More optimizaation passes (pending jll) (#1138)

* More optimizaation passes (pending jll)

* [Compiler] Add global option `SUM_TO_CONV`

* Update Reactant_jll and bumb version

---------

Co-authored-by: Mosè Giordano <[email protected]>
Co-authored-by: Mosè Giordano <[email protected]>

Author: wsmoses

Project.toml

@@ -1,7 +1,7 @@
 name = "Reactant"
 uuid = "3c362404-f566-11ee-1572-e11a4b42c853"
 authors = ["William Moses <[email protected]>", "Valentin Churavy <[email protected]>", "Sergio Sánchez Ramírez <[email protected]>", "Paul Berg <[email protected]>", "Avik Pal <[email protected]>", "Mosè Giordano <[email protected]>"]
-version = "0.2.64"
+version = "0.2.66"
 
 [deps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
@@ -87,7 +87,7 @@ PythonCall = "0.9"
 Random = "1.10"
 Random123 = "1.7"
 ReactantCore = "0.1.9"
-Reactant_jll = "0.0.123"
+Reactant_jll = "0.0.125"
 ScopedValues = "1.3.0"
 Scratch = "1.2"
 Sockets = "1.10"

src/Compiler.jl

@@ -459,6 +459,7 @@ end
 
 const WHILE_CONCAT = Ref(false)
 const DUS_TO_CONCAT = Ref(false)
+const SUM_TO_CONV = Ref(false)
 
 # Optimization passes via transform dialect
 function optimization_passes(;
@@ -654,8 +655,19 @@ function optimization_passes(;
         "elementwise_licm(0)",
         "concatenate_licm(0)",
         "slice_broadcast",
+        "while_pad_induction_reduction",
+        "while_licm<1>(1)",
+        "associative_common_mul_op_reordering",
+        "slice_select_to_select_slice",
+        "pad_concat_to_concat_pad",
+        "slice_if",
+        "dus_to_i32",
     ]
 
+    if SUM_TO_CONV[]
+        push!(transform_passes_list, "sum_to_conv")
+    end
+
     if WHILE_CONCAT[]
         push!(transform_passes_list, "while_concat")
     end