[Official] DX9 vs DX11 performance difference

Status
Not open for further replies.
dadaboomda said:
Stonec... you make a very good job.
You're the user forum of the year ! :)
Click to expand...

Thanks, I do this for fun :). FPS is better in this build, but unfortunately rain effects consume a lot of FPS as well, so most probably don't notice the FPS improvement. I did another quick test with start grid FPS at Silverstone, 20 GT3 cars. What is confusing is the big loss of FPS from dry to rain with raindrops off, I find it hard to understand. FPS difference with raindrop levels is small.

rF2 rain.png
 
Suzukinol said:
mmm... No. Sorry but I don't agree with this. DX11 is having less perfomance in most of the systems using exactly the same settings as DX9, something that reflects that its not properly optimized. At this momment, I can say that the most important work that should be focused on is to fix the FPS issues with the new rain as using Medium Rain Drops is taking 50 or 60% of the fps compared to dry. I know its not easy for you but I don't want better graphics if they will come with less perfomance, give me better physics and I will enjoy them much more than some rain drops in the body of the car.

And... sorry but I think that a R9 280 graphic card which can run 2014-2016 games on ultra settings is beastly enough to move it at least 60 stable fps, specially when we remember this game was launched in 2012.
Click to expand...
the simulation has many mathematical calculations on physics, besides the calculations by graphics, and it is not as easy as for another kind of games, that spend a lot on graphics but little on physics, really everything is mathematical calculations ...
 
stonec said:
What is confusing is the big loss of FPS from dry to rain with raindrops off, I find it hard to understand. FPS difference with raindrop levels is small.
Click to expand...
I'm not sure how you set your "no rain", but clouds consume fps compared to no clouds.
 
stonec said:
Thanks, I do this for fun :). FPS is better in this build, but unfortunately rain effects consume a lot of FPS as well, so most probably don't notice the FPS improvement. I did another quick test with start grid FPS at Silverstone, 20 GT3 cars. What is confusing is the big loss of FPS from dry to rain with raindrops off, I find it hard to understand. FPS difference with raindrop levels is small.

View attachment 11676
Click to expand...

Yeah, rain fps is a disaster, unfortunately :(
 
Suzukinol said:
mmm... No. Sorry but I don't agree with this. DX11 is having less perfomance in most of the systems using exactly the same settings as DX9, something that reflects that its not properly optimized. At this momment, I can say that the most important work that should be focused on is to fix the FPS issues with the new rain as using Medium Rain Drops is taking 50 or 60% of the fps compared to dry. I know its not easy for you but I don't want better graphics if they will come with less perfomance, give me better physics and I will enjoy them much more than some rain drops in the body of the car.

And... sorry but I think that a R9 280 graphic card which can run 2014-2016 games on ultra settings is beastly enough to move it at least 60 stable fps, specially when we remember this game was launched in 2012.
Click to expand...
If I run the same exact settings in dx11 as I did in dx9,regardless of fps the image still looks way better than dx9 did.
So dx11 seems to be running better graphics than dx9 with the same settings on my system and with my eyes.
From what you’re saying,dx11 doesn’t run any improved graphics when set to the same graphics settings as dx9.
Is that true because my eyes are telling me different

Example dx9 min settings vs dx11 min settings
 
Suzukinol said:
mmm... No. Sorry but I don't agree with this. DX11 is having less perfomance in most of the systems using exactly the same settings as DX9, something that reflects that its not properly optimized. At this momment, I can say that the most important work that should be focused on is to fix the FPS issues with the new rain as using Medium Rain Drops is taking 50 or 60% of the fps compared to dry. I know its not easy for you but I don't want better graphics if they will come with less perfomance, give me better physics and I will enjoy them much more than some rain drops in the body of the car.

And... sorry but I think that a R9 280 graphic card which can run 2014-2016 games on ultra settings is beastly enough to move it at least 60 stable fps, specially when we remember this game was launched in 2012.
Click to expand...

I'm sorry you can't accept what the dev says (not me remember!), it would be good if you could respond to my earlier request and post your benchmark results from DX9 and then DX11.
 
SPASKIS said:
I know I have to upgrade my hardware. It is already 5 years old. This is more or less what I will be getting.
Click to expand...
That’s a beast of a pc,like it very much,Are you building it yourself ,save a few hundred.
I’m thick as pig shit and found building my pc easy,your a clever bloke so should find it easier.lol
 
patchedupdemon said:
That’s a beast of a pc,like it very much,Are you building it yourself ,save a few hundred.
I’m thick as pig shit and found building my pc easy,your a clever bloke so should find it easier.lol
Click to expand...
I already ordered it. I am not good with PC hardware. Hopefully next week I will be able to drive rF2 with nice graphics and no fps problems any more.
 
Foodie said:
Why cant rF2 have a standard benchmark mode so it will be much easier to compare results? Load a field of stock content and let rF2 do the work?
Click to expand...

It's not much use to have dedicated benchmark when each new piece of content that gets released runs at different rate, and usually is heavier than previous content. For example USF grid of 20 cars at Matsusaka runs around 90 FPS at start, with GT3 cars it's 55 FPS.
 
Rates? No idea what it is. I have no idea what track or car that would suit such benchmark test. Just coming with an idea becuase the amount of different numbers etc in this thread is just not very usable and people get more confused.

The result files could include everything the devs need, like apps running in the background, plugins all that stuff they have to ask for over and over again...
 
Foodie said:
Rates? No idea what it is. I have no idea what track or car that would suit such benchmark test. Just coming with an idea becuase the amount of different numbers etc in this thread is just not very usable and people get more confused.
Click to expand...

Sorry I meant frame rates (FPS). It's a bit pointless to make a benchmark when the next content they release will probably again run at poorer FPS and demand more from the hardware. Frame rate drop from Nissan GT500 to GT3 pack was something around 30-40%, so they'd need to update the benchmark with each new car released and results would not be compatible anymore.
 
I7-3770k. 2x780Ti, PCI- e@3.0 x16, x16 > GT3 at Silverstone, 40 AI, PP Ultra, AA Level 5 all settings in-game settings MAX.
SLI at Nola 2.jpg

Bus loads, 92/85%, RAM, 11533/19439 MB :(
 
@Ari Antero I'm not sure I understand whats the problem.. it seems like good SLI scaling, considering how SLI is not good optimized in RF2. Or are you pointing out at high RAM usage of 11 gb?
 
Ari Antero said:
No 11 gb used of 16 gb, 19gb is vram usage
Click to expand...
Can’t be,780ti only had 3gb vram,and having two doesn’t need mean 6gb vram either.
I’m confused

From that screen shot you are maxing our your gpus vram,of just over 3gb.
Could that mean your gpus are bottlenecking your cpu,which is why it only says 21% usage
 
Prodigy said:
@Ari Antero I'm not sure I understand whats the problem.. it seems like good SLI scaling, considering how SLI is not good optimized in RF2. Or are you pointing out at high RAM usage of 11 gb?
Click to expand...

Does not mater if you are running single card or SLI. Bus usage +-90% and PC ram usage is huge if max settings are used. SLI gives you +40% better performance with DX11 then with DX9.
Reason I made this test is that many are talking about CPU running out of real time but I believe that huge system usage is the reason who so many has problems with rF2 when too high game settings and to may AI`s are used.
 
Status
Not open for further replies.
Back
Top