Comparable FPS results

[Suchacz]

Hello my sim pilot colleagues :pilotfly:

 

many people here are discussing their FPS results in BS2, but their lonesome numbers are useless for others, it has no comparable value, because you don't know their hardware, graphics settings and resolution and part of game where it was measured.

 

So i decided to make a downloadable track, which can be used for common FPS measurement.

 

Procedure

 

1/ download my track

2/ start BS2

3/ go to graphic settings and set it to HIGH

4/ load my track

5/ right after the start of the track press "RCtrl+Pause/Break" to see FPS rate in top left corner of your screen. You can also use FRAPS

6/ write down FPS values when the circle is drawn around it with cursor

- now we have same track playing and same graphic settings :thumbup:

 

You will get four values

 

A/ cold & dark machine

B/ machine with running engines

C/ ...with map on Abris

D/ ...with image on Shkval

 

7/ last and most important: Post your results here and don't forget to mention your hardware!

 

My results:

 

A/ 23

B/ 20

C/ 19

D/ 18

 

HW: Phenom II X4 955 BE @ 3.2 GHz, 4GB Ram, Ati 4870 512MB VRam, res. 1680x1050, W7 x64

 

Edit: Don't forget to mention also your Windows vesion (like I did :smilewink:)

 

Edit 2:

My new Sapphire 7870 OC just arrived today, so this test was the first thing, that i made ;)

 

My results #2

 

- same settings

A/ 34

B/ 31

C/ 28

D/ 24

 

- everything maxed out

A/ 31

B/ 28

C/ 26

D/ 21

 

As you can see, if I want to have better results, I must upgrade also my CPU, I think that i72600 is the right choice...

 

Edit 3:

 

BS2 track doesnt work in World anymore. But stay tuned, I will make a new one, maybe SOON(*), maybe LATER(*)

 

(*) - ED trademark :D

 

 

But I can tell you, that there is no difference between new and old Black Shark

graphic test.trk

Edited May 19, 2012 by Suchacz

[spikenet]

A/ 40

B/ 39

C/ 38

D/ 34

 

2500K @4.6 + ATI6950 2gig res 1980x2280 on, win7 x64

 

Just upgraded couple of days ago from Q6600, quite a big difference! Interesting the FPS drop running map, I didnt realise this b4.

 

cheers

 

Edit: FYI My BS2 has the high detail and grass mod as well as running eyefinity dual monitor.

Edited December 23, 2011 by spikenet
update info

[Suchacz]

Wow, what a great results!

 

The Abris map is also 3D, but it has zero height. So it takes some graphic computing time...

[213]

if it has zero height, then it's 2d.

[BS_Fighter]

A/ 45

B/ 38

C/ 38

D/ 35

 

2600K@4,8Ghz, 8GB Corsair Vengeance, GTX580 AMP2 (1920x1080*8xMSAA*TSAA*16xAF with NV-Inspector), Win 7 64Bit.

[Suchacz]

if it has zero height, then it's 2d.

 

 

In 2D space objects have two dimensions, in 3d space has objects three dimensions, even if one of them equals zero. Its a little difference :smartass:

Edited January 4, 2012 by Suchacz

[Xxx]

a) 60

b) 41

c) 40

e) 39

 

For myself, the biggest test of "frame rate" is the perceived "smoothness", ie, the effect the movement in the sim has upon the human eye. We are not supposed to see anything faster than about 25 fps. We do notice frame rate variations (stutters).

 

Even with figures like the above, I see "some" stutters with large terrain loads and vehicles are firing? Again, for me its sthe way the sim can deal with this "load" which I look for.;)

 

PS Fraps will slow my machine down....:cry:, its great for recording tracks but always system hungry.

[EtherealN]

Accipter, FRAPS should not slow you down notcieably as long as you don't actually record video - remember, it runs on a separate core and all it does is check how often the screenbuffer gets updated.

[Kunter7]

A/ 45

B/ 38

C/ 38

D/ 35

 

2600K@4,8Ghz, 8GB Corsair Vengeance, GTX580 AMP2 (1920x1080*8xMSAA*TSAA*16xAF with NV-Inspector), Win 7 64Bit.

 

We are very close on this one Fighter:

 

A/ 47

B/ 42

C/ 39

D/ 30

 

i7 2600k - 16 Gig memory - HD6870 (about to move my HD6950 to this box and then OC both the CPU and GPU)

Win 7 64 bit

2 Crucial M4's in RAID O for OS and games.

3 TB in RAID 0 for video editing/Data

1 TB E-SATA for backups

Saitek X52 Pro HOTAS

Saitek Pedals

Track IR5

LG Monitor @ 1920x1080

[spikenet]

A/ 47

B/ 42

C/ 39

D/ 30

 

a) 60

b) 41

c) 40

e) 39

 

 

 

Quite a massive drop between A and D.

