[Various] Ashes of the Singularity DX12 Benchmarks - Page 141 - Overclock.net - An Overclocking Community

Forum Jump: 

[Various] Ashes of the Singularity DX12 Benchmarks

 
Thread Tools
post #1401 of 2682 (permalink) Old 08-30-2015, 10:33 PM
New to Overclock.net
PirateZ Level 1
 
PirateZ's Avatar
 
Join Date: Mar 2014
Posts: 356
PirateZ is offline  
Sponsored Links
Advertisement
 
post #1402 of 2682 (permalink) Old 08-30-2015, 10:44 PM
New to Overclock.net
Dudewitbow Level 1
 
Dudewitbow's Avatar
 
Join Date: Oct 2012
Location: Pittsburg California
Posts: 1,736
Quote:
Originally Posted by PirateZ View Post

So who's winning? biggrin.gif

It's not about whose winning, its more about the true potential of current gen hardware, and its potential future within the next few years for those who bought into it and not looking to upgrade.

Dudewitbow is offline  
post #1403 of 2682 (permalink) Old 08-30-2015, 10:50 PM
New to Overclock.net
Forceman has a spectacular aura about Forceman has a spectacular aura about Forceman has a spectacular aura about Forceman has a spectacular aura about Forceman has a spectacular aura about Forceman has a spectacular aura about Forceman has a spectacular aura about
 
Forceman's Avatar
 
Join Date: Sep 2012
Posts: 10,869
Quote:
Originally Posted by PirateZ View Post

So who's winning? biggrin.gif

No one knows, because we only have a single data point and it's hard to get a trend line from a single point.

Forceman's Law: Any AMD/Nvidia GPU thread, no matter what the topic, will eventually include a post referencing the GTX 970.
Forceman is offline  
Sponsored Links
Advertisement
 
post #1404 of 2682 (permalink) Old 08-30-2015, 11:03 PM
New to Overclock.net
Kollock Level 1
 
Kollock's Avatar
 
Join Date: Aug 2015
Posts: 39
Quote:
Originally Posted by Mahigan View Post

@Kpjoslee the oxide dev told us the reason behind the poor NVIDIA performance.

They turned off async compute because it performed even worse than the result we see now.

The results we see now are from a Vendor ID specific path. This means that once the Ashes of the Singularity detects NVIDIA hardware, it runs a non-async path.

Since Ashes of the Singularity makes relatively little use of Post Processing effects the boost we see for AMD hardware, due to Async, is minor. That being said it is enough to push a 290x into a parity position with a GTX 980 Ti. The GTX 980 Ti is just running a traditional path. That's why the performance is similar to DX11.

DX11 is, at times, a bit faster because of driver interventions.

People find this weird, as did I. But this pretty much is what we're going to expect from DX12 gaming if Async is used.

Pretty much... There won't be asynchronous compute tier 2 on Maxwell 2. The hardware has a defect. A defect which drives performance down when Async is turned on.

I've looked into it. As you've all read. And I think it is either a defect with the second DMA engine or with the L2 Cache.

For all those who bought Maxwell 2 cards... It seems you won't benefit from Async Compute.

I think you deserve a response from NVIDIA. Maybe you disagree. To me it seems that with the level of Async compute titles on the horizon, you deserve an explanation. Especially when you consider that review sites, quoting NVIDIA, now have their credibility in question.

That's not honest imo.

I think you are confusing a few issues. Tier 2 vs Tier 3 binding is a completely separate issue from Async Compute. It's has to do with the number of root level descriptors we can pass. In tier 3, it turns out we can basically never update a descriptor during a frame, but in tier 2 we sometimes have to build a few . I don't think it's a significant performance issue though, just a techinical detail.

In regards to the purpose of Async compute, there are really 2 main reasons for it:

1) It allows jobs to be cycled into the GPU during dormant phases. In can vaguely be thought of as the GPU equivalent of hyper threading. Like hyper threading, it really depends on the workload and GPU architecture for as to how important this is. In this case, it is used for performance. I can't divulge too many details, but GCN can cycle in work from an ACE incredibly efficiently. Maxwell's schedular has no analog just as a non hyper-threaded CPU has no analog feature to a hyper threaded one.

