Re: LLVM and supercollider. Was Re: symposium thought

OK,

   LLVM is a C++ library that could be linked with sclang and scsynth that would provide highly optimised just in time compiled native code. It's getting used in "nextgen" scripting languages like the rubinius ruby vm (google it if interested).

   In sclang what would have to happen is either

   1)the current bytecode would have to be translated into llvm bytecode inside sclang before passing it over to the llvm compilation process.

   2)the sclang virtual machine could be ditched entirely and sclang code compiled directly to LLVM bytecode, before being handed over to the llvm compilation process.

   either way would result in a very big speed boost for sclang.

   For scsynth:

      How'd you like to write unit generators in sclang? They could probably be made to run faster than those written in C++: instead of each unit generator being optimised by a C++ compiler in isolation, all the graphs on scsynth could be just in time compiled and optimised as one whole program.  You could do the single sample stuff without anything like the compilation/link/load into process latency mentioned as the killer in that keynote (that alas I didn't see being in Melbourne).

   It would also speed up block based ugen graphs, but who knows how much by.

   For both scsynth and sclang it would take quite a bit of work, but the results could be spectacular I think.

Lorien

Andrea Valle wrote:

(Thanks James)

Lorien, humble question: really cannot understand a word, but curious....

-a-

On 14 Apr 2009, at 15:47, Lorien Dunn wrote:

A comment here as you've turned them off there:

I read about the single sample SC in computer music journal just yesterday and thought to myself what a wonderful thing the Low Level Virtual Machine project (http://llvm.org) with it's optimisers and JIT compilers could be for  both sclang and scsynth. No need for C code these days :)

cheers,

Lorien

James Harkins wrote:

+1  -- I really appreciated seeing the hugely divergent ways we are using this fantastic tool.

In response to Andrea's question about notes on the symposium: I've started on something here:

http://www.dewdrop-world.net/words/

hjh

On Apr 14, 2009, at 1:34 AM, Sam Pluta wrote:

hey all,

i just wanted to write a quick note about the symposium this past weekend.  i said this to a couple of people, but what impressed me most about this gathering was how wonderfully different everyone's work was.  nothing presented was at all similar to anything else - with people doing everything from live coding to installations to interface driven performance to live manipulation -  and this is a sign that we are on to something as a group.

: H. James Harkins

: jamshark70@... <jamshark70@...>

: http://www.dewdrop-world.net

.::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:

"Come said the Muse,

Sing me a song no poet has yet chanted,

Sing me the universal."  -- Whitman

_______________________________________________

sc-users mailing list

info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml

archive: http://www.listarc.bham.ac.uk/marchives/sc-users/

search: http://www.listarc.bham.ac.uk/lists/sc-users/search/

--------------------------------------------------

Andrea Valle

--------------------------------------------------

CIRMA - DAMS

Università degli Studi di Torino

--> http://www.cirma.unito.it/andrea/

--> http://www.myspace.com/andreavalle

--> http://www.flickr.com/photos/vanderaalle/

-->  http://www.youtube.com/user/vanderaalle

--> andrea.valle@... <andrea.valle@...>

--------------------------------------------------

" This is a very complicated case, Maude. You know, a lotta ins, a lotta outs, a lotta what-have-yous." (Jeffrey 'The Dude' Lebowski)

_______________________________________________

sc-users mailing list

info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml

archive: http://www.listarc.bham.ac.uk/marchives/sc-users/

search: http://www.listarc.bham.ac.uk/lists/sc-users/search/