 

I heard Peter Jackson talking about perceived framerate, he says the eye can detect 60. They are shooting "the hobbit" at 48fps at 5k. He said its not perfect but thats where our technology is.

 

 

Accipiter, whats your system spec?

[EtherealN]

Accipiter, whats your system spec?

 

Try his signature. ;)

 

Regarding eye detection of FPS, there is no such thing as a specific FPS that the human eye can perceive. Everything depends on context, and the big deal is the brain, not the eye. In some circumstances you cannot see the difference between 24 and 60 fps, in other cases even 60 will look stuttery.

Edited December 23, 2011 by EtherealN

[213]

in the case of computer monitors, 60 is ample. any more is a waste of resources. ka-50 looks and plays smoothly capped at just 30.

 

people who claim they can notice differences up to 100 fps are of course liars and ignoramuses.

[BS_Fighter]

We are very close on this one Fighter:

 

A/ 47

B/ 42

C/ 39

D/ 30

 

i7 2600k - 16 Gig memory - HD6870 (about to move my HD6950 to this box and then OC both the CPU and GPU)

Win 7 64 bit

2 Crucial M4's in RAID O for OS and games.

3 TB in RAID 0 for video editing/Data

1 TB E-SATA for backups

Saitek X52 Pro HOTAS

Saitek Pedals

Track IR5

LG Monitor @ 1920x1080

 

I had my ultra high.lua for this test still active :doh:.

With the original lua:

 

A/ 50

B/ 44

C/ 43

D/ 36

[Kuky]

I'll chuck mine in also

 

In game settings:

Textures ....... HIGH

Scenes ......... HIGH

Civ Traffic ...... OFF

Water ........... HIGH

Vis Range ...... HIGH

Heat Blur ........ OFF

Shadows ....... HIGH

MSAA ............. 4X

HDR .............. OFF

Trees .........9000m

Cockpit shadow ON

Tree shadow .. OFF

TSSAA .......... OFF

Mirrors .......... OFF

 

Resolution 2720x1200 (3 monitors) - Grass/Bushes 300m (10x denser/mod) - Res of Cockpit Displays: 1024 every frame

A) HD5870: 26 HD7970: 42

B) HD5870: 24 HD7970: 39

C) HD5870: 23 HD7970: 36

D) HD5870: 19 HD7970: 30

 

Resolution 1920x1200 (1 monitor) - Grass/Bushes 300m (10x denser/mod) - Res of Cockpit Displays: 1024 every frame

A) HD5870: 28 HD7970: 42

B) HD5870: 26 HD7970: 40

C) HD5870: 26 HD7970: 37

D) HD5870: 20 HD7970: 31

 

Resolution 1920x1200 (1 monitor) - Grass/Bushes 300m (10x denser/mod) - Res of Cockpit Displays: 512

A) HD5870: 28 HD7970: 42

B) HD5870: 26 HD7970: 40

C) HD5870: 26 HD7970: 38

D) HD5870: 23 HD7970: 35

 

Resolution 1920x1200 (1 monitor) - Grass/Bushes 0m - Res of Cockpit Displays: 512

A) HD5870: 35 HD7970: 49

B) HD5870: 32 HD7970: 43

C) HD5870: 31 HD7970: 42

D) HD5870: 26 HD7970: 37

 

PS: running single monitor Windowed Mode vs Full Sceen I get same FPS

 

HD5870 1GB, i7 950 @ 4.0GHz, 12GB DDR3, Widows 7 64bit Home Premium

HD7970 3GB, i7 950 @ 4.2GHz, 12GB DDR3, Widows 7 64bit Home Premium

Edited February 12, 2012 by Kuky

[Spetz]

With everything as per the test instructions:

 

a) 22

b) 20

c) 20

d) 18

 

With my usual settings:

 

a) 28

b) 25

c) 25

d) 23

 

GTX 260 (768mb), E8400 Oc'd to 3.6 Mhz, 8GB Ram, Win7 Ultimate x64, Corsair 850 watt PSU, running at 1920x1200.

[Suchacz]

I heard Peter Jackson talking about perceived framerate, he says the eye can detect 60. They are shooting "the hobbit" at 48fps at 5k. He said its not perfect but thats where our technology is.

 

You and Peter are right. From 24-25 fps above, human eye will see a fluent image, but you will also see the difference between 25 and 60... :smilewink:

 

Btw, what does the MAJOR and minor letters mean?

[HiJack]

In game settings:

Textures ........: HIGH

Scenes ..........: HIGH

Civ Traffic ......: OFF

Water ............: HIGH

Vis Range .......: HIGH

Heat Blur ........: OFF

Shadows .........: HIGH

MSAA ..............: 8X

HDR ................: OFF

Clutter/Bushes..: 0m

Trees ...............:9000m

Cockpit shadow.: ON

Tree shadow ....: OFF

TSSAA ............: OFF

Mirrors ............: OFF

 

ATI Radeon HD 5870 3D application settings set to default profile: High Quality

I am also using the D3DOverrider v2.0.1 to force VSync.

