On Sun, Jan 06, 2013 at 12:10:19AM -0500, Timothy Normand Miller wrote:
> > Case in point: Who's paying Timothy's bills, and why has he been so
> > motivated to keep pushing OpenShader along?
> >
> >
> I was motivated to start the OGP in 2004 because at the time, there wasn't
> any off-the-shelf graphics card that had FOSS driver support.  ATI, for
> instance, had just decided to stop publishing docs for the Radeon 9200 (or
> some similar model at the time, I forget exactly which).  I thought it
> would be productive and fun to use my skills as a graphics chip designer to
> solve this problem.
> 
> I decided to change the direction of the project (the earlier approach had
> only very limited success anyhow) in early 2012 because I discovered that
> there were no particularly good GPU simulators and there was no
> open-architecture GPU at all.  It's completely different for CPUs, with
> plenty of open-architecture designs and really good cycle-accurage and
> energy-accurate simulators.  To get tenure, I'm going to have do something
> that makes a substantive impact on my field.  Why not do it an area I'm
> really passionate about?  Why not leverage something already going on?  Why
> not have a positive impact on both the FOSS community and the academic
> community at the same time, with the same project?  Also, I think it would
> be a lot of fun and a really compelling challenge solve this gaping problem.
> 
> Also, I go to work every day to pay my bills.  And I certainly don't think
> OpenShader is going to change that (at least not for the better).

I can think of about 15 different reasons why open-architecture GPU simulators
would be not only be a tenure-worthy project, but a great DOE Exascale computing
sub-project. Everyone is on the GPU bandwagon for HPC it seems.

Before you spend to much time writing generate statements, also take a look
at http://milkymist.org/3/migen.html, and http://networkx.github.com/ (which
it uses). It might be worth contacting some of the Networkx users for 
collaboration... see:
http://networkx.github.com/documentation/latest/bibliography.html#ba02

I think you got lucky in that you're in a position to do interesting open
hardware work that ALSO lines up well with some interesting research problems.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to