***THE BF4 BENCHMARKS THREAD***

[humbug]

1080 200% is 4k

How's the vram situation with mantle at ultra (no aa) at that res?

1920 x 1080P @ 200% = 3840 x 2160 (4K) thats right, its not all 'that' high really is it?

i might try it, see how i get on

 

[pgi947]

Nope its just the fear factor all those numbers bring

I can go as high as 200% 2xaa then I'm fresh out of vram, sure looks pretty too, a shame more games don't have internal resolution scaling.

 

[shankly1985]

1080 200% is 4k

How's the vram situation with mantle at ultra (no aa) at that res?

Will let you know tomorrow that was just quick run around that map. Must say though very playable. At that res you don't need any AA I would say.

 

[humbug]

Nope its just the fear factor all those numbers bring

I can go as high as 200% 2xaa then I'm fresh out of vram, sure looks pretty too, a shame more games don't have internal resolution scaling.

I think i would be fresh out of GPU muscle before i hit the Memory buffers. i will try it though. later on.

 

[pgi947]

A pair of 780's (and by default pair of 290's) handle it surprisingly well, full ultra at 4k is unplayable but only due to hitting the vram limit, frames are still in the 35-50 range. No aa is an easy 100+, 2x is around 70-75+ iirc.

How's the 780 hunt coming along anyway humbug? You missed out on 8packs stash!

Vram leaks still an issue with mantle (hence asking about vram usage at 4k no aa)?

 

[shankly1985]

A pair of 780's (and by default pair of 290's) handle it surprisingly well, full ultra at 4k is unplayable but only due to hitting the vram limit, frames are still in the 35-50 range. No aa is an easy 100+, 2x is around 70-75+ iirc.

How's the 780 hunt coming along anyway humbug? You missed out on 8packs stash!

Vram leaks still an issue with mantle (hence asking about vram usage at 4k no aa)?

I believe the Vram leak is just 7000 series cards. I get no issue with Vram on my cards. Not 100% sure on 4k though but I did spend about ten mins running about and it was fine.

 

[LtMatt]

Yeah theres no leak as such just at the moment Mantle uses more vram than DX. As humbug noted it appears to be because it does not get cleared when needed.

 

[shankly1985]

Yeah theres no leak as such just at the moment Mantle uses more vram than DX. As humbug noted it appears to be because it does not get cleared when needed.

yeah
I have also am sure seen 6GB+ on my cards full ultra @1080p dont quote me on that. Will let you know maybe Hwinfo64 no reporting correct results?

Also I wounder if mantle does add up the two cards Vram? 4+4 = 8GB?

 

[shankly1985]

No spiking for me in DX @ 4K x

Who getting spiking? becasue I sure am not.

No spiking for me in Mantle @ 4K

 

[LtMatt]

yeah
I have also am sure seen 6GB+ on my cards full ultra @1080p dont quote me on that. Will let you know maybe Hwinfo64 no reporting correct results?

Also I wounder if mantle does add up the two cards Vram? 4+4 = 8GB?

Render.DrawScreenInfo 1 into the console tells you accurately how much vram Mantle uses.

 

[shankly1985]

Render.DrawScreenInfo 1 into the console tells you accurately how much vram Mantle uses.

Ok seen a post by Rusty I should half what Hwinfo says so if it was 6GB then I was using around 3GB Vram.

 

[shankly1985]

More Mantle VS DirectX

Check the Gain here I would like to see some you Guys with Nvidia post some screens from this location.

Also just look at that Graph on DirectX wow!

Settings

Spoiler

DirectX - jumping between 105/109fps

Spoiler

Mantle - jumping between 155/158

Spoiler

 

[pgi947]

780 SLI @ 1150/6800

Wiggly line things aren't the bottom aren't as smooth as the mantle screen but my frame rates are matching your mantle scores.

How come one of your cards in the DX shot is only registering 5% usage?

Spoiler

 

[shankly1985]

780 SLI @ 1150/6800

Wiggly line things aren't the bottom aren't as smooth as the mantle screen but my frame rates are matching your mantle scores.

How come one of your cards in the DX shot is only registering 5% usage?

Spoiler

Not quite the same location, can you head to where I was.

Its how RadeonPro reads with the R9 cards. Once sec it will be 100% next %5 just jumps all over the place. Msi AB giving you 145fps and BF4 is giving you 152fps lol

 

[LtMatt]

I would love to get a comparison up but can't atm. Hopefully on tuesday or wednesday ill stick mine in.

 

[shankly1985]

I would love to get a comparison up but can't atm. Hopefully on tuesday or wednesday ill stick mine in.

Matty does AB also give your GPU readings like mine it will jump all over place?

 

[pgi947]

Not quite the same location, can you head to where I was.

Its how RadeonPro reads with the R9 cards. Once sec it will be 100% next %5 just jumps all over the place. Msi AB giving you 145fps and BF4 is giving you 152fps lol

Will give it a blast now, I think the mismatched FPS readings is due to either AB or the engine reading frame rates quicker/slower (delete as applicable). iirc FRAPS reads around the same as the game based read out. Think its that 'hardware polling period' thing in AB's settings.

 

[LtMatt]

Matty does AB also give your GPU readings like mine it will jump all over place?

Yes its a bug with the sensor. If you're using ab beta 18 its not too bad, but still not perfect.

 

[shankly1985]

Will give it a blast now, I think the mismatched FPS readings is due to either AB or the engine reading frame rates quicker/slower (delete as applicable). iirc FRAPS reads around the same as the game based read out. Think its that 'hardware polling period' thing in AB's settings.

Yeah very well could be polling rate.

Yes its a bug with the sensor. If you're using ab beta 18 its not too bad, but still not perfect.

Good to know I not alone.

 

[LtMatt]

Yeah think pg has it spot on. Also think Unwinder slowed down the rate at which its read on beta 18 so it should be more consistent.