NVIDIA’s VXAO (voxel based ambient occlusion) To Be Featured In Rise of the Tomb Raider

[patrickjp93]

nnnnnnnnooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo

 

here we go

 

Let the flame war begin. Here, I'll kick us off with a reminder of the one that was ended before it even began:

 

[Sauron]

 

I wish more people could see this

[Guest Kloaked]

Let the flame war begin. Here, I'll kick us off with a reminder of the one that was ended before it even began:

 

Screenshot 2016-01-03 21.42.00.png

 

fuckin lol

[Prysin]

The other guy is right, you sound ridiculous, and there isn't freedom of speech here, there is a TOS 

 

 

Besides, what good does personally attacking an inanimate object do you? a peice of plastic and silicon can't really be a "fag" can it?

[Prysin]

Funny because the AOTS benchmarks say Nvidia still does it better anyways.

after they heavily optimized their drivers. also AOTS is still alpha/beta.

 

if you own a Nvidia GPU, bench it. ill pull out my R9 295x2.

 

AOTS can be bought on steam quite cheaply

[patrickjp93]

after they heavily optimized their drivers. also AOTS is still alpha/beta.

 

if you own a Nvidia GPU, bench it. ill pull out my R9 295x2.

 

AOTS can be bought on steam quite cheaply

 

Currently on a 750m, so kinda pointless.

 

And what's your point? AMD can optimize drivers too, but to this day, there's still no single GPU solution better than the water force 980TI. Some people breaking into the 1500MHz area even beat the 295X2 overclocked.

[Prysin]

Currently on a 750m, so kinda pointless.

 

And what's your point? AMD can optimize drivers too, but to this day, there's still no single GPU solution better than the water force 980TI. Some people breaking into the 1500MHz area even beat the 295X2 overclocked.

a 295x2 is a bitch to OC. not just because you need a hefty PSU, but because you need tougher grade wires. cooling for the power connectors themselves (they hit 60c under normal stock load with no airflow over them).

 

then you need to make sure you can get enough cold air past the rad.

 

and then you need to bios mod it to get rid of the 75c temp limit.

 

THEN you can start overclocking it.

 

so all in all, i doubt that the 295x2 was any more then 8-10% overclocked at best

[Guest Kloaked]

a 295x2 is a bitch to OC. not just because you need a hefty PSU, but because you need tougher grade wires. cooling for the power connectors themselves (they hit 60c under normal stock load with no airflow over them).

 

then you need to make sure you can get enough cold air past the rad.

 

and then you need to bios mod it to get rid of the 75c temp limit.

 

THEN you can start overclocking it.

 

so all in all, i doubt that the 295x2 was any more then 8-10% overclocked at best

 

....and why were people recommending this abomination? Oh because the price came down, so it must be a good buy.

[patrickjp93]

a 295x2 is a bitch to OC. not just because you need a hefty PSU, but because you need tougher grade wires. cooling for the power connectors themselves (they hit 60c under normal stock load with no airflow over them).

 

then you need to make sure you can get enough cold air past the rad.

 

and then you need to bios mod it to get rid of the 75c temp limit.

 

THEN you can start overclocking it.

 

so all in all, i doubt that the 295x2 was any more then 8-10% overclocked at best

 

1300MHz. Wccftech had a benchathon. The winner had a 1600MHz 980TI Waterforce.

[Prysin]

1300MHz. Wccftech had a benchathon. The winner had a 1600MHz 980TI Waterforce.

1300Mhz...

 

ive hit 1170 on mine just briefly and 1102 stable (during winter) and given my ASIC values, mine is probably just OK. not good or bad...

 

that is truly a golden pair.

[Sharkyx1]

Snip

It's not like saying how little fucks you give gives you any upper hand at all.

 

"act like a dumbass"

 

get called out

 

"Oh look i don't care :^D"

 

I just think it's dumb to throw around insults like you can hurt the feelings of a games console.

