Changes between Version 24 and Version 25 of Commentary/Compiler/IntegratedCodeGen

Timestamp:

08/26/08 05:57:53 (5 years ago)

Author:

dias

Comment:

--

Commentary/Compiler/IntegratedCodeGen

@@ -30 +30 @@
 our design are as follows:
 
-  0. Build two big hammers, and hit as many nails as possible.  (The big hammers are the '''dataflow optimization engine''' and a '''coalescing register allocator.''')  The hammer itself may be big and complicated, but '''using a big hammer should be easy''' and should give easily predictable results.
+  0. Build two big hammers, and hit as many nails as possible.  (The big hammers are the '''dataflow optimization engine''' and a '''coalescing register allocator.''' For more on their uses, see our [wiki:Commentary/Compiler/IntegratedCodeGen#Designphilosophy design philosophy].)  The hammer itself may be big and complicated, but '''using a big hammer should be easy''' and should give easily predictable results.
   0. Load all back ends into every instance of the compiler, and '''treat every compilation as a cross-compilation.'''  Despite having been used in production compilers for at least twenty years, this technique is still seen as somewhat unorthodox, but it removes many {{{#ifdef}}}s and saves significant complexity at compiler-configuration time. Removing {{{#ifdef}}}s also mitigates problems with  validating the compiler under different build configurations.
 
@@ -47 +47 @@
    optimizations to create an integrated "superoptimization" that is
    strictly more powerful than any sequence of individual optimizations,
-   no matter how many times they are re-run
-   [http://portal.acm.org/citation.cfm?id=503298 (Lerner, Grove, and Chambers 2002)].
+   no matter how many times they are re-run.
+   The dataflow engine is based on 
+   [http://citeseer.ist.psu.edu/old/lerner01composing.html (Lerner, Grove, and Chambers 2002)];
+   you can find a functional implementation presented in
+   [http://www.cs.tufts.edu/~nr/pubs/zipcfg-abstract.html (Ramsey and Dias 2005)].
 
  * '''Coalescing register allocator:''' The back end can use fresh temporaries and register-register moves
@@ -54 +57 @@
    will eliminate almost all move instructions.
 
- * Our ultimate goal is to make adding a new back end easy as well.
+ * '''Back ends:''' Our ultimate goal is to make adding a new back end easy as well.
    In the long run, we wish to apply John Dias's dissertation work to GHC.
    In the short run, however, we