feat(phase21.5): MirBuilder optimization prep + crate EXE infrastructure

Phase 21.5 optimization readiness - C-level performance target:
- MirBuilder: JsonFrag purify toggle (HAKO_MIR_BUILDER_JSONFRAG_PURIFY=1)
- Normalizer: extended f64 canonicalization + dedupe improvements
- loop_opts_adapter: JsonFrag path refinement for crate EXE compatibility

Infrastructure improvements:
- provider_registry: add diagnostics + ring-1 providers (array/console/map/path)
- mir_interpreter: add normalization/purify feature gates
- tools/selfhost_exe_stageb.sh: new end-to-end Stage-B→crate EXE pipeline
- tools/perf/microbench.sh: performance measurement tooling

Smoke tests (phase2100):
- Extend timeout 15s→120s for heavy crate EXE builds
- Add stageb_loop_jsonfrag_crate_exe_canary_vm.sh (target test)
- Add s3_backend_selector_crate_exe_vm_parity_return42_canary_vm.sh

Documentation:
- ENV_VARS.md: add Phase 21.5 optimization toggles
- README updates: clarify crate backend strategy
- phase215-optimization.md: new optimization roadmap

This commit sets the stage for Phase 21.5 critical optimization:
achieving C-level performance to decide hakorune's future viability.

src/backend/mir_interpreter/exec.rs

@@ -257,6 +257,11 @@ impl MirInterpreter {
             if Self::trace_enabled() {
                 eprintln!("[vm-trace] inst bb={:?} {:?}", block.id, inst);
             }
+            // Dev counters: count non-phi instructions and compares
+            self.inst_count = self.inst_count.saturating_add(1);
+            if let MirInstruction::Compare { .. } = inst {
+                self.compare_count = self.compare_count.saturating_add(1);
+            }
             self.execute_instruction(inst)?;
         }
         Ok(())
@@ -281,6 +286,8 @@ impl MirInterpreter {
                 then_bb,
                 else_bb,
             }) => {
+                // Dev counter: count branch terminators actually evaluated
+                self.branch_count = self.branch_count.saturating_add(1);
                 let cond = self.reg_load(*condition)?;
                 let branch = to_bool_vm(&cond).map_err(VMError::TypeError)?;
                 let target = if branch { *then_bb } else { *else_bb };

src/backend/mir_interpreter/mod.rs

@@ -37,6 +37,10 @@ pub struct MirInterpreter {
     pub(super) static_boxes: HashMap<String, crate::instance_v2::InstanceBox>,
     // Static box declarations (metadata for creating instances)
     pub(super) static_box_decls: HashMap<String, crate::core::model::BoxDeclaration>,
+    // Lightweight dev counters (opt-in print via NYASH_VM_STATS=1)
+    pub(super) inst_count: u64,
+    pub(super) branch_count: u64,
+    pub(super) compare_count: u64,
 }
 
 impl MirInterpreter {
@@ -51,9 +55,17 @@ impl MirInterpreter {
             last_inst: None,
             static_boxes: HashMap::new(),
             static_box_decls: HashMap::new(),
+            inst_count: 0,
+            branch_count: 0,
+            compare_count: 0,
         }
     }
 
+    /// Return (inst_count, branch_count, compare_count)
+    pub fn stats_counters(&self) -> (u64, u64, u64) {
+        (self.inst_count, self.branch_count, self.compare_count)
+    }
+
     /// Register static box declarations (called from vm.rs during setup)
     pub fn register_static_box_decl(&mut self, name: String, decl: crate::core::model::BoxDeclaration) {
         self.static_box_decls.insert(name, decl);