2) It allows jobs to be cycled in completely out of band with the rendering loop. This is potentially the more interesting case since it can allow gameplay to offload work onto the GPU as the latency of work is greatly reduced. I'm not sure of the background of Async Compute, but it's quite possible that it is intended for use on a console as sort of a replacement for the Cell Processors on a ps3. On a console environment, you really can use them in a very similar way. This could mean that jobs could even span frames, which is useful for longer, optional computational tasks.

It didn't look like there was a hardware defect to me on Maxwell just some unfortunate complex interaction between software scheduling trying to emmulate it which appeared to incure some heavy CPU costs. Since we were tying to use it for #1, not #2, it made little sense to bother. I don't believe there is any specific requirement that Async Compute be required for D3D12, but perhaps I misread the spec.

Regarding trying to figure out bottlenecks on GPUS, it's important to note that GPUs do not scale simply by adding more cores to it, especially graphics tasks which have alot of serial points. My $.02 is that GCN is a bit triangle limited, which is why you see greater performance on 4k, where the average triangle size is 4x the triangle size of 1080p.

I think you're also being a bit short-sighted on the possible use of compute for general graphics. It is not limited to post process. Right now, I estimate about 20% of our graphics pipeline occurs in compute shaders, and we are projecting this to be more then 50% on the next iteration of our engine. In fact, it is even conceivable to build a rendering pipeline entirely in compute shaders. For example, there are alternative rendering primitives to triangles which are actually quite feasible in compute. There was a great talk at SIGGRAPH this year on this subject. If someone gave us a card with only compute pipeline, I'd bet we could build an engine around it which would be plenty fast. In fact, this was the main motivating factors behind the Larabee project. The main problem with Larabee wasn't that it wasn't fast, it was that they failed to be able to map dx9 games to it well enough to be a viable product. I'm not saying that the graphics pipeline will disappear anytime soon (or ever), but it's by no means certain that it's necessary. It's quite possible that in 5 years time Nitrous's rendering pipeline is 100% implemented via compute shaders.
Kollock is offline  
post #1405 of 2682 (permalink) Old 08-30-2015, 11:13 PM
New to Overclock.net
Noufel Level 1
 
Noufel's Avatar
 
Join Date: Apr 2012
Location: Constantine, Algeria
Posts: 1,548
Quote:
Originally Posted by Kollock View Post

Quote:
Originally Posted by Mahigan View Post

@Kpjoslee the oxide dev told us the reason behind the poor NVIDIA performance.

They turned off async compute because it performed even worse than the result we see now.

The results we see now are from a Vendor ID specific path. This means that once the Ashes of the Singularity detects NVIDIA hardware, it runs a non-async path.

Since Ashes of the Singularity makes relatively little use of Post Processing effects the boost we see for AMD hardware, due to Async, is minor. That being said it is enough to push a 290x into a parity position with a GTX 980 Ti. The GTX 980 Ti is just running a traditional path. That's why the performance is similar to DX11.

DX11 is, at times, a bit faster because of driver interventions.

People find this weird, as did I. But this pretty much is what we're going to expect from DX12 gaming if Async is used.

Pretty much... There won't be asynchronous compute tier 2 on Maxwell 2. The hardware has a defect. A defect which drives performance down when Async is turned on.

I've looked into it. As you've all read. And I think it is either a defect with the second DMA engine or with the L2 Cache.

For all those who bought Maxwell 2 cards... It seems you won't benefit from Async Compute.

I think you deserve a response from NVIDIA. Maybe you disagree. To me it seems that with the level of Async compute titles on the horizon, you deserve an explanation. Especially when you consider that review sites, quoting NVIDIA, now have their credibility in question.

That's not honest imo.

