ci: also run benchmarks as tests

For `nextest`'s parsing of `cargo criterion --list --format terse` to
work, benchmarks must not print to standard out.

.github/workflows/coverage.yml

@@ -31,7 +31,11 @@ jobs:
         run: cargo run --bin constraint-evaluation-generator
 
       - name: Collect coverage data
-        run: cargo llvm-cov nextest --lcov --output-path lcov.info --features integration-tests
+        run: >
+          cargo llvm-cov nextest
+          --all-targets
+          --features integration-tests
+          --lcov --output-path lcov.info
 
       - name: Upload coverage to coveralls.io
         uses: coverallsapp/github-action@v2

.github/workflows/main.yml

@@ -43,10 +43,7 @@ jobs:
         run: cargo clippy --all-targets -- -D warnings
 
       - name: Run tests
-        run: cargo nextest run --no-fail-fast --tests
-
-      - name: Run examples
-        run: cargo nextest run --no-fail-fast --examples
+        run: cargo nextest run --no-fail-fast --all-targets
 
         # doctests are special [^1] but this step does not incur a performance penalty [^2]
         #

triton-vm/benches/proof_size.rs

@@ -214,18 +214,18 @@ fn print_proof_size_breakdown(program_name: &str, proof: &Proof) {
     let proof_size_breakdown = break_down_proof_size(proof);
     let proof_size_breakdown = sort_hash_map_by_value_descending(proof_size_breakdown);
 
-    println!();
-    println!("Proof size breakdown for {program_name}:");
-    println!(
+    eprintln!();
+    eprintln!("Proof size breakdown for {program_name}:");
+    eprintln!(
         "| {:<30} | {:>10} | {:>6} |",
         "Category", "Size [bfe]", "[%]"
     );
-    println!("|:{:-<30}-|-{:->10}:|-{:->6}:|", "", "", "");
+    eprintln!("|:{:-<30}-|-{:->10}:|-{:->6}:|", "", "", "");
     for (category, size) in proof_size_breakdown {
         let relative_size = (size as f64) / (total_proof_size as f64) * 100.0;
-        println!("| {category:<30} | {size:>10} | {relative_size:>6.2} |");
+        eprintln!("| {category:<30} | {size:>10} | {relative_size:>6.2} |");
     }
-    println!();
+    eprintln!();
 }
 
 /// Create `num_iterations` many proofs for the program with the supplied source code and

triton-vm/benches/prove_fib.rs

@@ -17,8 +17,7 @@ fn prove_fib(criterion: &mut Criterion) {
     fib_benchmark_group(criterion, &claim, &aet);
 
     let report = prover_timing_report(&claim, &aet);
-    println!("Writing report ...");
-    println!("{report}");
+    eprintln!("{report}");
 }
 
 fn fib_benchmark_group(criterion: &mut Criterion, claim: &Claim, aet: &AlgebraicExecutionTrace) {

triton-vm/benches/prove_halt.rs

@@ -30,15 +30,14 @@ fn prove_halt(criterion: &mut Criterion) {
     });
     group.finish();
 
-    println!("Writing report ...");
     let padded_height = proof.padded_height().unwrap();
     let fri = stark.derive_fri(padded_height).unwrap();
     let report = profiler
         .report()
         .with_cycle_count(aet.processor_trace.nrows())
         .with_padded_height(padded_height)
         .with_fri_domain_len(fri.domain.length);
-    println!("{report}");
+    eprintln!("{report}");
 }
 
 criterion_group! {

triton-vm/benches/verify_halt.rs

@@ -41,7 +41,7 @@ fn verify_halt(criterion: &mut Criterion) {
     });
     group.finish();
 
-    println!("{report}");
+    eprintln!("{report}");
 }
 
 criterion_group! {

triton-vm/src/profiler.rs

@@ -103,19 +103,6 @@ impl TritonProfiler {
         );
 
         let mut report: Vec<TaskReport> = vec![];
-        let total_tracked_time = (self.total_time.as_nanos()
-            - self
-                .profile
-                .iter()
-                .filter(|t| t.task_type == TaskType::AnyOtherIteration)
-                .map(|t| t.time.as_nanos())
-                .sum::<u128>()) as f64
-            / 1_000_000_000f64;
-        println!(
-            "total time: {}s and tracked: {}s",
-            self.total_time.as_secs_f64(),
-            total_tracked_time,
-        );
 
         // collect all categories and their total times
         // todo: this can count the same category multiple times if it's nested
@@ -189,6 +176,7 @@ impl TritonProfiler {
             tasks: report,
             name: self.name.clone(),
             total_time: self.total_time,
+            tracked_time: self.tracked_time(),
             category_times,
             cycle_count: None,
             padded_height: None,
@@ -310,6 +298,14 @@ impl TritonProfiler {
             self.plain_stop();
         }
     }
+
+    pub fn tracked_time(&self) -> Duration {
+        let is_other_iteration = |t: &&Task| t.task_type == TaskType::AnyOtherIteration;
+        let all_tasks = self.profile.iter();
+        let other_iterations = all_tasks.filter(is_other_iteration);
+        let untracked_time = other_iterations.map(|t| t.time).sum();
+        self.total_time.saturating_sub(untracked_time)
+    }
 }
 
 impl Profiler for TritonProfiler {
@@ -391,6 +387,7 @@ pub struct Report {
     name: String,
     tasks: Vec<TaskReport>,
     total_time: Duration,
+    tracked_time: Duration,
     category_times: HashMap<String, Duration>,
     cycle_count: Option<usize>,
     padded_height: Option<usize>,
@@ -444,6 +441,9 @@ impl Display for Report {
         let title = format!("{title:<max_width$}");
 
         let total_time_string = Report::display_time_aligned(self.total_time).bold();
+        let tracked_time = Report::display_time_aligned(self.tracked_time);
+        writeln!(f, "tracked time: {tracked_time}s")?;
+
         let share_string = "Share".to_string().bold();
         let category_string = match self.category_times.is_empty() {
             true => ColoredString::default(),