SIGGRAPH 2009: Beyond Programmable Shading papers

There's an id Tech5 one!

 

I love the "We are hiring senior developers" slide in the DICE presentation. I want an office with a view like that.

 

Nice, thanks.

 

Yep, many thanks for the headsup man.

DICE's presentation was also very nice. I thought BF:BC2 was using the new version of the engine though.

 

I'd thought so too, but it looks like it's more of a 1.5 than a 2.0. The buildings will look prettier when they fall down, and that's all that matters to me.

I wish there was more information on how the job graph worked. A couple years ago I worked on an app that was my first foray into multi-threading, and I found it very hard to come up with a good way parallelize the workload. I could probably draw a state machine for the work to be performed on each piece of data. Looking at that job graph made me thing back to that app right away.

 

thanks reading now!

 

These job graphs are pretty simple. Each job has a list of other jobs that need to be done before it can be start. It mostly the same as having threads that needs to wait for other threads finished. One reason why game engines uses jobs instead of threads is that the overhead with a job based system is much lower. The other reason is that threads are normally could not execute on other “Processors”. This is currently mostly a PS3 problem but with the new compute features of the GPUs it becomes more interesting to execute such jobs on them.

 

That is the reaction I wanted, its a photo of my old desk. Since then I've moved a few steps and have _way_ more monitors

http://twitpic.com/bkqpu
http://twitpic.com/bkikv

 

There are tons of improvements to BFBC2, but many of them doesn't come from pure new tech but instead the team learning how to really utilize the engine, which will show.

Is there anything specific you are wondering about our job system?

 

That's wicked. What are the sizes of these monitors? I assume the hulky thing on the left is a TV?

I considered myself a freak for buying a 30" monitor, and wasn't sure if it would be worth the cash, but now I couldn't live without it.

 

Colour me jealous. I should take a picture of my cubicle at work. The contrast would be hilarious.

 

I'll see if I can come up with a couple questions, but most likely I'll take it to PM if I think they're too far off topic. I'm very new to the world of parallelism.

 

Since this is Stockholm in summer time that photo might be around 9-10pm or so, so I guess the main take away is that it's a lot of overtime.

 

Haha, no those images were take quite a few months ago. But yes I also do work quite long hours, but that is far from typical and not mandated in general.

 

PDFs 07, 08 & 09 seem to be corrupted, being only 17.3 kB each. Did anyone grad them before they went bad?

 

Just re-downloaded all of them successfully from the site. Sure it just isn't some download corruption from your side?

 

Thank you,

looking forward to giving those a closer look

BTW: Have anyone seen the slides from Cryteks talk, by Anton Kaplanyan, in the "Advances in Real-Time Rendering in 3D Graphics and Games I"?

 

Yes for those that aren't on twitter-time (hint hint) it can be found here

Kaplanyan, A.: Light Propagation Volumes in CryEngine 3
http://www.crytek.com/technology/presentations/
Abstract: "This is a SIGGRAPH 2009 talk from "Advanced Real-time Rendering Course" about new ligting technique in CryEngine 3. The new technique called "Light Propagation Volumes" is used for Global Illumination and Massive Lighting, which significantly improves the lighting quality. The concept and its implementation are explained in the paper. Slides from the talk with talk notes as well as the video are provided. "


Also Tim Sweeney's keynote "The End of the GPU Roadmap" at HPG (which was colocated with Siggraph) can be found here:
http://bit.ly/bniNt
Abstract: "Mainstream GPUs emerged 12 years ago and rapidly evolved from fixed-function graphics accelerators to semi-programmable computing devices. The speaker will argue that the GPU roadmap is now coming to an end, and that future PCs and game consoles will be powered by general computing devices, utilizing multi-core vector-processing units to run all code, graphics and non-graphics, uniformly, at Teraflop performance levels. He'll then give an overview of the design dimensions impacting such an architectures, such as caches, threads, vector sizes, and programming models."

 

I wonder why Mr. Sweeney likes analytical anti-aliasing so much. An a-buffer is one thing, but a BSP per pixel? Seriously? Just for edge AA? (It's not like say a micropolygon edge is the real edge anyway.)