ASM-coder
Member
- Jan 12, 2014
- 193
- 0
- 0
Oops you are right, that was fixed in a BF4 patch. But it's still old.It's not 14.1 that was the problem it was BF4, but it has been patched a bit ago which fixed that problem.
Oops you are right, that was fixed in a BF4 patch. But it's still old.It's not 14.1 that was the problem it was BF4, but it has been patched a bit ago which fixed that problem.
Is the bug with terrain decoration still present? It changed the quality every time you started the game despite the setting remaining the same...
I'm saying Mantle is not identical to DX as far as IQ. It is up to the end user as far as the differences.
I'm saying Mantle is not identical to DX as far as IQ. It is up to the end user as far as the differences.
http://forum.hardwarepal.com/battlefield-4-mantle-vs-directx-image-quality-comparison/
Earlier you said Ultra? I have a feeling there are some in-depth reviews coming on this from more than one place. We will also get to see Thief tomorrow as well.
Gloomy, Mantle has an APE on that billboard while DX has a WOLF!
Clearly the WOLF is a more dangerous animal, so Mantle is weak!..
/sarc
Thanks for the screens, seems to be the usual case of looking too hard for a fault when Mantle gives double the minimum FPS during intense scenes. Seems very strange some users aren't happy BF4 is now devoid of those terrible fps chokes.
3DVagabond
less viewing distance = far out mountain range are missing
less texture filtering = the leaves on the ground is a blended mess
less lighting = missing tree shadows
less terrain decoration = missing a whole lot of grass
less ambient occlusion = missing that glow where light bounces off object
did 14.2B fix all this? if so - good for amd.
Glommy
both are those comparison are near identical. too bad you did not engage (1) perfoverlay.drawfps and (2) perfoverlay.drawgraph and (3) render.drawscreeninfo.
OC Guy
can you also do a comparison on caspain border 2014 (that is the harshest map) with ultra setting.
I know someone is going to do something offensive next like suggest I just took two screenshots without changing the API. Going to curtail that right now: I logged in and out to change the API for both screenshots. There is no visual difference in-game. In the screenshots, it seems like the DX shot has some details missing esp. in the awning that is in front of the skyscraper. That's just dust and smoke getting in the way.
@UaVaj
I'll redo it later. Don't feel like it right now. I understand some people may want proof. I think it's offensive that you think I'm lying when there are a number of people here who can test it independently, but I understand being unable to test it yourself.
At low settings for high fps/120hz gaming my 7970 at 1050mhz had over 50fps higher minimums than my gtx780 at 1176mhz. I had a frame cap at 120 so I don't know exactly how much faster. Mantle never dropped below 120fps.
There are lots of Mantle Reviews on the net but none of them quite seem to reflect the point of Mantle and what it does.
My experience of those reviews seems typical in that they have a tendancy to use the best 12 thread Intel CPU's overclocked and then set very high in game settings, often also only testing in Campaign mode which has far less CPU implications.
What this does is make the Game completely GPU bound.
Mantle removes or betters CPU bottlenecking, so.
Anyone worth their knowledge knows the way you test for CPU bettlenecks is to run very low GFX setting, this moves the workload away from the GPU onto the CPU, this is what i have seen Anand do many times, and its right.
So with that principle in mind i have done some testing of my own, little of it, but didn't need to to show its effects right away.
So, here we go.
System used:
Motherboard: Asus Sabertooth 990FX, Rev-1.01, BIOS 1503.
CPU: AMD Phenom II x6 1090T @ 4.2Ghz
GPU: Radeon 7870XT (Tahiti LE) @ 1150 / 1575
RAM: Corsair XMS3 4x2GB @ 1600Mhz
SSD Samsung 830
PSU: XFX 750 Watt Pro Edition
Case: CM HAF 912+
Op: Shanghai:
I chose this map immediately as it has a well known CPU bottleneck; all the way down the main dual bridge road. Standing at the very top of the road my results are as follows.
48 Player server, with 40+ players active. (GFX setting = Medium preset @ 1080P)
In Direct X: 57 FPS.
In Mantle: 123 FPS
The performance difference in Mantle is +110% in that scenario.
This does not mean there is a FPS boost everywhere all the time, thats not the case at all.
I play BF4 all the time, since Mantle became available in BF4 i have switched to it and never looked back.
In Direct X my FPS get dragged down by a half regularly, often less than what you see in that screenshot.
With Mantle i'm constantly running roughly around 100 FPS Minimum on all High Settings 1080P. its like night and day.
I have made a couple of Youtube vids to show it in action, keep in mind that what i used to record it, Mirillis Action, reduces the performance by a good 20%.
Direct X
https://www.youtube.com/watch?v=3WFqX5k_GgA
Mantle
https://www.youtube.com/watch?v=pcEEuIvN8FI
4.8Ghz Sandy Bridge i7 and R9 290 xFire x2 test in this post. http://forums.anandtech.com/showpost.php?p=36169582&postcount=40
Some large performance increases being reported in this thread, much higher than what the reviews stated..
This one in particular made me
I would definitely need evidence to believe that. It's not like he was using a slow CPU either..
Mantle would impress me if if didn't require specific game support (or if support was widespread).
With the early problems (Dice's fault?) and extremely limited GPU support + extensive effort to make it work (see Thief's delayed Mantle implementation), I can't see this being relevant outside of rare instances over the next 12-18 months.
This kind of stuff is always really irritating in terms of fragmentation. Glide, Physx, S3 Metal, Redline, TressFX, it's all a bunch of bullcrap if you ask me. I wish the world revolved around OpenGL, or if need be, OpenGL and D3D. God help us if nVidia does something similar (and I wouldn't put it past them honestly). Then to make a PC game it would be just like the bad old days when you had to get the right version, right patch for this and that, right driver, oh your card is a different revision and isn't supported, yadda yadda.
All relevant : http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support
Somebody please settle this crap. I put a lot of this at the feet of Microsoft for letting DirectX sit around for freaking ages, and I blame Xbox for that. They screwed PC gamers, tried to force them to go with their cash-cow closed-market console by buying up exclusives, limiting PC versions, delaying PC versions, and just generally being bad for gaming overall. In that way, maybe Mantle will at least light a fire under their wannabe monopolistic asses. But give us a Mantle that will work with 95% of PCs and games, and not something only relevant for a few.