Jump to content

reliable comparisons between Amd Zen4, Zen4X3d, Intel 13th/14th gen for VR


twistking

Recommended Posts

Hello,
anecdotical evidence from reading forum signatures suggests, that AMD Zen4 CPUs are quite popular with DCS players these days.
I know that current gen AMD CPUs are generally considered very good CPUs for gaming and often beat intel 13th (and probably even 14th gen) in gaming focused benchmarks.
However, in some games Intel is stronger and generally speaking Intel CPUs seem to have slightly better multicore performance at a similar price point.

I'm currently deciding with which to go for my next build, which i want to specifically build around DCS VR.
Are there any hard facts which CPU might be better?
I've read many adulations from new AMD customers, but they of course will always only compare the new AMDs to an Intel chip that is several generations older, so no comparison really. Do we (the DCS community) have any reliable comaprisons on how those different chips compare to each other?

Thanks.

Link to comment
Share on other sites

47 minutes ago, twistking said:

Hello,
anecdotical evidence from reading forum signatures suggests, that AMD Zen4 CPUs are quite popular with DCS players these days.
I know that current gen AMD CPUs are generally considered very good CPUs for gaming and often beat intel 13th (and probably even 14th gen) in gaming focused benchmarks.
However, in some games Intel is stronger and generally speaking Intel CPUs seem to have slightly better multicore performance at a similar price point.

I'm currently deciding with which to go for my next build, which i want to specifically build around DCS VR.
Are there any hard facts which CPU might be better?
I've read many adulations from new AMD customers, but they of course will always only compare the new AMDs to an Intel chip that is several generations older, so no comparison really. Do we (the DCS community) have any reliable comaprisons on how those different chips compare to each other?

Thanks.

For flight sims x3D cpus seems to be simply the best:

 

dUnqqvmPBURYq8NPY5SW85.png

Link to comment
Share on other sites

3 minutes ago, maxsin72 said:

For flight sims x3D cpus seems to be simply the best:

dUnqqvmPBURYq8NPY5SW85.png

Thanks for the reply.
I would want to point out however, that (Microsoft) Flight Simulator is not necessarily comparable to DCS. VR in DCS is a bit of a diva concerning performance, not because virtual aircrafts are involved, but because the application bruteforces inefficient stereo rendering over a thinly multithreaded rendering thread carefully balanced ontop of a heap of vintage spaghetti code. The only valid similarity between the two - i would argue - is that they are both VR compatible games that are easily CPU bottlenecked.

Link to comment
Share on other sites

To add to the above, that Civ sim is I believe single threaded and it’s VERY easy to bog it down.

System: 9700, 64GB DDR4, 2070S, NVME2, Rift S, Jetseat, Thrustmaster F18 grip, VPC T50 stick base and throttle, CH Throttle, MFG crosswinds, custom button box, Logitech G502 and Marble mouse.

Server: i5 2500@3.9Ghz, 1080, 24GB DDR3, SSD.

Link to comment
Share on other sites

52 minutes ago, twistking said:

because the application bruteforces inefficient stereo rendering

Doesn't VR by its very definition require stereo rendering?


Edited by SharpeXB

i9-14900KS | ASUS ROG MAXIMUS Z790 HERO | 64GB DDR5 5600MHz | iCUE H150i Liquid CPU Cooler | 24GB GeForce RTX 4090 | Windows 11 Home | 2TB Samsung 980 PRO NVMe | Corsair RM1000x | LG 48GQ900-B 4K OLED Monitor | CH Fighterstick | Ch Pro Throttle | CH Pro Pedals | TrackIR 5

Link to comment
Share on other sites

44 minutes ago, SharpeXB said:

Doesn't VR by its very definition require stereo rendering?

well technically yes*. i'm really not an expert on this, but from my understanding DCS is a bit inefficient in its stereo rendering, because the CPU prepares two scenes completely seperately, one for each eye (obviously), while more efficient approaches would allow the pipeline to share much work needed for each eye, so that stereoscopy would not double the workload for the CPU (one scene with two views, instead of two scenes). Don't quote me on that though, i'm just an interested layperson.
What i wanted to express is that DCS's VR pipeline is not the most efficient and that that is most probably the main reason for the difficulties of getting optimal VR results.

