chore(ci): Add non determinism check and fixes (#6847)

compiler/noirc_driver/src/lib.rs

@@ -150,6 +150,10 @@ pub struct CompileOptions {
     /// A lower value keeps the original program if it was smaller, even if it has more jumps.
     #[arg(long, hide = true, allow_hyphen_values = true)]
     pub max_bytecode_increase_percent: Option<i32>,
+
+    /// Used internally to test for non-determinism in the compiler.
+    #[clap(long, hide = true)]
+    pub check_non_determinism: bool,
 }
 
 pub fn parse_expression_width(input: &str) -> Result<ExpressionWidth, std::io::Error> {

compiler/noirc_evaluator/src/brillig/brillig_ir/artifact.rs

@@ -27,7 +27,7 @@ pub(crate) struct GeneratedBrillig<F> {
     pub(crate) locations: BTreeMap<OpcodeLocation, CallStack>,
     pub(crate) error_types: BTreeMap<ErrorSelector, ErrorType>,
     pub(crate) name: String,
-    pub(crate) procedure_locations: HashMap<ProcedureId, (OpcodeLocation, OpcodeLocation)>,
+    pub(crate) procedure_locations: BTreeMap<ProcedureId, (OpcodeLocation, OpcodeLocation)>,
 }
 
 #[derive(Default, Debug, Clone)]
@@ -61,7 +61,7 @@ pub(crate) struct BrilligArtifact<F> {
     /// This is created as artifacts are linked together and allows us to determine
     /// which opcodes originate from reusable procedures.s
     /// The range is inclusive for both start and end opcode locations.
-    pub(crate) procedure_locations: HashMap<ProcedureId, (OpcodeLocation, OpcodeLocation)>,
+    pub(crate) procedure_locations: BTreeMap<ProcedureId, (OpcodeLocation, OpcodeLocation)>,
 }
 
 /// A pointer to a location in the opcode.

compiler/noirc_evaluator/src/ssa/opt/unrolling.rs

@@ -18,6 +18,8 @@
 //!
 //! When unrolling ACIR code, we remove reference count instructions because they are
 //! only used by Brillig bytecode.
+use std::collections::BTreeSet;
+
 use acvm::{acir::AcirField, FieldElement};
 use im::HashSet;
 
@@ -117,7 +119,7 @@ pub(super) struct Loop {
     back_edge_start: BasicBlockId,
 
     /// All the blocks contained within the loop, including `header` and `back_edge_start`.
-    pub(super) blocks: HashSet<BasicBlockId>,
+    pub(super) blocks: BTreeSet<BasicBlockId>,
 }
 
 pub(super) struct Loops {
@@ -238,7 +240,7 @@ impl Loop {
         back_edge_start: BasicBlockId,
         cfg: &ControlFlowGraph,
     ) -> Self {
-        let mut blocks = HashSet::default();
+        let mut blocks = BTreeSet::default();
         blocks.insert(header);
 
         let mut insert = |block, stack: &mut Vec<BasicBlockId>| {

compiler/noirc_frontend/src/ast/statement.rs

@@ -1,5 +1,4 @@
 use std::fmt::Display;
-use std::sync::atomic::{AtomicU32, Ordering};
 
 use acvm::acir::AcirField;
 use acvm::FieldElement;
@@ -841,10 +840,9 @@ impl ForRange {
         block: Expression,
         for_loop_span: Span,
     ) -> Statement {
-        /// Counter used to generate unique names when desugaring
-        /// code in the parser requires the creation of fresh variables.
-        /// The parser is stateless so this is a static global instead.
-        static UNIQUE_NAME_COUNTER: AtomicU32 = AtomicU32::new(0);
+        // Counter used to generate unique names when desugaring
+        // code in the parser requires the creation of fresh variables.
+        let mut unique_name_counter: u32 = 0;
 
         match self {
             ForRange::Range(..) => {
@@ -855,7 +853,8 @@ impl ForRange {
                 let start_range = ExpressionKind::integer(FieldElement::zero());
                 let start_range = Expression::new(start_range, array_span);
 
-                let next_unique_id = UNIQUE_NAME_COUNTER.fetch_add(1, Ordering::Relaxed);
+                let next_unique_id = unique_name_counter;
+                unique_name_counter += 1;
                 let array_name = format!("$i{next_unique_id}");
                 let array_span = array.span;
                 let array_ident = Ident::new(array_name, array_span);
@@ -888,7 +887,7 @@ impl ForRange {
                 }));
                 let end_range = Expression::new(end_range, array_span);
 
-                let next_unique_id = UNIQUE_NAME_COUNTER.fetch_add(1, Ordering::Relaxed);
+                let next_unique_id = unique_name_counter;
                 let index_name = format!("$i{next_unique_id}");
                 let fresh_identifier = Ident::new(index_name.clone(), array_span);
 

tooling/nargo_cli/build.rs

@@ -215,6 +215,9 @@ fn test_{test_name}(force_brillig: ForceBrillig, inliner_aggressiveness: Inliner
         // Set the maximum increase so that part of the optimization is exercised (it might fail).
         nargo.arg("--max-bytecode-increase-percent");
         nargo.arg("50");
+
+        // Check whether the test case is non-deterministic
+        nargo.arg("--check-non-determinism");
     }}
 
     {test_content}

tooling/nargo_cli/src/cli/compile_cmd.rs

@@ -216,6 +216,20 @@ fn compile_programs(
             cached_program,
         )?;
 
+        if compile_options.check_non_determinism {
+            let (program_two, _) = compile_program(
+                file_manager,
+                parsed_files,
+                workspace,
+                package,
+                compile_options,
+                load_cached_program(package),
+            )?;
+            if fxhash::hash64(&program) != fxhash::hash64(&program_two) {
+                panic!("Non deterministic result compiling {}", package.name);
+            }
+        }
+
         // Choose the target width for the final, backend specific transformation.
         let target_width =
             get_target_width(package.expression_width, compile_options.expression_width);