llvm dev - Apr 2012 - [LLVMdev] Potential Google Summer of Code Applicant

Hi, my name is Patrick Edwards, and I'm currently a CS major at Kent State
University. I have always been interested in doing work with compilers and
LLVM seems to be a perfect fit for me to learn more over the summer, and
also contribute to open-source projects at the same time. However, while
browsing through the project ideas, the only ideas I found accessible were
the code reduction and compile with/benchmark the LLVM compiler. I would
really love to help LLVM, as I have used C++ in quite a few of my classes
and side programs, not to mention learning other oddball languages as I
wanted. If possible, could someone please point me in the right direction
to contribute to LLVM in the best way I can?

Thank you in advance,
Patrick Edwards, potential GSoC applicant
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120401/e35713ed/attachment.html>

On 04/02/2012 05:41 AM, Patrick Edwards wrote:
> Hi, my name is Patrick Edwards, and I'm currently a CS major at Kent
> State University. I have always been interested in doing work with
> compilers and LLVM seems to be a perfect fit for me to learn more over
> the summer, and also contribute to open-source projects at the same
> time. However, while browsing through the project ideas, the only ideas
> I found accessible were the code reduction and compile with/benchmark
> the LLVM compiler. I would really love to help LLVM, as I have used C++
> in quite a few of my classes and side programs, not to mention learning
> other oddball languages as I wanted. If possible, could someone please
> point me in the right direction to contribute to LLVM in the best way I
can?
>
> Thank you in advance,
> Patrick Edwards, potential GSoC applicant
Hi Patrick,

besides the open project pages of the LLVM itself [1], there are also 
the ideas list of the subprojects:

- The clang open projects list [2]
- The SAFECode open projects list [3]
- The Polly todo list [4]

As you already realized, not all of them are suiteable for summer of 
code, but some of them definitely are.

A topic I personally would think might be an interesting summer of code 
project is a clang based gnu 'indent' replacement. Here the idea from 
the clang open projects page:

"Use clang libraries to implement better versions of existing tools: 
Clang is built as a set of libraries, which means that it is possible to 
implement capabilities similar to other source language tools, improving 
them in various ways. Three examples are distcc, the delta testcase 
reduction tool, and the "indent" source reformatting tool. distcc can
be
improved to scale better and be more efficient. Delta could be faster 
and more efficient at reducing C-family programs if built on the clang 
preprocessor, indent could do proper formatting for complex C++ 
features, and it would be straight-forward to extend a clang-based 
implementation to handle simple structural rules like those in the LLVM 
coding standards."

I am not an expert in clang itself, but on the clang mailing list, you 
there are plenty of people who are. I copied Manuel, as I think he was 
already planning to work on something like this.

Cheers
Tobi

[1] http://llvm.org/OpenProjects.html
[2] http://clang.llvm.org/OpenProjects.html
[3] http://sva.cs.illinois.edu/projects.html
[4] http://polly.llvm.org/todo.html

+cfe-dev, +chandlerc

On Mon, Apr 2, 2012 at 11:05 AM, Tobias Grosser <tobias at grosser.es>
wrote:
> On 04/02/2012 05:41 AM, Patrick Edwards wrote:
>>
>> Hi, my name is Patrick Edwards, and I'm currently a CS major at
Kent
>> State University. I have always been interested in doing work with
>> compilers and LLVM seems to be a perfect fit for me to learn more over
>> the summer, and also contribute to open-source projects at the same
>> time. However, while browsing through the project ideas, the only ideas
>> I found accessible were the code reduction and compile with/benchmark
>> the LLVM compiler. I would really love to help LLVM, as I have used C++
>> in quite a few of my classes and side programs, not to mention learning
>> other oddball languages as I wanted. If possible, could someone please
>> point me in the right direction to contribute to LLVM in the best way I
>> can?
>>
>> Thank you in advance,
>> Patrick Edwards, potential GSoC applicant
>
>
> Hi Patrick,
>
> besides the open project pages of the LLVM itself [1], there are also the
> ideas list of the subprojects:
>
> - The clang open projects list [2]
> - The SAFECode open projects list [3]
> - The Polly todo list [4]
>
> As you already realized, not all of them are suiteable for summer of code,
> but some of them definitely are.
>
> A topic I personally would think might be an interesting summer of code
> project is a clang based gnu 'indent' replacement. Here the idea
from the
> clang open projects page:
>
> "Use clang libraries to implement better versions of existing tools:
Clang
> is built as a set of libraries, which means that it is possible to
implement
> capabilities similar to other source language tools, improving them in
> various ways. Three examples are distcc, the delta testcase reduction tool,
> and the "indent" source reformatting tool. distcc can be improved
to scale
> better and be more efficient. Delta could be faster and more efficient at
> reducing C-family programs if built on the clang preprocessor, indent could
> do proper formatting for complex C++ features, and it would be
> straight-forward to extend a clang-based implementation to handle simple
> structural rules like those in the LLVM coding standards."
>
> I am not an expert in clang itself, but on the clang mailing list, you
there
> are plenty of people who are. I copied Manuel, as I think he was already
> planning to work on something like this.
While it is a topic where we'll want to see major stuff going on over
the coming year, I'm not sure that the underlying infrastructure will
meet the GSoC timeline - we're currently in the process of getting the
basics into mainline, and I think it would be a pain to now start
writing larger amounts of codes on something that we'll need to change
later.