*There is technically a way to make stereographic 3d from a single viewport render, by using the z-buffer and doing some post processing with it to create a stereo image. That gives you an image with depth, but it's a little bit fake and does not look as convincing as "true" stereo  rendering. I think it's dirt cheap to compute, but there must be a reason why it never really caught on.


Edited by twistking
Link to comment
Share on other sites

@twistking Better remove this chart from the thread because it is terribly misleading. VR is a scenario far far away from a galaxy that renders a flightsim on a flat monitor at 1920x1080.
All i can say is that a 13900K with a 4090 runs extremely well now with DCS 2.9 MT in VR, compared to the previous states. Every setting maxed out, DLSS enabled. I get between 60-80 fps (at resolution 3120x3120 per eye which is the middle standard for common VR panels today which feature a resolution around 2Kx2K per eye).
I was there from the beginning, going through several different hardware setups over the years and i am very happy atm.


Edited by RealDCSpilot
  • Like 1
  • Thanks 1

i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, Pico 4, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules

Link to comment
Share on other sites

3 hours ago, RealDCSpilot said:

@twistking Better remove this chart from the thread because it is terribly misleading. VR is a scenario far far away from a galaxy that renders a flightsim on a flat monitor at 1920x1080.
All i can say is that a 13900K with a 4090 runs extremely well now with DCS 2.9 MT in VR, compared to the previous states. Every setting maxed out, DLSS enabled. I get between 60-80 fps (at resolution 3120x3120 per eye which is the middle standard for common VR panels today which feature a resolution around 2Kx2K per eye).
I was there from the beginning, going through several different hardware setups over the years and i am very happy atm.

 

Here 7800x3d and 4090 with reverb g2: 90 fps on multiplayer dogfight server, cpu frame time about 3ms both with or without dlss. That chart tells the truth😀

Link to comment
Share on other sites

3 hours ago, maxsin72 said:

90 fps on multiplayer dogfight server

I personally prefer demanding servers like 4YA with 1000+ ground and air units and 40+ players. Having good fps on a server with nothing much else going on is pointless for any comparison.

i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, Pico 4, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules

Link to comment
Share on other sites

7 hours ago, RealDCSpilot said:

@twistking Better remove this chart from the thread because it is terribly misleading. VR is a scenario far far away from a galaxy that renders a flightsim on a flat monitor at 1920x1080.
All i can say is that a 13900K with a 4090 runs extremely well now with DCS 2.9 MT in VR, compared to the previous states. Every setting maxed out, DLSS enabled. I get between 60-80 fps (at resolution 3120x3120 per eye which is the middle standard for common VR panels today which feature a resolution around 2Kx2K per eye).
I was there from the beginning, going through several different hardware setups over the years and i am very happy atm.

 

i did not post that chart, so don't have the power to remove it. would also be rude;)

thanks for your reply. with my new pc i would want to aim at a similar experience to what you describe. i'll go cheaper on the GPU, because i can easily upgrade later and will use lower pixel density or more aggressive DLSS until i can afford a more powerful GPU. But that's the only real compromise i'm willing to make. I'm planning to use the CPU for many years (i'm still on a highly overclocked second gen intel i7-860... well, yeah...), so it should have some headroom ideally.
My big concern is: Are you CPU limited or GPU limited with you setup and settings? I would like to go big on CPU and RAM speed to get a system that is potentially capable of high quality, 90fps VR (disregard GPU limitation, where i'm willing to make a compromise).
It's of course difficult to see how CPU requirements will change with the dynamic campaign (I'm strictly COOP PvE and SinglePlayer, so i'm more concerned about that than those super complex public MP missions), but i would hope that ED can put AI on different threads before the dyn campaign arrives and even creates some sort of AI instancing to help with those big, dynamic scenarios.

TLDNR: Did MT and latest gen CPUs "solve" DCS HQ 90hz VR?