[SurvivorNVL]

Let the flame war begin. Here, I'll kick us off with a reminder of the one that was ended before it even began:

 

Screenshot 2016-01-03 21.42.00.png

That right there is both gorgeous and hilarious.  I am so sad that I wasn't there that night to watch that post go up, and see you freak out.  That would've just been, lol.. I can't imagine how it must've been in that thread.

[Notional]

If it is based on Asych Compute then it should run better on AMD, because AMD has hardware support for it and doesn't have to run an emulation layer to use it.

 

Yeah, but as long as it's not implemented as concurrent async compute, the AMD cards won't leave NVidia in the dust (just be ahead of them a bit).

 

Funny because the AOTS benchmarks say Nvidia still does it better anyways.

 

And we know why: The devs implemented an entire other NVidia programming line, that did not use concurrent async compute, but rather proprietary NVidia shaders, etc. AOTS still proves that NVidia cannot do concurrent async compute worth a damn (or async compute good at all).

[patrickjp93]

Yeah, but as long as it's not implemented as concurrent async compute, the AMD cards won't leave NVidia in the dust (just be ahead of them a bit).

 

 

And we know why: The devs implemented an entire other NVidia programming line, that did not use concurrent async compute, but rather proprietary NVidia shaders, etc. AOTS still proves that NVidia cannot do concurrent async compute worth a damn (or async compute good at all).

 

If you can achieve the same result using less computational power, then that's the better solution. I don't give a rat's ass if it's standard or not. AMD is competing against Nvidia's software engineers as well as its hardware engineers. Nvidia wins so far.

[Prysin]

If you can achieve the same result using less computational power, then that's the better solution. I don't give a rat's ass if it's standard or not. AMD is competing against Nvidia's software engineers as well as its hardware engineers. Nvidia wins so far.

Nvidia wins at software.

they have no chance in raw hardware.

 

if assuming equal driver overhead and application specific optimization, AMDs hardware is simply stronger.

[Fetzie]

If you can achieve the same result using less computational power, then that's the better solution. I don't give a rat's ass if it's standard or not. AMD is competing against Nvidia's software engineers as well as its hardware engineers. Nvidia wins so far.

Yep, if nVidia is worse at something than AMD it doesn't matter, and if nVidia is better at something than AMD then that is a deal-breaker issue.

 

i wish this was sarcasm

[Marinatall_Ironside]

Can't you guys just shut the fuck up on this AMD/Nvidia bullshit already. It's counterproductive and disheartening for someone like me to see all of you act like a bunch of 12 year olds over this stupid shit in this thread.

Now I will say that VXAO is very intriguing since we have seen SSAO and HBAO for a while. If game developers and software developers are pushing graphics technology into a new realm of reality, it's good because then games don't look dated or outdated, causing a stagnation in GPU development in raw power. And that goes to BOTH AMD and Nvidia.

With an increase in available computer resources such as Asynchronous Compute, Direct3D 12, and Vulkan, whereby it reduces overhead and improves utilization, that means you can do more now since there is that reduction in overhead on resources. Yes I also intend to push the limits of hardware specs to achieve advanced graphics and gameplay now that these are available. However optimization is still crucial so that games still know how to use resources properly and doesn't run like shit (Lookin' at you Arkham Knight) because of bad code.

And to clear things up, Ambient Occlusion is subtle and not pronounced since it only affects areas like crevices and other small gaps in 3D models by approximating the effects of lighting and applying soft shadows. This concept applies to all forms of Ambient Occlusion.

[Notional]

If you can achieve the same result using less computational power, then that's the better solution. I don't give a rat's ass if it's standard or not. AMD is competing against Nvidia's software engineers as well as its hardware engineers. Nvidia wins so far.

 

NVidia's hardware is vastly inferior. We can see that in the Maxwell cards, being utterly useless at anything new. No NVidia's force is and has been for a very long time, their software department. They have more means and use it to optimize each game for the cards. That might sound good and yield good results here and now, but as we can see with planned obsolescence and the 700 series, the performance craps out completely the second focus is removed from the cards.

 

We don't see the same when the focus is on hardware. That is why Hawaii is making a mockery of the 780 series now.

