I can't tell who did this latest GPU and without spending a bunch of time
digging through code, I can't tell what its features are.  Is there an
actual GPU in there, or is it just memory and video controllers?  What kind
of rendering can it do?  The blog seems to imply that 2D isn't even
working.  We could do a dumb framebuffer in 2007.  There were some people
from Silicon Spectrum (that bought the IP from Number Nine) who were
looking at putting their design into the open source, and I think they did
a kickstarter.  But I can't easily find any info on the identity of the
people who are doing this one.

Sometimes I worry that the OGP may have inadvertently caused some trouble,
because every time someone else tries to do the open source GPU thing, the
response from the community is "been there, didn't do it."

If we had a synthesizable design in Verilog that COULD perform at cutting
edge performance, if only we could put it in a cutting edge chip
technology, then that might help.  I have bits and pieces, but I haven't
had time to work on it.

Currently, we've got a paper written that we're submitting for publication.
 It involves an open source GPU developed by another member of OGML.  And
we're looking at doing some other research in that area.  Once we have a
few papers, we can apply for grants.  We did try to get one masters student
to work on OGA2's ALU, but that didn't work out very well.  If we can get
funding, then we can justify spending a lot more time on it ourselves,
along with funding students. There's lots of room for research on GPUs, and
one or more completed synthesizable GPUs will emerge as byproducts.
 Currently, we have a graduate student collecting the latest research on
GPU memory systems (virtual memory, exceptions, page faults, TLBs, etc.).




On Sun, Aug 3, 2014 at 12:28 PM, Dieter BSD <[email protected]> wrote:

> There is another system.
>
> > The "GPLGPU" source-code is out as a GPLv3-licensed Verilog design to a
> > graphics engine.
>
> > The Verilog was tested on a Synopsys VCS simulator and the
> > Mentor Graphics Modelsim Questasim while it should be possible to
> > develop an ASIC around it or run it off an FPGA.
>
> http://www.phoronix.com/scan.php?page=news_item&px=MTc1NDY
> http://gplgpu.com/
>
> Let the compare-and-contrast discussions and class assignments begin.
> _______________________________________________
> Open-graphics mailing list
> [email protected]
> http://lists.duskglow.com/mailman/listinfo/open-graphics
> List service provided by Duskglow Consulting, LLC (www.duskglow.com)
>



-- 
Timothy Normand Miller, PhD
Assistant Professor of Computer Science, Binghamton University
http://www.cs.binghamton.edu/~millerti/
Open Graphics Project
_______________________________________________
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