Link to comment
Share on other sites

4 hours ago, maxsin72 said:

Here 7800x3d and 4090 with reverb g2: 90 fps on multiplayer dogfight server, cpu frame time about 3ms both with or without dlss. That chart tells the truth😀

Thanks for the reply... although i still strongly disagree with the chart... or with the validity of the chart regarding DCS. That doesn't mean that i don't believe you getting very good eprformance from your CPU. But there is not neccessarily direct causality between the chart and your experience. By the way - and i'm not trying to be mean - your post sums up the "problematic" nature of anecdotal evidence in this forum regarding performance of different pc parts: You have a great experience with your CPU, but that prooves nothing for those who want to truly understand how different parts perform in comparison to each other.

MS Flight Simulator is not a good stand-in for DCS. 1080p flatscreen is not a good stand-in for HQ VR.

Your feedback is still valuable though!


Edited by twistking
Link to comment
Share on other sites

1 hour ago, RealDCSpilot said:

I personally prefer demanding servers like 4YA with 1000+ ground and air units and 40+ players. Having good fps on a server with nothing much else going on is pointless for any comparison.

When you have 30 players +3/4 AI planes flying at the same time and cpu frame time is still 3ms , cpu matters a lot, i've seen it after updating my system.

Just one question, which is your cpu frame time?

41 minutes ago, twistking said:

Thanks for the reply... although i still strongly disagree with the chart... or with the validity of the chart regarding DCS. That doesn't mean that i don't believe you getting very good eprformance from your CPU. But there is not neccessarily direct causality between the chart and your experience. By the way - and i'm not trying to be mean - your post sums up the "problematic" nature of anecdotal evidence in this forum regarding performance of different pc parts: You have a great experience with your CPU, but that prooves nothing for those who want to truly understand how different parts perform in comparison to each other.

MS Flight Simulator is not a good stand-in for DCS. 1080p flatscreen is not a good stand-in for HQ VR.

Your feedback is still valuable though!

 

You are welcome, just one question which is your cpu frame time?

Thank you


Edited by maxsin72
Link to comment
Share on other sites

5 minutes ago, maxsin72 said:

When you have 30 players +3/4 AI planes flying at the same time and cpu frame time is still 3ms , cpu matters a lot, i've seen it after updating my system.

Just one question, which is your cpu frame time?

You are welcome, just one question which is your cpu frame time?

Thank you

 

Don't know my frametime (and can't test right now), but i exclusively fly hand-made SP and Coop scenarios optimized for not punishing the CPU too hard. Also im on a monitor currently (no VR) and while i can hold 60fps mostly even on decent settings, i'm sure that i'm CPU bound 90% of the time. I would guess that this would mean that my cpu frame time is roughly 1/60 of a second? Or do i have my math wrong? I mean my CPU is over 10 years old. I find it astounding that DCS runs so well honestly while looking good (but again, no VR, 1080p, only smaller scale missions).

Link to comment
Share on other sites

47 minutes ago, twistking said:

Don't know my frametime (and can't test right now), but i exclusively fly hand-made SP and Coop scenarios optimized for not punishing the CPU too hard. Also im on a monitor currently (no VR) and while i can hold 60fps mostly even on decent settings, i'm sure that i'm CPU bound 90% of the time. I would guess that this would mean that my cpu frame time is roughly 1/60 of a second? Or do i have my math wrong? I mean my CPU is over 10 years old. I find it astounding that DCS runs so well honestly while looking good (but again, no VR, 1080p, only smaller scale missions).

To remain at least at 60 fps, you need a cpu frame time of not more than 16,66 ms. A 10 year old cpu could probably be the problem.

Link to comment
Share on other sites

  • 3 weeks later...

Just for info, i accidently found CPU frame time of my 13900K. It's between 2 ms to 3 ms. So yeah, not much benefit from a "3D" in the name. Choose whatever best possible CPU and GPU you can get for your budget. Intel or AMD - both deliver very good results with the best Nvidia GPU.

i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, Pico 4, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...