I think you are confusing a few issues. Tier 2 vs Tier 3 binding is a completely separate issue from Async Compute. It's has to do with the number of root level descriptors we can pass. In tier 3, it turns out we can basically never update a descriptor during a frame, but in tier 2 we sometimes have to build a few . I don't think it's a significant performance issue though, just a techinical detail.

In regards to the purpose of Async compute, there are really 2 main reasons for it:

1) It allows jobs to be cycled into the GPU during dormant phases. In can vaguely be thought of as the GPU equivalent of hyper threading. Like hyper threading, it really depends on the workload and GPU architecture for as to how important this is. In this case, it is used for performance. I can't divulge too many details, but GCN can cycle in work from an ACE incredibly efficiently. Maxwell's schedular has no analog just as a non hyper-threaded CPU has no analog feature to a hyper threaded one.

2) It allows jobs to be cycled in completely out of band with the rendering loop. This is potentially the more interesting case since it can allow gameplay to offload work onto the GPU as the latency of work is greatly reduced. I'm not sure of the background of Async Compute, but it's quite possible that it is intended for use on a console as sort of a replacement for the Cell Processors on a ps3. On a console environment, you really can use them in a very similar way. This could mean that jobs could even span frames, which is useful for longer, optional computational tasks.

It didn't look like there was a hardware defect to me on Maxwell just some unfortunate complex interaction between software scheduling trying to emmulate it which appeared to incure some heavy CPU costs. Since we were tying to use it for #1, not #2, it made little sense to bother. I don't believe there is any specific requirement that Async Compute be required for D3D12, but perhaps I misread the spec.

Regarding trying to figure out bottlenecks on GPUS, it's important to note that GPUs do not scale simply by adding more cores to it, especially graphics tasks which have alot of serial points. My $.02 is that GCN is a bit triangle limited, which is why you see greater performance on 4k, where the average triangle size is 4x the triangle size of 1080p.

I think you're also being a bit short-sighted on the possible use of compute for general graphics. It is not limited to post process. Right now, I estimate about 20% of our graphics pipeline occurs in compute shaders, and we are projecting this to be more then 50% on the next iteration of our engine. In fact, it is even conceivable to build a rendering pipeline entirely in compute shaders. For example, there are alternative rendering primitives to triangles which are actually quite feasible in compute. There was a great talk at SIGGRAPH this year on this subject. If someone gave us a card with only compute pipeline, I'd bet we could build an engine around it which would be plenty fast. In fact, this was the main motivating factors behind the Larabee project. The main problem with Larabee wasn't that it wasn't fast, it was that they failed to be able to map dx9 games to it well enough to be a viable product. I'm not saying that the graphics pipeline will disappear anytime soon (or ever), but it's by no means certain that it's necessary. It's quite possible that in 5 years time Nitrous's rendering pipeline is 100% implemented via compute shaders.
I hope all game devs can behave like you and come to tech forums to share their information and respond to people questions smile.gif
Bravo thumb.gif
Noufel is offline  
post #1406 of 2682 (permalink) Old 08-30-2015, 11:20 PM
Linux Lobbyist
Xuper Level 1
 
Xuper's Avatar
 
Join Date: Jan 2014
Location: None
Posts: 787

From Anandtech Silverforce11!

Wow I can't believe!

 

https://www.khronos.org/assets/uploads/developers/library/2015-gdc/Valve-Vulkan-Session-GDC_Mar15.pdf

 


CPU : AMD Ryzen 1600X | Memory : [Ripjaws V] F4-3200C16D-16GVKB  | Motherboard : Asus Prime X370 Pro | Graphic : XFX AMD Radeon R9 290 Double Dissipation | Monitor : AOC 931Sw | HDD : 1x120 GB SSD Samsung Evo , 1x2TB Seagate

 

 

Xuper is offline  
post #1407 of 2682 (permalink) Old 08-30-2015, 11:26 PM
New to Overclock.net
spacin9 Level 1
 
spacin9's Avatar
 
Join Date: Apr 2013
Posts: 400
Quote:
Originally Posted by Kpjoslee View Post

I need to see the DX12 game that heavily stresses the GPU in more traditional sense rather than the game that is heavy in number of units and AI going on, as it would put different stress on GPUs to get a better picture of the whole thing.

