D CPUs help A LOT with it though. People really underestimate the power of a good CPU in that game!
@@Omniwhatever Agreed. :)
I'm feeling so good with my 7800X3D (I don't play VRChat) having seen your video. 😃Looking forward to your Crystal review! I'm curious if you find the same pros and cons as I did. I even uploaded two very short videos about my issues, maybe you want to watch them. Thanks for the video and keep it up! 🙂👍
Thanks! The Crystal review is gonna be... Extremely, extremely long. Since there is so damn much to talk about. Easily will be the longest video I've ever done. I can at least confirm the tracking issues you saw in your videos seem much improved with the latest update that went out just a little over a week ago for beta testers.
@@Omniwhatever That's good to hear! The last regular firmware-update for the controllers a few days ago made no difference - just like the others before. Fingers crossed that the current beta-stuff will be released to public soon, also because I'm a little frustrated by the fact that I see DFR with so many RUclipsrs for many weeks now...
Yeah, I feel you there. I'm hoping it'll be out rather soon! Because it's honestly pretty incredible some of the stuff you can do with it. Gonna be making a dedicated video on DFR too at some point.
You and Drab Majesty both are so underated
What a coincidence that this video comes out the day before my 7800x3d arrives. lol.
Hope it's useful then! It's not exactly applicable since I tested the 7950x3D, but should give a ballpark idea.
Im on a 9700f. Looking at a 13600kf to upgrade to pair with 4060 ti 16gb. I was torn between a 5800x3d and i5 13600kf. I sometimes do 2d animation and video editing. I play mostly VR (soon on a Quest 3 via airlink) games and sometime go on to vr chat with slime trackers. After watching your video, im leaning towards i5 13600fk now. Thank you for in depth testing instead of just calling a product trash like a lot of people do.
Interesting, and thanks for sharing all this mate. I think it may have been better to compare with a newer 13th gen i9 13900k or i9 13900ks. Comparing a new cpu with a much older one isn't a fair comparison imho. In any case, I'm getting fantastic results with my 10mo QPro (mainly wireless pcvr with air link) and my i913900k/rtx4090/32Gb 5200ram/z790 mb. It will be interesting to see how much better the soon to be released 14th gen 14900k cpu is going to perform. If it's significant I may look at upgrading it, plus my 5200ram. Thanks again for your vids mate and best cheers from Australia.
Yesterday I decided to upgrade from my 5800x3d to 7800X3D to go with my new Crystal & 4090. Looks like there is good benefit!
are u feel big difference from 5800x3d to 7800x3d?
@@Titanum7899 for VR DCS and MSFS yes. For more simple games they are not as demanding
You were doing God’s work, my brother
Can’t wait to upgrade, but I’m waiting for the 9000 series to come out
did you used the --affinity=FFFF parameter for the 7950x3d? Tupper actually points that one out as necessary to use that processor at it's full potential
Yes, I did account for that. I covered it in detail from 3:04 to 5:31. That's not the only way to get a game to run exclusively on the vcache cores.
when you benchmark performance or something you must use identical specs and test parameters(test conditions). i think comparing in VRChat is never equal in any 10 minutes past as users move, avatars change, and people stand still or just log out or log in with that 300MB avatar sometimes and that's a huge frame drop.
That's exactly why we did short tests of only around a minute or so and in almost the exact same spot in VRChat, without moving around in the public lobbies. It's about like for like as you could get with the circumstances because two different systems were used for the public lobbies and we only compared the 5800x3D vs one other CPU. It's why I'd never test on a single system in REAL WORLD circumstances in VRChat since the very nature of it means you kind of need to have two systems benchmarking at the same time if you're in lobbies with people. I noted there were things we still couldn't perfectly control but comparison is close enough I think it's still valid to get the general point across.
We could only do a proper cross comparison with all 3 CPUs in the completely private lobbies where they'll be little to no changes because of that.
@@Omniwhatever 12:00 there's no way a 130 on the 7950x3D which is absurd as there are not many FPS unlocked VR headsets.
