[FoRK] Reconfigurable Code Re: Software hacks using timestamp counters

Noon Silk noonslists at gmail.com
Mon Oct 1 15:47:13 PDT 2012


On Tue, Oct 2, 2012 at 7:35 AM, Dr. Ernest Prabhakar
<drernie at radicalcentrism.org> wrote:
> Hi all,
>
> On Oct 1, 2012, at 9:23 AM, Stephen Williams <sdw at lig.net> wrote:
>> We need buffer, L2, core, device (CPU/GPU/DSP), and parameter management
>> and optimization while chaining a number of compute intensive modules on large
>> amounts of mostly use-once data in, for some modes, a highly repetitive
>> environment.  All while keeping the functional code clean and highly reconfigurable
>> at compile or runtime with several alternate versions.
>
> That reminds me -- have we discussed Halide here?
>
> Decoupling Algorithms from Schedules for Easy Optimization of Image Processing Pipelines
> Halide
>
> I've long though we need a way to decouple the "expression" of the code's intent from the
> "mechanism" of its efficient implementation.  Halide is the closest I've seen to a real-world articulation of that philosophy.
>
> Do you know anyone else who's attempted this? Any thoughts on whether this will prove useful?

Link is here, for anyone too lazy to search:
http://people.csail.mit.edu/jrk/halide12/

Yeah, I've read over the Halide paper; I've not done anything with it
in particular, but I do think the general idea is a very interesting
one, and probably worth considering in terms of decoupling different
things.


> -- Ernie P.

-- 
Noon Silk

Fancy a quantum lunch? https://sites.google.com/site/quantumlunch/

"Every morning when I wake up, I experience an exquisite joy — the joy
of being this signature."



More information about the FoRK mailing list