I've been playing it. 90 percent or better most of the time on the GPU. The GPU is being stressed. It sits about 50 percent on all cores on the CPU. I've never seen my Titan X temp so high under gaming load. And that's just with two players. Smooth as silk most of the time @ 4K. This is going to be a great game... when they get full strategic zoom going.

Whereas with Sins or Sup Com, you'll be lucky to get out of 2d clocks. GPU has nothing to do with those games... not like this game.
spacin9 is offline  
post #1408 of 2682 (permalink) Old 08-30-2015, 11:45 PM
Zen
Kpjoslee Level 1
 
Kpjoslee's Avatar
 
Join Date: Jan 2013
Location: Somewhere in US.
Posts: 938
Quote:
Originally Posted by spacin9 View Post

I've been playing it. 90 percent or better most of the time on the GPU. The GPU is being stressed. It sits about 50 percent on all cores on the CPU. I've never seen my Titan X temp so high under gaming load. And that's just with two players. Smooth as silk most of the time @ 4K. This is going to be a great game... when they get full strategic zoom going.

Whereas with Sins or Sup Com, you'll be lucky to get out of 2d clocks. GPU has nothing to do with those games... not like this game.

Ah, good to know. I meant to say in terms of games that heavily focuses on graphics in a traditional sense, like adventure, RPG, or FPS. +Rep for you biggrin.gif


Quote:
Originally Posted by Kollock View Post

I think you are confusing a few issues. Tier 2 vs Tier 3 binding is a completely separate issue from Async Compute. It's has to do with the number of root level descriptors we can pass. In tier 3, it turns out we can basically never update a descriptor during a frame, but in tier 2 we sometimes have to build a few . I don't think it's a significant performance issue though, just a techinical detail.

In regards to the purpose of Async compute, there are really 2 main reasons for it:

1) It allows jobs to be cycled into the GPU during dormant phases. In can vaguely be thought of as the GPU equivalent of hyper threading. Like hyper threading, it really depends on the workload and GPU architecture for as to how important this is. In this case, it is used for performance. I can't divulge too many details, but GCN can cycle in work from an ACE incredibly efficiently. Maxwell's schedular has no analog just as a non hyper-threaded CPU has no analog feature to a hyper threaded one.

2) It allows jobs to be cycled in completely out of band with the rendering loop. This is potentially the more interesting case since it can allow gameplay to offload work onto the GPU as the latency of work is greatly reduced. I'm not sure of the background of Async Compute, but it's quite possible that it is intended for use on a console as sort of a replacement for the Cell Processors on a ps3. On a console environment, you really can use them in a very similar way. This could mean that jobs could even span frames, which is useful for longer, optional computational tasks.

It didn't look like there was a hardware defect to me on Maxwell just some unfortunate complex interaction between software scheduling trying to emmulate it which appeared to incure some heavy CPU costs. Since we were tying to use it for #1, not #2, it made little sense to bother. I don't believe there is any specific requirement that Async Compute be required for D3D12, but perhaps I misread the spec.

Regarding trying to figure out bottlenecks on GPUS, it's important to note that GPUs do not scale simply by adding more cores to it, especially graphics tasks which have alot of serial points. My $.02 is that GCN is a bit triangle limited, which is why you see greater performance on 4k, where the average triangle size is 4x the triangle size of 1080p.

I think you're also being a bit short-sighted on the possible use of compute for general graphics. It is not limited to post process. Right now, I estimate about 20% of our graphics pipeline occurs in compute shaders, and we are projecting this to be more then 50% on the next iteration of our engine. In fact, it is even conceivable to build a rendering pipeline entirely in compute shaders. For example, there are alternative rendering primitives to triangles which are actually quite feasible in compute. There was a great talk at SIGGRAPH this year on this subject. If someone gave us a card with only compute pipeline, I'd bet we could build an engine around it which would be plenty fast. In fact, this was the main motivating factors behind the Larabee project. The main problem with Larabee wasn't that it wasn't fast, it was that they failed to be able to map dx9 games to it well enough to be a viable product. I'm not saying that the graphics pipeline will disappear anytime soon (or ever), but it's by no means certain that it's necessary. It's quite possible that in 5 years time Nitrous's rendering pipeline is 100% implemented via compute shaders.