 

Resolution 1920x1080 - Res of Cockpit Displays: 512 every frame

A) 26

B) 24

C) 22

D) 19

 

Computer spec:

Casing: ASUS M4A89TD PRO, Antec Nine Hundred Two, Passive cooling on board and CPU

Processor: AMD Phenom II X6 1090T Processor, 3200 Mhz, 6 Core(s), 6 Logical Processor(s)

Memory: 8GB RAM, 4 x OCZ DDR3 2 GB Value Memory Gold i5, 1600MHz. (8-8-8-24)

Display adapter: ATI Radeon HD 5870, 1GB RAM, Catalyst 12.1 Graphics Software

System disk: OCZ 120 GB SSD 2,5" S-ATA II TRIM/GC, Vertex 2, Max Read/Write 285/275MB/sec

Storage disk: WD 2TB Black 3,5", SATA64MB Cache, Dual Processor, 7200RPM

Operating System: Windows 7 64-bit English, Trend Micro Internet Security antivirus

Monitor: Samsung 27" LCD Syncmaster P2770H, 1920x1080, DVI, 70000:1, 2ms

Flight gear: Saitek X52Pro, TrackIR 5, VAC, TS3, Fraps

 

(HJ)

Edited December 31, 2011 by HiJack

[Xxx]

EtherealN, thanks for that. Was not aware Fraps ran on a separate core, for our DCS. I know it annoys me greatly in FSX:doh:

 

Spikenet sys. specs in my sig.

[Wyverex]

Civ Traffic: Off

Resolution: 1920 x 1080

Displays: 512 every frame

MSAA: 8x

TSSAA: On

Cockpit Shadows: On

HDR: Normal

Shadow Tree: On

Trees: 9000 m

 

a) 29

b) 26

c) 25

d) 20

 

Core2 Quad 9550 @2.83 GHz, MSI GTX570 TwinFrozr III, 4GB RAM, Win7 x64

 

I was a bit stunned that the framerate is that low so I ran a CPU and GPU monitor alongside. I found that at the specified times, total CPU usage was between 36-39% and total GPU usage about 50%. Those values never climbed above that (only below in the fast-forward section).

 

So apparently BS2 doesn't utitilize the full power that is given. I suppose my system could easily handle 50-60 frames given that configuration.

 

Can anyone confirm that usage pattern?

[Suchacz]

So apparently BS2 doesn't utitilize the full power that is given. I suppose my system could easily handle 50-60 frames given that configuration.

 

Can anyone confirm that usage pattern?

 

Multi-core usage http://forums.eagle.ru/showthread.php?t=82567

[Wyverex]

Ah didn't know that. Ok then BS2 uses only two cores. That still doesn't explain why neither those cores nor the GPU is topped out. Today's games are usually GPU-bound (with some exceptions being CPU-bound).

Maybe BS2 doesn't use the full capability of my graphics card but the features that are used are used at 100% (that could explain why the GPU usage is pretty constant at 56-58%).

 

Intriguing...

[diveplane]

TECHNICAL

 

Q: Does DCS: A-10C use all cores on my CPU?

A: Not fully. It does include the new sound engine which operates in a separate thread, and does take advantage of some DirectX threading features when running on Windows Vista or 7.

[Suchacz]

Ah didn't know that. Ok then BS2 uses only two cores. That still doesn't explain why neither those cores nor the GPU is topped out. Today's games are usually GPU-bound (with some exceptions being CPU-bound).

Maybe BS2 doesn't use the full capability of my graphics card but the features that are used are used at 100% (that could explain why the GPU usage is pretty constant at 56-58%).

 

Intriguing...

 

I agree. I know something about programming, so I can say, that well-programmed aplication uses all resources it can get. I hope that ED will work on optimalisation in some future patches...

[Wyverex]

I happen to be a game developer myself so I have a particular interest in this topic :)

Did some more testing with various settings. Even when setting everything as low as it can be, the frame rate is only between 35 and 54 frames per second.

At the same time, GPU usage still goes down (40%) so we can definitely rule out that BS2 is GPU-bound (which can be expected since sims generally don't put that much stress on the GFX card).

Resolution also doesn't have a noticeable impact, so no fill-rate problems.

It rather looks like being somewhat CPU-bound. At least one of the CPUs goes up to about 85-90% usage which is probably the main thread. Still no explanation why it doesn't utilize 100%. Might be that BS2 is cache-unfriendly and sends large chunks of memory around, so that the CPU spends a considerable time waiting for memory transfers. Or there's just a hidden sleep() somewhere :lol:

 

Bottom line is we'll probably never know but this is definitely something that should be addressed. Getting only 20-30 frames on a quad core that isn't fully utilized is... improvable, to put it mildly :smartass:

[spikenet]

Thought I read somewhere from Wags that the engine is being re-written for dx11 ?