tests: latency_measure: reduce the chance of cycles underflow

Sometimes there's an unusually large cycles for tests that are
known to complete with just a few cycles. Upon some testing,
I found that it was because the overhead cycles was larger
than the cycles taken by tests, causing the cycles to
underflow.

To workaround that, make sure that the overhead measurement
thread runs uninterrupted, repeat the measurement for a few
times, and take the minimum value.

Signed-off-by: Yong Cong Sin <[email protected]>
Signed-off-by: Yong Cong Sin <[email protected]>

Author: ycsin

tests/benchmarks/latency_measure/src/timing_sc.c

@@ -19,6 +19,8 @@ BENCH_BMEM uint64_t timestamp_overhead;
 BENCH_BMEM uint64_t user_timestamp_overhead;
 #endif
 
+#define OVERHEAD_CALC_ITER 10
+
 timing_t z_impl_timing_timestamp_get(void)
 {
 	return timing_counter_get();
@@ -37,17 +39,23 @@ static void start_thread_entry(void *p1, void *p2, void *p3)
 	uint32_t  num_iterations = (uint32_t)(uintptr_t)p1;
 	timing_t  start;
 	timing_t  finish;
+	uint64_t min_cycles = UINT64_MAX;
 
 	ARG_UNUSED(p2);
 	ARG_UNUSED(p3);
 
-	start = timing_timestamp_get();
-	for (uint32_t i = 0; i < num_iterations; i++) {
-		timing_timestamp_get();
+	/* Repeat the overhead measurements for a few times to obtain the minimum overhead */
+	for (int n = 0; n < OVERHEAD_CALC_ITER; n++) {
+		start = timing_timestamp_get();
+		for (uint32_t i = 0; i < num_iterations; i++) {
+			timing_timestamp_get();
+		}
+		finish = timing_timestamp_get();
+
+		min_cycles = MIN(min_cycles, timing_cycles_get(&start, &finish));
 	}
-	finish = timing_timestamp_get();
 
-	timestamp.cycles = timing_cycles_get(&start, &finish);
+	timestamp.cycles = min_cycles;
 }
 
 void timestamp_overhead_init(uint32_t num_iterations)