Cheers,
/Manuel

On 4/1/12 10:41 PM, Patrick Edwards wrote:
> Hi, my name is Patrick Edwards, and I'm currently a CS major at Kent 
> State University. I have always been interested in doing work with 
> compilers and LLVM seems to be a perfect fit for me to learn more over 
> the summer, and also contribute to open-source projects at the same 
> time. However, while browsing through the project ideas, the only 
> ideas I found accessible were the code reduction and compile 
> with/benchmark the LLVM compiler. I would really love to help LLVM, as 
> I have used C++ in quite a few of my classes and side programs, not to 
> mention learning other oddball languages as I wanted. If possible, 
> could someone please point me in the right direction to contribute to 
> LLVM in the best way I can?
Here are two random ideas I've had:

1) Run tests to see how well Google's Go language works with Dragonegg.  
Fix bugs that you find.  Measure performance of Dragonegg-compiled Go 
programs compiled over GCC-compiled and 6g compiled Go programs.

2) Create a Clang front-end for Google Go.  This would enable static 
analysis to be built for Google Go just as they are built for 
C/C++/Objective-C now.

If you haven't already, you might want to investigate the Open Projects 
pages of LLVM sub-projects such as compiler-rt, SAFECode, polly, klee, etc.

-- John T.
>
> Thank you in advance,
> Patrick Edwards, potential GSoC applicant
>
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120402/e737d6eb/attachment.html>

I would really like to see someone work on LLVM's garbage collection
support - it hasn't been updated in 4 years, and while there's been a
lot
of talk about ways that it could be improved, there's been no action.

On Sun, Apr 1, 2012 at 8:41 PM, Patrick Edwards <pedwar11 at kent.edu>
wrote:
> Hi, my name is Patrick Edwards, and I'm currently a CS major at Kent
State
> University. I have always been interested in doing work with compilers and
> LLVM seems to be a perfect fit for me to learn more over the summer, and
> also contribute to open-source projects at the same time. However, while
> browsing through the project ideas, the only ideas I found accessible were
> the code reduction and compile with/benchmark the LLVM compiler. I would
> really love to help LLVM, as I have used C++ in quite a few of my classes
> and side programs, not to mention learning other oddball languages as I
> wanted. If possible, could someone please point me in the right direction
> to contribute to LLVM in the best way I can?
>
> Thank you in advance,
> Patrick Edwards, potential GSoC applicant
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>
>

-- 
-- Talin
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120405/d873c4a4/attachment.html>

On 4/6/12 2:21 AM, Talin wrote:
> I would really like to see someone work on LLVM's garbage collection 
> support - it hasn't been updated in 4 years, and while there's been
a
> lot of talk about ways that it could be improved, there's been no
action.
That is *sooo* true! :-) I'm one of the authors of an LLVM backend for 
Erlang (ErLLVM [1]); we have tested and measured our backend  and 
noticed that with the current GC infrastructure we see 20-40% 
performance degradation (because of the loads/stores on the stack for 
all gcroots). It is clear to me and the rest of the team that with this 
infrastructure the LLVM might not be suitable for languages with 
explicit garbage collection, like Erlang. I've also studied the way the 
Vmkit project handles GC and they seem to face the same deficiency too.

offtopic: I am working on an email (more like an RFC) with all the 
details and patches to the LLVM project in order to support our Erlang 
backend. I hope I will be able to send it by next week. Note, that we 
have already talked with the Ericsson/OTP team about integrating our 
work in a future release of Erlang/OTP (as a new HiPE backend).

Yiannis

[1]: http://erllvm.softlab.ntua.gr

-- 
Yiannis Tsiouris
Ph.D. student,
Software Engineering Laboratory,
National Technical University of Athens
WWW: http://www.softlab.ntua.gr/~gtsiour