P3D v5 - how is VRAM handling now?

[TurboKen 221]

This is my results shown in another thread, but I wanted to share them here so you can see the potential going to V5. Again V5 has some lighting and weather issues(Storms), but I'm sure they are working on that currently. Please note I was using higher settings on V5 which is why you actually see lower FPS in the same area, but I assure you there is a noticeable gain going to V5, and as you can see from my results once you go to a more powerful video card you will be rewarded massively in V5.

Prepar3d V4
SEATAC w/ mostly high settings and ORBX PNW and DD Seattle
RTX2080  34FPS
RTX3090  36FPS
DT Seattle
RTX2080  40FPS
RTX3090  45FPS

Prepar3d V5
SEATAC w/ mostly higher settings and ORBX True Earth
RTX2080  31FPS
RTX3090  50FPS
DT Seattle 
RTX2080  35FPS
RTX3090  55FPS

Here it really shows the work LM have done with off loading much of the work to the GPU. V5 w/True earth is really stunning and the smoothness is something to be seen.

Edited February 26, 2021 by TurboKen

[Ray Proudfoot 8,903]

@TurboKen, I agree they’re very impressive numbers. There’s just one tiny problem. I’d have a greater chance of buying a flux capacitor than a series 20 or 30 graphics card. I’m amazed you’re not running in 4K with a 3090.

My humble 1080Ti will have to make do for the foreseeable future.

Your signature shows you still have v4.5. Any reason why?

[Christopher Low 5,324]

The framerates with an RTX 2080 are actually lower in P3Dv5 than P3Dv4 when using high density scenery? Does this mean that any graphics cards less powerful than this would also result in lower framerates in P3Dv5?

Edited February 27, 2021 by Christopher Low

[Afterburner 373]

The good thing about LM is that you can take advantage of their 14-day refund policy. You can buy P3Dv5, install it to have it side by side with P3Dv4 and compare both with regard to their performance. (And if you use the XML method for add-on sceneries, it is quite easy to run them on both sims). If you like the sim, you can keep it, if not, remove it and apply for a refund...

[Ray Proudfoot 8,903]

I've seen enough reports indicating there is a significant performance boost in v5. DX12 processes graphics more efficiently resulting in around 20% so for 50GBP it seems a logical choice.

Yes, add-on scenery is easy to import but I still have a lot of scenery.cfg stuff but I'm well versed in that. Just a case of deciding when.

[kevinfirth 1,392]

6 hours ago, Christopher Low said:

The framerates with an RTX 2080 are actually lower in P3Dv5 than P3Dv4 when using high density scenery? Does this mean that any graphics cards less powerful than this would also result in lower framerates in P3Dv5?

Not necessarily, it could be quite settings dependent as v4 settings are not equivalent to v5 settings

[GSalden 4,650]

41 minutes ago, Ray Proudfoot said:

I've seen enough reports indicating there is a significant performance boost in v5. DX12 processes graphics more efficiently resulting in around 20% so for 50GBP it seems a logical choice.

Yes, add-on scenery is easy to import but I still have a lot of scenery.cfg stuff but I'm well versed in that. Just a case of deciding when.

Yes, v5 gives a good performance boost with a fast graphics card as with DX12 there is a little less load on the cpu and more on the gpu.

In my case :

V4 / AG both Dense / LOD High / only clouds + terrain shadows / 25 hertz

V5 / AG both Very Dense / LOD Max + HR text / clouds + terrain + buildings shadows / 30 hertz

[Ray Proudfoot 8,903]

3 minutes ago, GSalden said:

Yes, v5 gives a good performance boost with a fast graphics card as with DX12 there is a little less load on the cpu and more on the gpu.

I suppose it all boils down to whether the 1080Ti is still considered a “fast graphics card”. 🤔 After all, it only cost around 750GBP 2.5 years ago.

[GSalden 4,650]

1 hour ago, Ray Proudfoot said:

I suppose it all boils down to whether the 1080Ti is still considered a “fast graphics card”. 🤔 After all, it only cost around 750GBP 2.5 years ago.

That is absolutely correct.

About 2,5 ya I bought 2 1080Ti cards, 1,5 ya I was able to sell them well and bought 2 2080Ti cards. One new and one second handed.

Half a year ago I sold the 2080Ti cards, bought a new 3090 and 2 months later a second handed one.

Two months ago I bough 2 1Tb M2 SSD’d extra as the others ( 2 2Tb SSD ) where filled up to 85%.

🥳

[Noel 4,483]

28 minutes ago, GSalden said:

That is absolutely correct.

About 2,5 ya I bought 2 1080Ti cards, 1,5 ya I was able to sell them well and bought 2 2080Ti cards. One new and one second handed.

Half a year ago I sold the 2080Ti cards, bought a new 3090 and 2 months later a second handed one.

Two months ago I bough 2 1Tb M2 SSD’d extra as the others ( 2 2Tb SSD ) where filled up to 85%.

🥳

Good Lord!  No wonder we can't buy a 3090 GSalden bought them all up!

[Ray Proudfoot 8,903]

@GSalden, a man who lives on the cutting edge of technology. I wish I’d gone for a 3090 when they were available although I have read that some have expired which is concerning.

[GSalden 4,650]

FYI : at one time I 3  3090s... 🤠

Inno3D X3 for my left view client

Inno3D X4 for my server

And I was able to purchase a AORUS Xtreme. The idea was to put that one in my server and return the X4. However, when I tried the AORUS it was broke and they did not have another one.

However, I am fully satisfied with both current cards... 😁

regards, Gerard

 

[neilhewitt 493]

On 2/27/2021 at 6:54 AM, Christopher Low said:

The framerates with an RTX 2080 are actually lower in P3Dv5 than P3Dv4 when using high density scenery? Does this mean that any graphics cards less powerful than this would also result in lower framerates in P3Dv5?

TurboKen mentioned he was using higher settings in P3D v5 and TrueEarth vs non-TE. So it's not a like-for-like comparison. I think the point was that he can manage near enough the same FPS with higher settings and more complex scenery in v5. Which tracks with my experience. I've not found v5 to be slower than v4 in any respect, short of having literally all sliders right including EA (which v4 of course never had).

Which is not to say that I get huge FPS with all sliders right in highly complex scenery with complex add-ons; I don't. But I get better FPS than I would in v4 in the same scenario. And with careful trimming of the settings I have got good visuals and very flyable FPS in London and other high-demand areas which I just could not do in v4.5.

[TurboKen 221]

On 2/25/2021 at 4:38 AM, Ray Proudfoot said:

So other than eating more VRAM what makes DX12 so much better than DX11?

The response from ASOBO was once it moved to DX12 it will give them the ability to add more eye candy. LOL more you say....

[TurboKen 221]

On 2/26/2021 at 10:54 PM, Christopher Low said:

The framerates with an RTX 2080 are actually lower in P3Dv5 than P3Dv4 when using high density scenery?

Sorry that's not true. I did mention above the results shown that I was using a higher scenery and draw distance settings then I was on V4. However that was simply to show the power of V5, I should have spent more time making the settings closer to equal. Unfortunately I sold my RTX2080 so I cannot perform anymore side by side testing. But there was probably 15% or better improvement between the two. With more Vram even more gains then that.