You must isolate the variables or you get nonsensical results like these.
Benchmarking apex legends is hard for this reason, but channels like Hardware Unboxed(?) and Gamers Nexus were able to by making a private lobby with placed bots/characters/other computers in specified locations that move at specified intervals specified distances and take pre-written action. They should have more details in the videos where they mention the methodology.
The difference actually exists, you got your point across. But titles for VR are generally not CPU heavy but GPU heavy.
One more suggestion I have is lowering the steamVR resolution for future tests so it does not reach 90% or above to avoid being GPU bound.
@@harurosech.4848 No, that is correct. Sorry, you are misunderstanding these benchmarks and the metrics they take. You can absolutely have frametime which is below your headset's hz limit, because it means the part is capable of rendering faster than the game demands to run at that hz, be it on the GPU or CPU as those frametimes are measured independently by FPSVR. But unlike flatscreen you can't pre-render more frames than your screen can handle for things like latency reasons with how VR is. If my GPU could hypothetically keep up to deliver 130~ FPS, then that's what it would be able to deliver BEFORE the CPU became the limiting factor. These results are not at all nonsensical. Broadly speaking for what the user can see without ripping apart the entire render pipeline but you'd have to be the game dev for that. If the CPU frame time is higher than the GPU you're generally looking at the CPU being the limiting factor where the GPU is the limiting one if the frametimes are higher on that.
In that scenario the GPU was the limiting factor. The entire point of this was to show a real world situation and compare how a CPU upgrade could or could not improve your actual frames. In that situation the 7950x3D was SO much better than the 5800x3D running at identical settings and practically the same res that it removed the bind from the CPU to the GPU, with significant headroom at that. Kind of the point to show. The comparison is totally valid and those frametimes are accurate. We even did some quick private room tests with nobody else in them too. The tests in the lobbies were meant to illustrate something that's actually what people are gonna run into since not many just run in a purely empty world.
Hell, even in GPU bound situations you can have frame times below 8.3ms in FPSVR on a 120hz headset. Because the GPU only takes that amount of time to render the frame, but then it's just kind of sitting there resting since it can't deliver more than 120hz. In a situation like that, if you got a higher hz HMD it would then be able to deliver those frames, all else identical.
The point of using the "100%" res is because VR has far less standardized resolutions than flatscreen so it's be stupid to lower the resolution significantly just to try and force a CPU bind with how variable it can be from headset to headset. The entire point was trying to show real world differences in as reasonable identical conditions. It's not like flatscreen benchmarks where you only have 1080p, 1440p, and 4k as the main used ones with the rare(but still standardized) outlier.
Do you still recommend the 7950x3d with a 4090 a year later? I am shopping for this rig right now and keen hear your thoughts, noting I do lots of VR and some content creation (early days). Thanks in advance!
7950x3D is still the best you can buy right now if you do a lot of gaming and productivity or run a metric ton of programs on the side. I know some people really like to run a lot of stuff with VRChat, for instance.
The 9000 series seems kind of like a letdown so might not expect much of a change with the 9000x3D chips, but if you got some patience may be worth waiting to see how they pan out just in case. But could also be a good time to snag the top end of last gen for cheap without much of a loss.
4090, I'd wait a month to see if Nvidia announces the 5000 series. They announced their next gen in September, give or take a month, usually.
@@Omniwhatever I did it, just setting it all up now. V excited to test!
Pretty cool video! thnx for the info. i do a ton of Sim racing in VR. i specifically play Iracing and im super cpu bound right now. current setup, ryzen 5 5600 and a rtx 4080. looking to get the Ryzen 7 7800X3D. do you recommend just spending the extra cash on the 7950x3d? i appreciate everyones opinion. thnx in advance
They are both very good cpus I just got the i9 12900k for my pc and still getting other parts for this pc and I can’t wait to get it up and running I play Vr chat all the time on a gtx 1650 laptop rn😅 and I get 70 fps with ppl in the world
Thanks you so much for this test.
