Jump to content

2.5 Settings Comparisons and GPU Impact


Sideslip

Recommended Posts

Great topic... and good job.

 

Looking at comparisons, and making decisions with the best FPS / RAM results from these examples i got smooth DCS with 45-60fps. One more time, smooth on my GTX780ti.

 

 

Thanks!

 

I need to clean up the initial post to make it clear, but don't read too much into the RAM amounts. I had Photoshop running at the same time (and music) to make organizing the shots quicker, so the RAM was probably fluctuating for reasons other than DCS. I have never observed more than 12GB usage from DCS alone, generally around 9 or 10GB. This was also true in multi-player, but only on the aerobatics server with a dozen players and only for 10-15 minutes.

System specs: i7 3820 @4.75Ghz, Asus P9X79LE, EVGA GTX1080SC @2100mhz, 16GB Gskil DDR3 @ 2000mhz, 512GB 960EVO m.2, 2 X 512GB 860EVO SATA3 in RAID0, EVGA Supernova 850W G2, Phantek Entho Luxe White. CPU and GPU custom water-cooled with 420mm rad and lots of Noctua fans.

ASUS PG348Q. VKB Gladiator Pro w/MCG, X-55 throttle and MFG Crosswind.

Link to comment
Share on other sites

I need to clean up the initial post to make it clear, but don't read too much into the RAM amounts....

 

Don`t worry... Basically, you gave me good guidance. :thumbup:

Quote

Немој ништа силом, узми већи чекић!

MSI Tomahawk MAX | Ryzen 7 3700x | 32GB DDR4 3200MHz | RX 5700 XT OC Red Dragon 8GB | VPC Throttle CM3 + VPC Constellation ALPHA on VPC WarBRD Base | HP Reverb G2

 Youtube Follow Me on TWITCH! 

Link to comment
Share on other sites

Important Update: I read that you should delete the "metashaders" and "fxo" folders from your saved games DCS folder. Doing so reduced my vram usage from 8GB to 5GB. I have not noticed a difference in performance otherwise and the visual comparisons should be unchanged, but the vram shown in the comparison images may not be accurate.

 

Where I read to delete those folders: https://forums.eagle.ru/showthread.php?t=200906


Edited by Sideslip

System specs: i7 3820 @4.75Ghz, Asus P9X79LE, EVGA GTX1080SC @2100mhz, 16GB Gskil DDR3 @ 2000mhz, 512GB 960EVO m.2, 2 X 512GB 860EVO SATA3 in RAID0, EVGA Supernova 850W G2, Phantek Entho Luxe White. CPU and GPU custom water-cooled with 420mm rad and lots of Noctua fans.

ASUS PG348Q. VKB Gladiator Pro w/MCG, X-55 throttle and MFG Crosswind.

Link to comment
Share on other sites

Again...outstanding work. I know how long it takes to line up the shots, stitching it together etc. One quick tip. Even if you don't need audio, record it. This way, every time you need a reference point, you clap once. It causes a visible spike in the audio track and makes it very easy to line up the exact point in time between the various tracks. I do this when I create Wireshark videos. If I make a mistake, I just keep rolling, but I clap to demarc what needs to be cut. Although you may already be doing it seeing as how everything is nicely lined up. awesome job.

hsb

HW Spec in Spoiler

---

 

i7-10700K Direct-To-Die/OC'ed to 5.1GHz, MSI Z490 MB, 32GB DDR4 3200MHz, EVGA 2080 Ti FTW3, NVMe+SSD, Win 10 x64 Pro, MFG, Warthog, TM MFDs, Komodo Huey set, Rverbe G1

 

Link to comment
Share on other sites

Outstanding analysis Sideslip! There are a lot of varying facts and opinions on these boards concerning quality and performance gains. This demystifies the situation.

 

I put together a VR specific settings guide: 2.5 Video Settings - VR Quality/Performance Balance in the Virtual Reality thread. Of course, there are different considerations for VR, but I've linked your post in my OP so people can look at base line visual comparisons.

 

Thanks for taking the time to do this for the community!


Edited by StrongHarm

It's a good thing that this is Early Access and we've all volunteered to help test and enhance this work in progress... despite the frustrations inherent in the task with even the simplest of software... otherwise people might not understand that this incredibly complex unfinished module is unfinished. /light-hearted sarcasm

Link to comment
Share on other sites

Great topic... and good job.

 

Looking at comparisons, and making decisions with the best FPS / RAM results from these examples i got smooth DCS with 45-60fps. One more time, smooth on my GTX780ti.

