Advertisement Jump to content
Sign in to follow this  
Jason Z

BUILD 2014 Sessions

This topic is 1736 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Advertisement


Direct3D 12 API Preview - This is probably the most anticipated talk for me, so hopefully we will find out more about the low level details on D3D12...

 

It's the talk I gave at GDC two weeks ago, just for the BUILD audience.  I didn't have a cold at GDC so I thought I did better the first time :).

Share this post


Link to post
Share on other sites

 


Direct3D 12 API Preview - This is probably the most anticipated talk for me, so hopefully we will find out more about the low level details on D3D12...

 

It's the talk I gave at GDC two weeks ago, just for the BUILD audience.  I didn't have a cold at GDC so I thought I did better the first time smile.png.

 

 

Didn't notice the cold; you did well.

 

I think we all are eagerly waiting to see the eventual answer to that backport to Windows 7 question ;)

Share this post


Link to post
Share on other sites

Do people still think they will target Windows 7? laugh.png


Not think; hope.

Win7 remains the largest market for AAA games on Steam, people are not jumping to Win8 and will probably continue to resist to even with the update, and while Win9 is around the same time period that might not see a huge jump either; thus DX12 support needs to go back downstream to Win7 from a AAA point of view - failure to do so will just result in a lot of people sighing loudly and continuing with DX11.

Now, this doesn't mean that WDDM2 needs to be back ported; I could see a situation where D3D12 sits on top of the WDDM version on Windows which means D3D12 is still limited to the exposed feature set of that OS; Win7 = D3D12_FEATURE_LEVEL_11 at most, Win8 = D3D12_FEATURE_LEVEL_11_1 and so on. I think most people would accept that as a pragmatic way forward (although I'd like to see things like multidrawindirect appear as that is useful to have in the pocket) even if the WDDM submission part takes a bit longer on the older OSes/versions.

Share this post


Link to post
Share on other sites

Not think; hope.


Win7 remains the largest market for AAA games on Steam, people are not jumping to Win8 and will probably continue to resist to even with the update, and while Win9 is around the same time period that might not see a huge jump either; thus DX12 support needs to go back downstream to Win7 from a AAA point of view - failure to do so will just result in a lot of people sighing loudly and continuing with DX11.

Now, this doesn't mean that WDDM2 needs to be back ported; I could see a situation where D3D12 sits on top of the WDDM version on Windows which means D3D12 is still limited to the exposed feature set of that OS; Win7 = D3D12_FEATURE_LEVEL_11 at most, Win8 = D3D12_FEATURE_LEVEL_11_1 and so on. I think most people would accept that as a pragmatic way forward (although I'd like to see things like multidrawindirect appear as that is useful to have in the pocket) even if the WDDM submission part takes a bit longer on the older OSes/versions.

 

Yes, this is the ideal and realistic best-case scenario.  I don't think anyone would really expect WDDM 2 to be backported.  What one would hope to expect is to be able to use the same API (i.e D3D12) for the bulk of one's program, excepting some D3D12-specific features (of which it seems there are relatively few), with only the DXGI layer being appreciatively different.

 

One would also expect that the performance improvements would come through even in cases where it's running on WDDM 1.x, otherwise it also seems pointless.  A D3D12 that runs at the same performance as 11 but with considerably more cost in code complexity on ~80% to ~90% of one's target audience isn't going to get much uptake.

Share this post


Link to post
Share on other sites

One would also expect that the performance improvements would come through even in cases where it's running on WDDM 1.x, otherwise it also seems pointless.  A D3D12 that runs at the same performance as 11 but with considerably more cost in code complexity on ~80% to ~90% of one's target audience isn't going to get much uptake.


You should still, in theory, get most of the performance improvement by virtue of the drivers being simpler, doing less work, going wide and all that so I don't see any issue there; sure on a trace the WDDM portion on Win7 might be (makes up number) 5% bigger than Win9/WDDM2.0 but that's not going to massively hurt you so you still get the wins smile.png

Share this post


Link to post
Share on other sites

 

One would also expect that the performance improvements would come through even in cases where it's running on WDDM 1.x, otherwise it also seems pointless.  A D3D12 that runs at the same performance as 11 but with considerably more cost in code complexity on ~80% to ~90% of one's target audience isn't going to get much uptake.


You should still, in theory, get most of the performance improvement by virtue of the drivers being simpler, doing less work, going wide and all that so I don't see any issue there; sure on a trace the WDDM portion on Win7 might be (makes up number) 5% bigger than Win9/WDDM2.0 but that's not going to massively hurt you so you still get the wins smile.png

 

 

I'd be happy with just being able to shove in a few tens-of-thousands (equivalent of) draw calls without needing to do any complex sorting or batching to be honest.  And on reflection, a D3D12 that let me do that at even 75% the perf of a complex sorted/batched D3D11 would even be acceptable.  Everything else would be gravy.

Share this post


Link to post
Share on other sites

There have been multiple people from Microsoft that publicly acknowledged the fact that Win7 is highly in demand for D3D12 - so I think the message is getting through.  Now we have to wait and see what happens...

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!