Hi Charles,
Charles Cui <charles.cui1984@xxxxxxxxx> writes:
[...]
> This is Charles Cui from Tsinghua University, China. I am pretty
> interested in the ideas listed in GSOC 2016 pages for PCP, especially
> the idea of "performance metric extraction using go language agents".
> As a PhD student in Tsinghua, I used C for roughly 4 years. At the
> same time, I have golang experience for the production code. Besides,
> I speak fluent English and am available the full summer. I am not sure
> whether I can join in the development of PCP? In case my project is
> selected by others, I am able to pick others.
Thanks for your interest in PCP! Currently, no proposed project has
been taken yet (official applications start on March 14), so no
worries there. It's great to see that you've subscribed to the
mailing list, we've also got a irc channel on freenode.net (#pcp) if
you prefer that form of communication.
Sounds like you've got some relevant programming language experience,
have you contributed to any open source projects, or Google Summer of
Code before?
A few more details about the task itself:
So, from a metrics collection standpoint, Performance Co-Pilot has two
components. The first is the pmcd (1) (Performance Metrics Collector
Daemon), which orchestrates client requests, and gathers the requested
metrics from the individual pmdas (Performance Metric Domain Agents).
Each of which define and collect a certain subset of metrics from the
system in question. The pmcd is written in C, and currently we have
pmdas (the agents), written in C/C++, perl, and python. We're looking
to extend pmdas to Golang as well, which will include writing some of
the standard pmda functions, as well as libraries, to interface between
pmdas written in Golang, to the C level apis and libraries which
currently exist.
If you're particularly keen to read more about the technical
implementation details within PCP, we've got plenty of documentation on
the website[1], specifically the Programmer's Guide[2] (Chapter Two is
specifically for writing a pmda).
Please feel free to reach out with any more specific questions you might
have, regardless if it's about the task, or the PCP community as a whole.
Cheers,
Lukas
[1] - http://pcp.io/documentation.html
[2] - http://pcp.io/books/PCP_PG/pdf/PCP-3-pcp-programmers-guide-en-US.pdf
|