 

blIkBP7.png

 

Thanks!

 

Thanks, it's good.

My computer : i5-2500, GTX960, 8 GB RAM.

 

Anyway, I edited it a bit.

c625ab06c396.jpg

 

I set MSAA from 8x to 4x because there's no visible difference and takes less fps.

Terrain Objects Shadows - I turned it off because I've not got fps drops now.

AF set from 2x to 4x because now the runway looks better, sharper.

 

 

Usually I've got about 45 fps. Sometimes 60, sometimes about 30, but without fps drops.


Edited by Valium
Link to comment
Share on other sites

Ok ,i Want to say that fps aré stongest impact with high resolution monitor and then your test

Is limited for your resolution.and it is not benchmarking with others

 

Perhaps.

So in your opinion if the resultion is higher, all these graphic settings take more fps?

 

Once I changed my resolution in game to 1024 x ... but I even lost a few fps instead of gaining them.

Link to comment
Share on other sites

For those who are getting different results than others

 

I have finally concluded what should have always been obvious: Ultimately your performance is either limited by your CPU or GPU (I won't go into other system performance like memory or HDD). Think of it as 2 guys running a race but tied together by a rope. If one is slow, he limits the speed of the other. In DCS, when the CPU gets maxed out then the gpu becomes underutilized and vice versa.

 

Which is why people here get different results when playing with different settings (for some, the video card might be holding them back and for others it is their cpu) AND the limiting factor might change in different game conditions - it is dynamic.

 

e.g. On my rig, while sitting on the runway at Batumi my cpu was underutilized and gpu was the limiting factor (using SLI and reducing eye candy settings could give me a huge performance boost). However, once I flew to a different area and entered intense combat (lots of objects, AI, TGP, etc.) then one or two of the cpu cores became maxed out and the gpu was underutilized (eye candy settings and SLI did NOT impact fps); however reducing cpu load (changing visual range or overclocking the cpu) would boost performance.

 

So, ultimately, your individual fps performance will come down to what is limiting your overall fps (cpu or gpu?)

 

If it is cpu, then change settings that impact cpu load (i.e. visual range, etc.)

If it is gpu, then change settings that impact gpu load (i.e. eye candy or use SLI if you have it).

 

edit: another perspective, if you are already cpu maxed, then you can probably start cranking up the graphics settings without lowering your fps

 

You can determine the limiting factor under different conditions by using various monitoring tools. (I use EVGA PrecisionX, HWmonitor, etc.)


Edited by prccowboy
Link to comment
Share on other sites

Valium, this is probably because you're loading textures and objects into memory as you encounter them. I like to speed this along by hitting F11 to change view to a few airfields and spin my point of view around before taking off. This used to be very useful several years ago before ED optimized object load.

It's a good thing that this is Early Access and we've all volunteered to help test and enhance this work in progress... despite the frustrations inherent in the task with even the simplest of software... otherwise people might not understand that this incredibly complex unfinished module is unfinished. /light-hearted sarcasm

Link to comment
Share on other sites

Wow should be stick for further reference. Thank you for your hard work Slideslip

 

Once we are on the subject I have a few questions on the graphic setting that some might have the answer.

 

Heat Blur. Do we have any penalty using them?

 

Res of cockpit displays. What is it doing? I never seen any difference and if so, again what is the penalty.

 

Depth of field Another area where I can't see jack shit as difference, visually nor performance wise.

 

 

Disable aero interface Same here.

 

 

o7


Edited by Decibel dB
Link to comment
Share on other sites

Valium, this is probably because you're loading textures and objects into memory as you encounter them. I like to speed this along by hitting F11 to change view to a few airfields and spin my point of view around before taking off. This used to be very useful several years ago before ED optimized object load.

 

Thank you, I'm going to check this. Perhaps, it'll help.

 

 

 

Yes, hear blur takes fps when it's on. When it set on high, it takes more.

Anyway, it doesn't take too much fps in my opinion. Sometimes when you're staying at the airport, got heat blur on high and turn on the external camera, you may have fps drops or something like that so my suggestion is to set in on low.

 

Depth of field - I haven't tested it in 2.5 but even in 1.5.8 it took too much.

Link to comment
Share on other sites

Very interesting post sideslip!

 

I've done some math, tell me if I'm wrong:

 

- You play 3440x1440=4953600 pixels and with your settings you obtain an average 70 fps with a gtx 1080.