 

Async compute is the future, as you can have more hardware ressources for compute, if more effects and graphics are compute based. It simply gives a lot more flexibility to the devs.

 

I'm sure NVidia will put in the extra money to replace concurrent async compute with their own specific solutions for each card, but the second they remove focus on a card, it will tank in performance. This will likely happen to the 970 the most, as games will treat it like a 4GB card and use it as such. So if the drivers aren't constantly patched to tell new games how to handle it, we might see that bottleneck get worse and worse. Oh and of course the horrible async compute performance and lack of proper concurrent async compute.

[BuckGup]

Yeah! More optimization!

[Sauron]

Nvidia wins at software.

they have no chance in raw hardware.

 

if assuming equal driver overhead and application specific optimization, AMDs hardware is simply stronger.

 

actually their drivers have been a clusterfuck lately - definitely not better than amd's ones, except for the crimson driver burning a couple of cards

[Snadzies]

I have a GTX 980 and I end up turning the various Nvidia graphical settings off as the visual improvements are so minor that I generally have to do a side by side comparison to notice the difference.

I'm sure this will be no different.

[patrickjp93]

Nvidia wins at software.

they have no chance in raw hardware.

 

if assuming equal driver overhead and application specific optimization, AMDs hardware is simply stronger.

 

Nvidia will win at hardware when it has to, not before. AMD needs to learn to be competitive at software.

[patrickjp93]

Yep, if nVidia is worse at something than AMD it doesn't matter, and if nVidia is better at something than AMD then that is a deal-breaker issue.

 

i wish this was sarcasm

 

1)BS

2)It should be.

 

Just because Nvidia lacks a hardware feature that isn't even in the DX 12 standard (it's not required for Tier 3 support) doesn't mean they can't still have the superior cards and drivers in combination. AMD can have the best hardware ever and I won't give a damn if their drivers are so bad Nvidia cleans their clocks!

[WolfDeville]

Kobathor, on 07 Jan 2016 - 06:36 AM, said:

Everybody is going to turn it off anyways, why bother?

I won't. And I'm sure other people with expensive rigs with powerful GPU's won't either...

[patrickjp93]

NVidia's hardware is vastly inferior. We can see that in the Maxwell cards, being utterly useless at anything new. No NVidia's force is and has been for a very long time, their software department. They have more means and use it to optimize each game for the cards. That might sound good and yield good results here and now, but as we can see with planned obsolescence and the 700 series, the performance craps out completely the second focus is removed from the cards.

 

We don't see the same when the focus is on hardware. That is why Hawaii is making a mockery of the 780 series now.

 

Async compute is the future, as you can have more hardware ressources for compute, if more effects and graphics are compute based. It simply gives a lot more flexibility to the devs.

 

I'm sure NVidia will put in the extra money to replace concurrent async compute with their own specific solutions for each card, but the second they remove focus on a card, it will tank in performance. This will likely happen to the 970 the most, as games will treat it like a 4GB card and use it as such. So if the drivers aren't constantly patched to tell new games how to handle it, we might see that bottleneck get worse and worse. Oh and of course the horrible async compute performance and lack of proper concurrent async compute.

Even though in DX 12 benchmarks Nvidia seems to win more than it loses with Maxwell. Such logic...

 

The 780/TI still trade blows with the 290/X and 390/X, even if a little behind. 

 

I don't believe it until AMD proves it because on paper to me it doesn't seem that way at all. It's extremely rare that you can come up with a workload that a GPU is good for where synchronicity is a requirement. That's exactly why most good AIs still run on CPUs. Sure you can speed up learning with a GPU, but the decisions are all at the CPU.

 

Nvidia builds the hardware for its time and builds the software to properly support it for multiple years. No one reasonable can complain about that. If AMD doesn't get its software act together, it will go down in flames eventually. Nvidia's engineers are just as good at what they do. They wouldn't have lasted this long if such wasn't the case. AMD cannot continue to let software be on the back burner. And I'd argue the Fury X was a garbage design. Slice the SPs down by 15% and double the ROPs. I don't care if it would have required more time. It needed it.