If it possible I need you to test between 12900k with 7950x3d with high polygons workload avatars.
because I still thinking about I should upgrade from intel to Amd with x3d series or not.
Thanks you!
Hey, I've already got the 12900k packed up as i was selling it to a friend to afford the x3D CPU. And I can't really test the 12900k vs 7950x3D in a real world situation with full lobbies becuase of how much the content can vary. I only have 1 PC to use at a time, least with identical specs, sorry.
yo, wheres the link so I can copy that fit, ninja style looks cool.
what software do you use to show the stats in the top left for desktop and in vr? could you do a video on this?
It's FPSVR on SteamVR. Don't think there's really a lot worth covering in a dedicated video since I think it's rather self-explanatory with some of the stuff I've already covered.
@@Omniwhatever ooh ok thanks, i'll check it out
Do you still have the possiblity to do a test like this in dcs? from what ive heard it is very intensive on the cpu and loves 3dcache so itd guess the difference would be very big
Unfortunately, I don't sorry. I sold the 12900k set up as part of the reason to afford the 7950x3D for this test. Not quite a big enough channel to keep a bunch of expensive gear around solely for benchmarking.
Do you think an i5 9400f would be able to run blade & sorcery (heavily moded) at medium to good quality?
Does your conclusion for non-VRChart games extend to the 5800x3d as well, that is not noticeable difference? I have a 4090 and Crystal on the way. Don't play VRChat, main focus is sim racing and will want to use DFR to reach 120fps for 120hz mode. Many thanks in advance!
The 5800x3D couldn't be tested outside of VRChat, so I can't really comment, sorry. VRChat's a pretty unique game so I'd be hesitant to draw conclusions for other games concerning it. If it's even within 10% of the 12900k though like 90% of games won't struggle on it. Not sure on sims, but I know some sims, like ACC, REALLY love Vcache(At least with the 7000x3D series), seeing even notable benefits at 4k in flatscreen even when most stuff is long GPU bound.
@@Omniwhatever thank you! I'm new to VR (only tested a Pico 4 recently which has lead me down the VR rabbit hole) so a noob question to ask but correct me if I'm wrong, I get the sense from your video that DFR can cause extra CPU load which can then lead to overall worse performance than the benefit gained from reducing GPU load with the number of pixels needed rendering? And that can cause performance to be worse in some cases? Also, is if fair to look at 4k 4090 benchmarks of the race sims I'm interested in between the 7800x3d and 5800x3d and any differences in performance (if any) should translate across with the Crystal as well?
@@DrR1pper So, I plan on making a dedicated video concerning DFR as there's a lot to talk about there and discuss, along with some benchmarks, but that's not quite how it works. So, it depends on the kind of game you play. Most of the games I tested had a negligible CPU load impact. There were just a few anomolies(Like Skyrim) which saw a MASSIVE increase. In those edge cases, yes you can end up CPU bound, but it should be rare.
As for your 4k question... Kind or? But also not really. It'll probably give you a general idea of what to expect in terms of a bottleneck. VR is much, MUUUUCH higher resolution and the vast majority of the time you won't be CPU bound because of that, as long as you've already got something that's upper end. There's just a handful of games which're gonna hit your CPU hard. Sims do appear to be one of those from what I have heard though. So you may want to look at the 7800x3D. It'll last you longer too since you might be able to upgrade to a better CPU in the future since the socket will be good for longer. While AM4 is EOL.
@@Omniwhatever thanks again!
Nice test! do yow can do a test with DCS?
I'm really not into flight sims at all, especially one as hyper realistic as DCS, so I really don't think I'd be able to do a good benchmark for it, sorry.
@@Omniwhatever
You should have optimized the 5800x3d. PBO2 brings out its full power.
lol no, the 5800X3D is NOT power limited. It's clock constrained.
@Wasmachineman Actually, it is power limited dumby, funny enough that's not how I was using the word anyways. I meant you can bring out the full potential of the cpu, "Power!" Also, PBO2 increased my 5800x3d performance by nearly x1.25 in certain games and applications. When setup correctly.
