Caporegime
That's the whole point of the star swarm benchmark. To showcase the limitations of DX11 and the benefits of Mantle. Lots of draw calls and other things that are above my station.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Star Swarm - Oxide bench thread
- Thread starter Gregster
- Start date
That's the whole point of the star swarm benchmark. To showcase the limitations of DX11 and the benefits of Mantle. Lots of draw calls and other things that are above my station.
Oh well, it is a showcase for AMD and if it makes the AMD owners happy, that is cool with me
@ Humbug.
Give over and stop saying everything I type is baiting or trolling. Carry on and I will start reporting you.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,147
Oh well, it is a showcase for AMD and if it makes the AMD owners happy, that is cool with me
@ Humbug.
Give over and stop saying everything I type is baiting or trolling. Carry on and I will start reporting you.
There will hopefully become a time when a game is designed from the ground up with both mantle and dx used to there full potential so we can see what the real difference is. I bet the difference would be pretty big as atm all company's are doing atm is using it as an add in for extra performance.
I wanna see a game with added detail and the performance to back it up. This demo shows what added detail does to dx and thats slow it to a crawl.
Having motion blur on should not cripple a card and make any card run @ 40% usage.
As has been said, this is to showcase mantle benefits, as long as its adapted by all gfx chip makers then it will be a good thing.
If it never becomes open and is locked to amd cards only , then , well, "Revolutionary Mantle" will harm pc gaming.
asking a completely noob question as i often get 50% gpu and and 50%ish cpu utilization where is the dx/mantle bottlenecks?? its not bottle necked in the gpu as its still got plenty of gas in the tank same with my cpu,, so its the memory not transferring stuff fast enough or the electrons in the wires not going fast enough??
You dont know but you still post for what? You dont know as you said so stop posting about the same subject again and again.
They are already so many games that cant pass 30-50% gpu usage. Try sc2,total war,company of heroes,mmos.If the dx11 thread cant keep up with the draw calls no matter what you have you will never gonna have higher gpu usage.Dx11 use a thread for its own work so the bottleneck is there. Even if your engine is multithread you will gonna get a draw call bottleneck or dx11 thread bottleneck call it w/e.Dx11 creates a data structure thats why it needs an additional thread vs mantle.Too many draw calls and you know the rest
I have no idea what you have just said.
I cant explain it easier.sorry
Ok, I think I have got the gist of what you are saying.
How come when I change deferred contexts to enabled (switching the settings in profiles to 1), my GPU usage went to 80% and is hammering all the Mantle scores?
If they dont have DC in their graphic engine which i doubt then its gonna help.But its not optimal for a engine like frostbite. I dont know yet about this one
Well, they obviously do have DC in their engine, else it would make no difference at all.
Not that I care but why is there no Mantle scores as high as mine?
I have seen 80-90 fps mantle. Wait until crossfire/sli works.Also the mantle drivers atm are alpha quality.
We dont know if it is dx11 DC or its done on the graphic engine
Proof of those frames? Not that I doubt you but looking through here, I couldn't see any (maybe I missed them).
If the most I can get my GPU to is 40% (without having to mess about), I can't see how having SLI working will help?
Doesn't it seem strange that Oxide have this bench and state that they have spent all this time on DX and only a fraction on Mantle and yet, I can change a 0 to a 1 and get double the performance I previously had?
This is why I call BS to this bench and claim it is aimed at making Mantle look good (which isn't a bad thing) but at least do it honestly.
Soldato
- Joined
- 30 Nov 2011
- Posts
- 11,375
Right, the reason DX cant handle lots of drawcalls is because it is poor at multi threading them, you know this because its in the AMD presentation about mantle, so poor threading = maxing out at least 1 cpu core and then bottleneck
so why is it NONE of my cpu cores hits 100% with star swarm, but that is still a cpu bottleneck?
Soldato
Oh I'm so tempted to ramp this up 
Soldato
If this was something where PhysX could be run on AMD cards but it was severely gimped in the benchmark / demo due to reason whatever... can you imagine the uproar. Different nuances but same principle effectively.
Soldato
Oh wait. I can't. It crashes out instantly to desktop...Wrist slit worthy stuff this is.
Caporegime
There is also no DX deferred contexts score as high as yours despite the fact that one of them has a CPU faster than yours.
Also, have a look at Andy's 4930K @ 4.6 compared to my score in Mantle
Consistency in this is not 100%, its just been updated again to fix that.
You got extremely lucky. thats it.
Last edited:
Everything is blocking on a single thread. This is why nothing is being utilized 100%. If you profile the application, or even look at CPU monitoring programs, you'll see you have one CPU core being used and the others sitting mostly idle. What you are seeing is the game waiting on the driver thread.
Since the game has to wait on one core to process this one thread, it just hangs around doing nothing. This is why both the CPU and GPU are both underutilized.
Since the game has to wait on one core to process this one thread, it just hangs around doing nothing. This is why both the CPU and GPU are both underutilized.