Mantle and Frame Latency

flopper · 4 Oct 2014 at 10:59

Nexus18 said:
Yup 100% correct, once again mantle is utterly useless.....

Battlefield 4 with max settings except no AA at 2560x1080 on my rig as specced in sig, single player:

DX

Mantle

min 30fps more is godlike like amd.
30fps at such is a whole new generation die shrink performance increase.
no refreshed 780 aka 980 etc...

shankly1985 · 4 Oct 2014 at 11:45

Can we get this thread back on topic? And start posting results? Same old users

shankly1985 · 4 Oct 2014 at 11:49

Gregster said:
Lost me? I was comparing Shankly's results with Jwpalmers. Like I said, this isn't something I want to get into as for some the game is still broken and for others it runs great. I occasionally still get a BF4 has stopped working error but apart from Pearl Market, the game flows very smoothly.

I fully expect it to be smooth on AMD as well as nVidia and the time it has been out, there is little excuse for it not being. None of the graphs are in the 40+'s and this is where you start to feel the stutter.

Not being funny but if you don't want to add anything please do leave the thread.

Edit
Sorry Greg maybe I was abit harsh. I just dont want seeing this thread closed and I want to see a discussion and results shown.

Rusty0611 · 4 Oct 2014 at 11:55

Oh dear - that attitude to a member of the moderating team. Disagree with peoples POVs and thoughts all you like but telling a member of the mod team to leave the thread isn't the most intelligent thing you can do.

shankly1985 · 4 Oct 2014 at 12:20

jwrpalmer99 said:
I ran over same mission with SLI and GSync - it felt very smooth

Thanks for taking time.. It would seem AMD still win here on Smoothness. You link is all over the place, it should show a flat line as possible to show case how smooth each frame is getting rendered to the player.

Here my Crossfire result's had too spikes start of cut scene and end
Has you can see apart from the spike, look how flat the line is.

1080p - ultra - 80fov - mantle - no frame rate cap

shankly1985 · 4 Oct 2014 at 12:23

Here about with updated ms range 5ms added to graph

Wow
1080p - ultra - 80fov - mantle - no frame rate cap

shankly1985 · 4 Oct 2014 at 12:25

Rusty0611 said:
Oh dear - that attitude to a member of the moderating team. Disagree with peoples POVs and thoughts all you like but telling a member of the mod team to leave the thread isn't the most intelligent thing you can do.

I created this thread to see peoples results.. the first thing Greg said was he dont want to get involved thats fine I respect that. But why carry on posting after that?

andybird123 · 4 Oct 2014 at 12:36

shankly1985 said:
I created this thread to see peoples results.. the first thing Greg said was he dont want to get involved thats fine I respect that. But why carry on posting after that?

Thing is you are posting up a thread and subject that is subjective - mantle feels smoother to you - you are posting graphs to "prove" a feeling and then arguing with anyone that disagrees or has data that disagrees with your data

Obviously any dx vs mantle thread is going to draw comparisons of DX on Amd vs dx on nvidia in to the mix, some people even have both rigs to do comparisons, but people without both rigs are also sayingn that side by side testing is invalid

You must also be one of the last people to tell someone not to comment in one of "your" threads, as you take it upon yourself to go in to every gsync thread to tell people why their experience of that tech is invalid because of contrived reason x, just like you are trying to say that your opinion is the only valid one here

shankly1985 · 4 Oct 2014 at 12:39

Results dont lie, you can see clearly that the render time and frame latency is a different league on mantle vs dx

Edit
I edited my post, guess i was abit harsh

andybird123 · 4 Oct 2014 at 12:42

Thing is, most people posting results are running 60hz and vsync, so as long as the frane is rendered in under 16.6ms it makes absolutely no difference to their perception of smoothness

Some peoples results are actually showing a higher percentage of >16ms results on Mantle than on DX, which would be picked up as a skipped frame

shankly1985 · 4 Oct 2014 at 12:44

andybird123 said:
Thing is, most people posting results are running 60hz and vsync, so as long as the frane is rendered in under 16.6ms it makes absolutely no difference to their perception of smoothness

Some peoples results are actually showing a higher percentage of >16ms results on Mantle than on DX, which would be picked up as a skipped frame

You have link? so far I see only one result in here.. and that was Vsync but Gsync his avg Frame rate was 92fps not 60hz well not locked to 60hz

andybird123 · 4 Oct 2014 at 12:49

Even your own results show that when you do have a skipped frame, it ranges in to the 40+ms range, where as Nvidia dx results show a few just over 16, a 3 frame skip should be more noticeable than a single frame skip

shankly1985 · 4 Oct 2014 at 12:53

andybird123 said:
Even your own results show that when you do have a skipped frame, it ranges in to the 40+ms range, where as Nvidia dx results show a few just over 16, a 3 frame skip should be more noticeable than a single frame skip

No, No his spikes are much higher than 16ms the next part off graph on his is 16ms and then 20ms

Like I said my spike was start and end of the loading on cut scene.

andybird123 · 4 Oct 2014 at 12:59

so his graph goes 5, 10, 16?
Dont think so, it goes 5,10,15,20,25 clearly, so he has only a few result over 16 and even where they are he has none ranging to 40+ unlike you

If results dont lie then the nvidia graph very clearly shows less skipping and skipping of a shorter duration

shankly1985 · 4 Oct 2014 at 13:10

Ok his shows 5x over 15ms mine show 4x over 15ms

The only slow down I got was end of the cut scene.. This more a crossfire issue than Mantle as you can go look at my single result and see it didn't happen.

Gregster · 4 Oct 2014 at 13:13

shankly1985 said:
I created this thread to see peoples results.. the first thing Greg said was he dont want to get involved thats fine I respect that. But why carry on posting after that?

Maybe you misunderstood my comments and I wasn't interested in getting in a 'too and fro-ing' argument. I have already said that Mantle is smooth, so not sure why you felt the need to have a pop at me and not happy with it.

andybird123 · 4 Oct 2014 at 13:19

shankly1985 said:
Ok his shows 5x over 15ms mine show 4x over 15ms

The only slow down I got was end of the cut scene.. This more a crossfire issue than Mantle as you can go look at my single result and see it didn't happen.

Crucially, his "over 15's" are still under 16.6, which would be a critical cut off point for anyone running 60hz e.g. 90+% of gamers, where as yours are not

You are being very cute with your "interpretation" of the results

shankly1985 · 4 Oct 2014 at 13:27

andybird123 said:
Crucially, his "over 15's" are still under 16.6, which would be a critical cut off point for anyone running 60hz e.g. 90+% of gamers, where as yours are not

You are being very cute with your "interpretation" of the results

How is 20ms under 16.6 ???

see his image again, let alone how spikey all his results are. My line is very flat resulting in smoother, rendering times. his are all over the place

shankly1985 · 4 Oct 2014 at 13:34

Anyway can we have more results shown please? maybe Greg and Andy would like to share?

andybird123 · 4 Oct 2014 at 13:36

Oh...my...god...he has 2 at 20ms and not five as you previously alleged, you have 4 over 16.6 (the cut off for smoothness at 60hz) And another 2 at over 38ms which would be a full three frames skip

You have a total of 10 repeated frames vs 2

Objectively, the nvidia graph is 66% or 80% perceptually smoother