Thank you for some clarifications. thumb.gif

My home PC
(15 items)
CPU
AMD Threadripper 1950x
Motherboard
Gigabyte Aorus X399 Gaming 7
GPU
EVGA Geforce RTX 2080 Ti XC Ultra
RAM
G.Skill DDR4 3600 CL16
Hard Drive
Samsung Evo 840 500GB
Hard Drive
Samsung 960 Pro 500GB
Power Supply
EVGA SuperNova G2 1300W
Cooling
Noctua NH-U14S TR4
Case
Corsair Carbide Air 540
Operating System
Windows 10 Pro
Monitor
Dell U2711
Monitor
Samsung 55" 4k
Keyboard
Corsair K70
Mouse
Logitech G502
Audio
Denon AVR-X3300W
▲ hide details ▲
Kpjoslee is offline  
post #1409 of 2682 (permalink) Old 08-30-2015, 11:56 PM
New to Overclock.net
spacin9 Level 1
 
spacin9's Avatar
 
Join Date: Apr 2013
Posts: 400
Quote:
Originally Posted by Kpjoslee View Post

Ah, good to know. I meant to say in terms of games that heavily focuses on graphics in a traditional sense, like adventure, RPG, or FPS. +Rep for you biggrin.gif
Thank you for some clarifications. thumb.gif

I understand what you're saying. I guess you're looking for post-process and ambient performance... chromatic whatever and such.

But it is completely out of bounds for a huge scale RTS to be so GPU dependent.

If the devs are reading this, can't wait for the tech upgrades to be unlocked. Hope there are some neat super-weapons and experimentals...

I'm coming into it completely cold of course. I saw and I bought. I could use a play on words with the popular Roman equivalent of my last sentence, but I'd be banned [again] for it. smile.gif

I have no idea what this game is going to be. But looks great so far...thumb.gif
spacin9 is offline  
post #1410 of 2682 (permalink) Old 08-31-2015, 12:07 AM
Zen
Kpjoslee Level 1
 
Kpjoslee's Avatar
 
Join Date: Jan 2013
Location: Somewhere in US.
Posts: 938
Quote:
Originally Posted by spacin9 View Post

I understand what you're saying. I guess you're looking for post-process and ambient performance... chromatic whatever and such.

But it is completely out of bounds for a huge scale RTS to be so GPU dependent.

If the devs are reading this, can't wait for the tech upgrades to be unlocked. Hope there are some neat super-weapons and experimentals...

I'm coming into it completely cold of course. I saw and I bought. I could use a play on words with the popular Roman equivalent of my last sentence, but I'd be banned [again] for it. smile.gif

I have no idea what this game is going to be. But looks great so far...thumb.gif

After all this long talk about technical stuff, I almost forgot this is a game biggrin.gif
Game certainly looks promising, definitely on the list of future purchases. thumb.gif

My home PC
(15 items)
CPU
AMD Threadripper 1950x
Motherboard
Gigabyte Aorus X399 Gaming 7
GPU
EVGA Geforce RTX 2080 Ti XC Ultra
RAM
G.Skill DDR4 3600 CL16
Hard Drive
Samsung Evo 840 500GB
Hard Drive
Samsung 960 Pro 500GB
Power Supply
EVGA SuperNova G2 1300W
Cooling
Noctua NH-U14S TR4
Case
Corsair Carbide Air 540
Operating System
Windows 10 Pro
Monitor
Dell U2711
Monitor
Samsung 55" 4k
Keyboard
Corsair K70
Mouse
Logitech G502
Audio
Denon AVR-X3300W
▲ hide details ▲
Kpjoslee is offline  
Closed Thread

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off