This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [GSoC Project Proposal] ISO C11 threads.h implementation in GNU C Library
- From: Juan Manuel Torres Palma <j dot m dot torrespalma at gmail dot com>
- To: Andrea Francesco Iuorio <andreafrancesco dot iuorio at gmail dot com>
- Cc: libc-alpha at sourceware dot org
- Date: Thu, 27 Feb 2014 23:48:36 +0100
- Subject: Re: [GSoC Project Proposal] ISO C11 threads.h implementation in GNU C Library
- Authentication-results: sourceware.org; auth=none
- References: <CAD82F-pDTRUi1Nh6YiQ-Mx21m6K0NqqvfqZ6ayU9aAACeaVP+w at mail dot gmail dot com> <1393531628 dot 23480 dot 1 dot camel at SamsungRF510> <20140227202310 dot GS184 at brightrain dot aerifal dot cx> <Pine dot LNX dot 4 dot 64 dot 1402272108280 dot 1305 at digraph dot polyomino dot org dot uk> <20140227212605 dot GT184 at brightrain dot aerifal dot cx> <Pine dot LNX dot 4 dot 64 dot 1402272132500 dot 1305 at digraph dot polyomino dot org dot uk> <CAMaJsKhK2-p1zXQ5Q5_Jzt1ugaWHK9Wzi+c3u6r+zgJ0=5_g5w at mail dot gmail dot com>
I just took some time trying to research and understand what Rich was
suggesting. To be honest I really like that idea of using a lower
level approach, due to a precious improvement on performance (that's
the pro I can difference) but I have to agree with Joseph cause there
are some disadvantages that are really important and we must be aware
of them:
- The project must be completed in two months.
- I have no experience in such a low level, so I would need extra time
to be comfortable using futex and atomic libraries, including that I
will also need some asm programming.
- The POSIX thread library is a good start to get a working API in
those two months, and I am already used to work with it.
On the other hand, I think it may be pretty interesting to make a
research about the possible differences and pros/cons on doing one or
another implementation, even though I know I won't be able to finish
it on time with a futex/atomic approach, so just as an academic
research. I consider that proposal very interesting so I can do that
research if more guys here support that idea. Maybe in future it could
be ported.
Thank you all guys for the feedback. As soon as you think this is an
acceptable program to work on the project I will start working in a
more specific and detailed solution description.
--
Juan Manuel "Lolo" Torres Palma.
Computer Science Student at Universidad de Granada.