@@Purified1k I find it incredibly hard to believe PBO2, which is literally increasing your power limits and nothing more, increases your performance by 25%.
@Wasmachineman It did when setup correctly. Also, I didn't increase the power. Matter affect I used the negative option. Which keeps the CPU cooler allowing the entire CPU to perform better. I also decreased the power limits a little. (PPT, TDC, and EDC). Also that 25% was game and application dependent. So overall, it was a 5-25% increase.
You need to compare with a 4090 against the 13900K or KS to get even a reasonably fair comparison. 8000MTs 48GB RAM as well. 14900K incoming very shortly and I believe a 14900KS in January which I will get to upgrade my 13900K.
If I had the money, I would. But I'm not a big enough channel to buy a bunch of high end gear like that for the sole purposes of benchmarking and sort of use what I have around. I disagree that it's a completely unfair comparison though. Intel benefits from memory scaling but not to a massive degree. It's much less memory sensitive than Zen 4 is, though the x3D chips are not very memory sensitive either due to how they work. There's not any game benchmark I've seen where the 13900k's been so far ahead of the 12900k to bridge the gap where the 7950x3D pulled noticeably ahead here. But those could also just be Vcache sensitive scenarios.
Im building a rig with a core i9 12900k and 32gbs ddr5 so within the nex year or 2 i can get a used 14900k. Upgrading from socket 1200
The Index 144Hz enjoyer knows to turn down settings and resolution to achieve buttery smooth performance on his high-end cutting-edge X3D CPU
No comparison with the 7800x3d?
I don't have a 7800x3D to use.
7800x3d will be ok? With 4090?
Easily. That's pretty much the best gaming CPU you can get without getting REAL high budget. If you're doing a bunch of multi-tasking stuff it may get strained depending on the workload, but you'll be running fine.
@@Omniwhatever i am buying it only for vr gaming. for everything else I use my mac
Lol wish I had a vr so that this thing actually would help me I love vrs im trying to save up for one but im not even close
yet I still watch this vid
i’ll gladly sell my quest 2 to you
@bhumelti636 Please don't discuss trying to sell anything in my comments, take it somewhere else. Any further discussion of details here will be deleted.
@@bhumelti636 whats your discord so that we can move this somewhere else
You need cyberpunk or flight simulator benchmarks, you promoting VR chat id obvious
Lol that's ridiculous. I'm not even big on VRChat outside the Vket events. It's THE game which always gets brought up as being a big deal with Vcache, above all else, to where lots of people even(wrongly) claim the 5800x3D is much better than even higher end Intel stuff because of it and everyone constantly brings it up. Or massively exaggerate the benefits, I've seen some people flatout say the 7950x3D is several times better than a 13900k because of it when that's also not true.
I even tested flat2VR mods like the Resident Evil VR mods, not using the upscaler beta branch which reduced CPU workload, and that's also really tough on the CPU and there was no real difference so I didn't want to waste time when it was the same story as virtually every game tested. And I'd be willing to bet you it's the same story there because numerous people who play MSFS, some which I know, got huge benefits from a 4090 even if they weren't on the best CPU in the world, which wouldn't happen if CPU bound.
This isn't like comparing a CPU from 2-3 generations ago with a current top of the line and you're more liable to find meaningful differences in more than like 2-3 games, these are already two very high end CPUs and just like with 4k resolution your GPU tends to be overwhelmingly more important. But those few, very rare, instances were part of what inspired the video.
@@Omniwhatever what your top 5 VR games? I'm on Concle VR, Standalone VR, and PCVR. In the past Gear VR...
@@WorldHayes-VR That's a tough one, haven't thought about it too much. If I were to discount Alyx and Skyrim simply because almost everyone choses those and that's 2/5 already, then of what I've personally played(At least mostly to completion for single player games) I've really enjoyed Project Wingman, Contractors VR(For the mods), Saints and Sinners, the RE4 Quest Port, and Until You Fall.
I don't have any experience with the PSVR line of things to comment much there.
😜 Promo`SM