Caporegime
You are correct.
Soon as you replied with the PerfOverlay.FrameFileLogEnable 1 business, I started to yawn.
I will post some DX vs Mantle results in due course.
-
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.
***THE BF4 BENCHMARKS THREAD***
- Thread starter pgi947
- Start date
I will post some DX vs Mantle results in due course.
But you'll be comparing how AMD handles DX, as we know Mantle was created to fix AMD's DX flaws.
Last edited:
Soldato
Watching a YouTube video back will be hard, to pin point a time when a user feels slow down. Unless it's a massive slow down.
It's this reason why when I recorded all my mantle videos I left the performance graph on, that shows the engine in real time. While Greg might not feel the stutter that the graph shows, I can understand this I was the same before using mantle. Once you use Mantle dx shows its problems.
It's all about frame change, having the frame rate drop and back up again that moment and give a slow down effect.
It's this reason why when I recorded all my mantle videos I left the performance graph on, that shows the engine in real time. While Greg might not feel the stutter that the graph shows, I can understand this I was the same before using mantle. Once you use Mantle dx shows its problems.
It's all about frame change, having the frame rate drop and back up again that moment and give a slow down effect.
Caporegime
Soldato
I'm not trying to say you're having a bad experience or you're suffering from crippling stutter. I'm merely trying to explain why those graphs look like they do and from looking at your video rendering lines i can see why they show up as variance in frame time delivery.
The green lines tell you how long the gpu's are taking to render the frame. The gpus are responsible for generating the image and managing the delivery of frames to your display. Spikes in those lines indicate that frame took longer to render than the previous and that will show itself as spiking or hitching and shows the problems that those pcper graphs show.
Unless the spike in rendering time dropped your fps below 30, which it didn't then it will not show up in your video. I'm not trying to say your video looks stuttery. Even if it did it will not show up in a video recorded at 30 fps.
Why don't you play a round of Conquest 64 large and use the command perfoverlay.framefilelogenable 1 to start and 0 to stop. That will measure your smoothness etc far better asking me to find areas of stutter in your video.
Agreed. The point i was trying to make is you might not notice a difference when you've only tried one, it's only once you try something else you realise how much better it is. It is most apparent on multi gpu.
Nothing wrong with that at all and i used to think like that too. However as above, once you try something better and switch between the two it becomes quite noticeable. Of course some people just can't sense these things, but the majority usually can. It depends how sensitive you are to stutter i guess. I find this technical aspect of it interesting personally. I like to find why it's so much smoother, why is it much more consistent. So i go in search of answers.
I've got multi-GPUs, I've played BF4 on Nvidia DX, on AMD DX and on AMD Mantle and I can't say I've noticed a difference, unless I turn on the fps counter or something. Even then I don't notice the difference I'm just aware there is one because the number tell me.
I do seem to be more sensitive to the noise the cards make than small framerate inconsistencies though (obviously I can notice big drops or low framerates).
The main visual annoyance I notice is screen tearing and it really bugs me.
Caporegime
Guys, it doesn't matter. I have been playing BF4 loads since it has been working better and apart from the spike I get once in a blue moon (maybe once an hour), it runs sweet as. Smooth and no matter what people tell me, I know the difference between smooth and not smooth and my experience is smooth.
If Mantle users want to tell me it isn't smooth, that's also fine but I am quite content playing it how it is and hopefully G-Sync will show me what I am missing out on
If Mantle users want to tell me it isn't smooth, that's also fine but I am quite content playing it how it is and hopefully G-Sync will show me what I am missing out on
Soldato
So wish we had an in-game benchmark for this.
Soldato
- Joined
- 14 Feb 2011
- Posts
- 7,490
Agreed ^^^
Caporegime
Also agreed. It's hard to test multiplayer because network issues can and do interfere with the results.
Caporegime
I don't think anyone is going to argue over the frames, when clearly both AMD and nVidia can hold exceptional frames in BF4 but the point being made is smoothness and I am just not seeing where it isn't smooth for me. Mantle might well be better at keeping low latency and it should be however, I am failing to see where I am getting this apparent hitching and stutter I am supposed to be seeing with DX.
Caporegime
If you actually read whats been written no one has actually said it's not smooth for you.
Soldato
Thing is you prob wont, till you try out Mantle. I didn't
Caporegime
If you actually read whats been written no one has actually said it's not smooth for you.
Might make you feel sick but recorded this slow spinning camera to see if there is stuttering/tearing/hitching.... Am I still not seeing it?
Last edited:
Soldato
We not going to be able to tell by watching
We have no control or feel of the game. Its like watching a film at 24fps and then playing a game at 24fpsget me?
But if its fine to you then that's great.
Caporegime
Was the same for me. I was blissfully ignorant with DiretX 11 too prior to this.
Did a couple of tests and believe it or not on this occasion DirectX11 provided a slightly better experience than Mantle.
There were some big cpu spikes on the server i was on, particularly on my second Mantle run. But both results were definitely tainted in this case. Nonetheless, DX11 is working well for AMD too, despite what some will have you believe.
DirectX 11
Lockers 32 CQ Large
1440P Ultra
290P Crossfire Stock
14.7
Mantle
Lockers 32 CQ Large
1440P Ultra
290P Crossfire Stock
14.7
Last edited:
Soldato
Was the same for me. I was blissfully ignorant with DiretX 11 too prior to this.
Did a couple of tests and believe it or not on this occasion DirectX11 provided a slightly better experience than Mantle.
There were some big cpu spikes on the server i was on, particularly on my second Mantle run. But both results were definitely tainted in this case. Nonetheless, DX11 is working well for AMD too, despite what some will have you believe.
DirectX 11
Lockers 32 CQ Large
1440P Ultra
290P Crossfire Stock
14.7
Mantle
Lockers 32 CQ Large
1440P Ultra
290P Crossfire Stock
14.7
Is it normal for the Mantle results to say DirectX on them?
In fact those results are very close.
Caporegime
I'll try explain what these frame time readings actually mean later, some of you are either getting confused or making it up as you go along as to what they actually mean and how to distinguish good readings from bad, it gos far far deeper then lower = better, as its not always the case.
Caporegime
We not going to be able to tell by watching
get me?
But if its fine to you then that's great.
You can't tell if that is smooth by watching? I am not sure what else I can say really and will leave it at that.
Caporegime
Consistency is the most important thing!