[Librecores Discussion] [GSoC2017] High-level bitstream format

Martin d'Anjou martin.danjou14 at gmail.com
Fri May 26 00:34:24 CEST 2017


Hi Stefan,

We had a short meeting today on the GSoC High-level bitstream format
project.

The school semester has just come to an end in Germany, and the project
is expect to be starting as planned. School continues through the GSoC
summer months, and Roland has organized his time to devote 30 hours per
week on average to the GSoC project.

The preferred mode of communication will be email rather than chat room.
We will be using the librecores discuss mailing list with the
[Librecores Discussion] [GSoC2017] High-level bitstream format subject
for discussion, questions, and answers. If this becomes too much
traffic, we should create a public mailing list to remain in the spirit
of GSoC open public communications.

We talked about the input format the tool will take. Two input formats
are possible: start from the binary format, or start from the text
format. The text is probably the best approach since all icestorm tools
work based on this format. Two approaches are possible regarding the
tool, either modify an existing icestorm tool to output the
user-understandable format, or create a new tool. The later makes more
sense at this time, and it may be possible to leverage some existing
icestorm library code.

I deferred a question to you regarding the FOSSi foundation (no rush).

I will be digging inside my PLL book to refresh my memory on practical
applications of phase-locked loops in FPGAs. Until I find my book, I
find this video to be useful:
https://www.youtube.com/watch?v=0jzLDe950AY. There is a short high-level
example at the top of this page:
http://digital.ni.com/public.nsf/allkb/07BC8D77D4E9AE258625708B007CE74F
(though I find the Figure 1 unclear). Lastly, page 3 of this very long
pdf shows how a PLL can multiply the frequency of an input reference
(http://www.cppsim.com/PLL_Lectures/day1_am.pdf), which is useful say
for double-data-rate interfaces (DDR) where the reference clock is half
the frequency of the data rate. I hope this helps!

The next meeting will be on June 8 at 13:00 UTC on mumble.

Best,

Martin




More information about the Discussion mailing list