llvm dev - Mar 2016 - [GSoC'16] Add a Scheme Frontend for LLVM

Hi, everyone

This proposal covers the OpenProjects idea 'Port the Bigloo Scheme compiler,
from Manuel Serrano at INRIA Sophia-Antipolis, to output LLVM bytecode.' 
There is not a popular functional language frontend yet on LLVM. To my
knowledge, this is mostly because LLVM is lack of runtime support for the
functional language needs, and C foreign function interface calls are relatively
hard since LLVM leaves memory management to be handled by program. Bigloo is a
high performance(according to [1] ) scheme implementation and it designed to be
mostly R5RS compatible. It uses Boehm GC(also used in llvm-gcc) to manage C FFI
calls memory. Therefore bigloo is a proper choice for the scheme frontend of
LLVM. On the other hand, compile scheme to LLVM IR will make use of LLVM's
optimization techniques. The LLVM backend will boost the performance of bigloo. 
In this summer, I would like to build a robust, high-performance scheme frontend
for LLVM.

About Me:
I am Zhengyang, a M.S. candidate at Beijing University of Posts and
Telecomunications working in the area of static program analysis. My CV can be
found in [2] . Most of my works are based on Clang.  I have contributed to
Static Analysis Suite(SAS) from CERN PH-SFT as Google Summer of Code 2015
project. About the preparation for this proposal, I have studied the source
codes, serveral research papers and user manual of bigloo. I am interested in
working with LLVM community, and hoping to contribute to LLVM.

Proposed Work:
(Before GSoC'16 start) Add support for calling LLVM IR API on bigloo.
(Week 1-3) Create a new LLVM backend for bigloo scheme to make bigloo produce
LLVM IR.
(Week 4-6) Integrate Boehm garbage collector to the runtime.
(Week 7) Create project build/test infrastructure.
(Week 8-9) Run benchmarks and analysis the performance on the new scheme
compiler.

I have some questions for this idea:
1) Is this a good idea for a GSoC project ?
2) If any one is available to mentor me for this ?

Please let me know your comments on this proposal as a GSoC project. Any
comments on how to refine this proposal are welcome. I will be very grateful if
someone puts me in the right direction for the project.


Thanks,
Zhengyang Liu,
Beijing University of Posts and Telecommunications,
zhengyang-liu at hotmail.com.

[1]
http://www.cs.utah.edu/~mflatt/benchmarks-20100126/log1/Gambit_20benchmarks.html
[2] http://zhengyangl.com/cv.pdf
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160318/0de9f449/attachment.html>

FYI - This is way more work than is appropriate for a single summer.  
Bringing up a new frontend and getting good performance out of it is 
general several man years of work.  I strongly suggest you identify a 
subproject of a more reasonable scope and propose that.

p.s. Also, I'm not sure that porting an external open source project is 
really appropriate for an LLVM GSoC project.  Have you considered 
proposing something to the Bigloo project?  They might be better able to 
provide mentoring for what you're interested in.  Or alternatively, you 
might do well to seek an active mentor from both projects.

Philip

On 03/17/2016 09:57 PM, Zhengyang Liu via llvm-dev
wrote:
> Hi, everyone
>
> This proposal covers the OpenProjects idea ‘Port the Bigloo Scheme 
> compiler, from Manuel Serrano at INRIA Sophia-Antipolis, to output 
> LLVM bytecode.’  There is not a popular functional language frontend 
> yet on LLVM. To my knowledge, this is mostly because LLVM is lack of 
> runtime support for the functional language needs, and C foreign 
> function interface calls are relatively hard since LLVM leaves memory 
> management to be handled by program. Bigloo is a high 
> performance(according to [1] ) scheme implementation and it designed 
> to be mostly R5RS compatible. It uses Boehm GC(also used in llvm-gcc) 
> to manage C FFI calls memory. Therefore bigloo is a proper choice for 
> the scheme frontend of LLVM. On the other hand, compile scheme to LLVM 
> IR will make use of LLVM’s optimization techniques. The LLVM backend 
> will boost the performance of bigloo.  In this summer, I would like to 
> build a robust, high-performance scheme frontend for LLVM.
>
> About Me:
> I am Zhengyang, a M.S. candidate at Beijing University of Posts and 
> Telecomunications working in the area of static program analysis. My 
> CV can be found in [2] . Most of my works are based on Clang.  I have 
> contributed to Static Analysis Suite(SAS) from CERN PH-SFT as Google 
> Summer of Code 2015 project. About the preparation for this proposal, 
> I have studied the source codes, serveral research papers and user 
> manual of bigloo. I am interested in working with LLVM community, and 
> hoping to contribute to LLVM.
>
> Proposed Work:
> (Before GSoC’16 start) Add support for calling LLVM IR API on bigloo.
> (Week 1-3) Create a new LLVM backend for bigloo scheme to make bigloo 
> produce LLVM IR.
> (Week 4-6) Integrate Boehm garbage collector to the runtime.
> (Week 7) Create project build/test infrastructure.
> (Week 8-9) Run benchmarks and analysis the performance on the new 
> scheme compiler.
>
> I have some questions for this idea:
> 1) Is this a good idea for a GSoC project ?
> 2) If any one is available to mentor me for this ?
>
> Please let me know your comments on this proposal as a GSoC project. 
> Any comments on how to refine this proposal are welcome. I will be 
> very grateful if someone puts me in the right direction for the project.
>
>
> Thanks,
> Zhengyang Liu,
> Beijing University of Posts and Telecommunications,
> zhengyang-liu at hotmail.com.
>
> [1] 
>
http://www.cs.utah.edu/~mflatt/benchmarks-20100126/log1/Gambit_20benchmarks.html
> [2] http://zhengyangl.com/cv.pdf
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160322/1c5db789/attachment.html>