- I have a gtx 1070. If you consider 1080 an average 25% faster than 1070 (this is approximately what all benchmarks show), I should obtain about 56 fps at your same resolution and with your same settings (125:70 = 100:X; X=56).

- But I play in VR with oculus rift, so I have 1080x1200 pixels per eye that is a total resolution of 2160x1200=2592000 pixels. That is 52.3% of your resolution, just over half of yours!

- Considering that I need at worst 45 solid fps to avoid stuttering with rift (assuming a pixel density = 1), I should be able to set your settings even with a pixel density of 1.4 - 1.6 without loosing fps, or alternatively keep pixel density = 1 but get solid 90 fps.

- With an increased pixel density maybe I should set no AA, since supersampling is a sort of antialiasing method (is it worth to use both?) and thus gaining another bunch of fps.

 

Are these maths correct? Now I should test...


Edited by nessuno0505
Link to comment
Share on other sites

Very interesting post sideslip!

 

I've done some math, tell me if I'm wrong:

 

- You play 3440x1440=4953600 pixels and with your settings you obtain an average 70 fps with a gtx 1080.

- I have a gtx 1070. If you consider 1080 an average 25% faster than 1070 (this is approximately what all benchmarks show), I should obtain about 56 fps at your same resolution and with your same settings (125:70 = 100:X; X=56).

- But I play in VR with oculus rift, so I have 1080x1200 pixels per eye that is a total resolution of 2160x1200=2592000 pixels. That is 52.3% of your resolution, just over half of yours!

- Considering that I need at worst 45 solid fps to avoid stuttering with rift (assuming a pixel density = 1), I should be able to set your settings even with a pixel density of 1.4 - 1.6 without loosing fps, or alternatively keep pixel density = 1 but get solid 90 fps.

- With an increased pixel density maybe I should set no AA, since supersampling is a sort of antialiasing method (is it worth to use both?) and thus gaining another bunch of fps.

 

Are these maths correct? Now I should test...

 

You maths forgot something fundamental: With VR you are not only changing resolution, but you also render the scene twice from two points of views, this change almost everything...

Link to comment
Share on other sites

Wow should be stick for further reference. Thank you for your hard work Slideslip

 

Once we are on the subject I have a few questions on the graphic setting that some might have the answer.

 

Heat Blur. Do we have any penalty using them?

 

Yes there is, but I have yet to measure it. You will only get a performance drop if it is actively blurring, which you wont see unless you are fairly close to another aircraft (A2A refuel, formation, maybe really close dog fighting). So the majority of the time it wont hurt FPS.

 

Res of cockpit displays

What is it doing? I never seen any difference and if so, again what is the penalty.

 

This is the resolution of things like the Shival TV. If you use trackir and lean into the screen, or zoom in on it, it will look better with a higher resolution. 512 is fine to use it as intended. The "per frame" I think refers to the framerate of the display. Don't quote me on this, but I think if it is not set to "per frame" it might run at like half frame rate or something. Again, really doesn't matter considering what it's designed for. Performance impact can be very high depending what you are looking at through the TV.

 

 

Depth of field

Another area where I can't see jack shit as difference, visually nor performance wise.

 

Do not use this setting. It does nothing in the cockpit, it only affects your exterior views. It blurs the SH*T out of everything. It also tanks my FPS from 70-80 to 20-30. It's only use is for making cinematic videos. There must be a way of controlling the depth, but I've never cared to look.

 

 

Disable aero interface

Same here.

 

 

o7

 

Just ignore this. I imagine this is an old setting from when Windows Vista first released (2005?). I can't believe that the aero interface (makes the edges of windows transparent and stuff) could have any impact on performance if you are even capable of running DCS.

 

Windows Aero Wikipedia


Edited by Sideslip

System specs: i7 3820 @4.75Ghz, Asus P9X79LE, EVGA GTX1080SC @2100mhz, 16GB Gskil DDR3 @ 2000mhz, 512GB 960EVO m.2, 2 X 512GB 860EVO SATA3 in RAID0, EVGA Supernova 850W G2, Phantek Entho Luxe White. CPU and GPU custom water-cooled with 420mm rad and lots of Noctua fans.

ASUS PG348Q. VKB Gladiator Pro w/MCG, X-55 throttle and MFG Crosswind.

Link to comment
Share on other sites

You maths forgot something fundamental: With VR you are not only changing resolution, but you also render the scene twice from two points of views, this change almost everything...

 

Yes, as stated above, the resolution is not 2160x1200 but 1080x1200 TWICE. In DCS settings, next to monitor choose "3 screen". You will have the same resolution as before but it is being split and rendered for 3 different views. You will notice an impact to your FPS. This is also similar to what happens with Mirrors and Res of cockpit displays.

