Optimization for miranda

• Optimize cava and edge detection schedules (CPU schedules are multi-core now).
• Add fork-extend pass to make fork dimensions multiple of some number (closes #28 (closed)).
• Make GCM add extra dimensions to per-thread arrays in fork-joins.
• Make GCM allocate extra memory for calls in parallel, make RT backend use different memories per parallel call.
• Make not finding xdot binary crash the compiler.
• Some misc. fixes.
• Add conv-like fork-join test.

juno_samples/fork_join_tests/src/gpu.sch

@@ -73,23 +73,7 @@ dce(auto.test8);
 simplify-cfg(auto.test8);
 dce(auto.test8);
 
-fork-split(auto.test9@filter_loop);
-fork-unroll(auto.test9);
-fork-unroll(auto.test9);
-dce(auto.test9);
-ccp(auto.test9);
-gvn(auto.test9);
-phi-elim(auto.test9);
-dce(auto.test9);
-fixpoint {
-  predication(auto.test9);
-  simplify-cfg(auto.test9);
-}
-ccp(auto.test9);
-gvn(auto.test9);
-phi-elim(auto.test9);
-lift-dc-math(auto.test9);
-dce(auto.test9);
+no-memset(test9@const);
 
 ip-sroa(*);
 sroa(*);