From: Josh Bleecher Snyder <josharian@gmail.com>
Date: Sat, 25 Jun 2016 22:11:25 +0000 (-0700)
Subject: testing: respect benchtime on very fast benchmarks
X-Git-Tag: go1.8beta1~1896
X-Git-Url: http://www.git.cypherpunks.su/?a=commitdiff_plain;h=31ad583ab24263f9dbcb5cbcce849eed64e74040;p=gostls13.git

testing: respect benchtime on very fast benchmarks

When ns/op dropped below 1, the old code
ignored benchtime and reverted to 1s.

Change-Id: I59752cef88d8d73bfd5b085f5400ae657f78504e
Reviewed-on: https://go-review.googlesource.com/26664
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Marcel van Lohuizen <mpvl@golang.org>
---

diff --git a/src/testing/benchmark.go b/src/testing/benchmark.go
index 5d58b85e78..4d45130516 100644
--- a/src/testing/benchmark.go
+++ b/src/testing/benchmark.go
@@ -263,10 +263,9 @@ func (b *B) launch() {
 	for n := 1; !b.failed && b.duration < d && n < 1e9; {
 		last := n
 		// Predict required iterations.
-		if b.nsPerOp() == 0 {
-			n = 1e9
-		} else {
-			n = int(d.Nanoseconds() / b.nsPerOp())
+		n = int(d.Nanoseconds())
+		if nsop := b.nsPerOp(); nsop != 0 {
+			n /= int(nsop)
 		}
 		// Run more iterations than we think we'll need (1.2x).
 		// Don't grow too fast in case we had timing errors previously.