 

However, I know Nvidia was supposed to work some black magic with up to 16 viewports (viewpoints really) with the current 1000 series but I don't know how effective it is or what is required for it to work.


Edited by Sideslip

System specs: i7 3820 @4.75Ghz, Asus P9X79LE, EVGA GTX1080SC @2100mhz, 16GB Gskil DDR3 @ 2000mhz, 512GB 960EVO m.2, 2 X 512GB 860EVO SATA3 in RAID0, EVGA Supernova 850W G2, Phantek Entho Luxe White. CPU and GPU custom water-cooled with 420mm rad and lots of Noctua fans.

ASUS PG348Q. VKB Gladiator Pro w/MCG, X-55 throttle and MFG Crosswind.

Link to comment
Share on other sites

Yes there is, but I have yet to measure it. You will only get a performance drop if it is actively blurring, which you wont see unless you are fairly close to another aircraft (A2A refuel, formation, may really close dog fighting). So the majority of the time it wont hurt FPS.

 

 

 

This is the resolution of things like the Shival TV. If you use trackir and lean into the screen, or zoom in on it, it will look better with a higher resolution. 512 is fine to use it as intended. The "per frame" I think refers to the framerate of the display. Don't quote me on this, but I think if it is not set to "per frame" it might run at like half refresh rate or something. Again, really doesn't matter considering what it's designed for. Performance impact can be very high depending what you are looking at through the TV.

 

 

 

 

Do not use this setting. It does nothing in the cockpit, it only affects your exterior views. It blurs the SH*T out of everything. It also tanks my FPS from 70-80 to 20-30. It's only use is for making cinematic videos. There must be a way of controlling the depth, but I've never cared to look.

 

 

 

 

Just ignore this. I imagine this is an old setting from when Windows Vista first released (2005?). I can't believe that the aero interface (makes the edges of windows transparent and stuff) could have any impact on performance if you are even capable of running DCS.

 

Windows Aero Wikipedia

 

:thumbup: Thank you sir

Link to comment
Share on other sites

Yes, as stated above, the resolution is not 2160x1200 but 1080x1200 TWICE. In DCS settings, next to monitor choose "3 screen". You will have the same resolution as before but it is being split and rendered for 3 different views. You will notice an impact to your FPS. This is also similar to what happens with Mirrors and Res of cockpit displays.

 

However, I know Nvidia was supposed to work some black magic with up to 16 viewports (viewpoints really) with the current 1000 series but I don't know how effective it is or what is required for it to work.

 

In fact with your settings and PD = 1 I obtain quite solid 45 fps and not always 90, maybe that's because 1080x1200 pixels x 2 is different from 2160x1200 pixels, even if the mere pixel count is the same.

Link to comment
Share on other sites

Nessuno0505, where VR is concerned, try kicking your PD up to 1.2, it makes a big difference. I wouldn't go over MSAAx2 though. The considerations for the various settings are very different for VR. Drill into quoted post for VR specific quality/performance analysis.

 

VR PD = 1.2

MSAA = 2x

Anisotropic Filtering = 2x

Deferred Shading = ON

Terrain Shadows = FLAT

Cockpit Global Illumination = ON

 

 

[TABLE]

__PD__

|

_MSAA_

|

_Quality_

|Performance

1|

0

|

0

|

10



1.2

|

0

|

2

|

10



1.4

|

0

|

4

|

10



1.6

|

0

|

4

|

10



1.8

|

0

|

4

|

8



2.0

|

0

|

6

|

6



1

|

2

|

6

|

10



1.2

|

2

|

8

|

8



1.4

|

2

|

8

|

6



1.6

|

2

|

10

|

4



1.8

|

2

|

10

|

2



2.0

|

2

|

10

|

2



1

|

4

|

8

|

4



1.2

|

4

|

8

|

4



1.4

|

4

|

8

|

2



1.6

|

4

|

10

|

0



1.8

|

4

|

10

|

0



2.0

|

4

|

10

|

0



1

|

8

|

8

|

2



1.2+

|

8

|

10

|

0

[/TABLE]

 

It's a good thing that this is Early Access and we've all volunteered to help test and enhance this work in progress... despite the frustrations inherent in the task with even the simplest of software... otherwise people might not understand that this incredibly complex unfinished module is unfinished. /light-hearted sarcasm

Link to comment
Share on other sites

  • Recently Browsing   0 members

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