[Assorted-commits] SF.net SVN: assorted: [269] sandbox/trunk/src/java/TimeBench.java

[<yan...@us...>]

Revision: 269
          http://assorted.svn.sourceforge.net/assorted/?rev=269&view=rev
Author:   yangzhang
Date:     2008-01-20 23:22:45 -0800 (Sun, 20 Jan 2008)

Log Message:
-----------
added timebench

Added Paths:
-----------
    sandbox/trunk/src/java/TimeBench.java

Added: sandbox/trunk/src/java/TimeBench.java
===================================================================
--- sandbox/trunk/src/java/TimeBench.java	                        (rev 0)
+++ sandbox/trunk/src/java/TimeBench.java	2008-01-21 07:22:45 UTC (rev 269)
@@ -0,0 +1,50 @@
+public class TimeBench {
+
+	public static void main(String[] args) {
+		// On my dual-core Intel machine, millis is *faster* than nano, despite the
+		// fact that FreeNode folk told me nano was more accurate (perhaps it is
+		// once you get to that granularity, but the cost of invocation is twice as
+		// high):
+		//
+		// $ for i in {1..10}; do java TimeBench; done
+		// 213
+		// 416283169
+		// 210
+		// 427611823
+		// 211
+		// 425782976
+		// 211
+		// 416913721
+		// 213
+		// 419326076
+		// 209
+		// 415857948
+		// 209
+		// 415659440
+		// 210
+		// 410968289
+		// 212
+		// 421450138
+		// 216
+		// 417043555
+		//
+		// Furthermore, I verified using strace, that none of these produces
+		// syscalls.
+
+		{
+			long t = System.currentTimeMillis(), t0 = System.currentTimeMillis();
+			for (int i = 0; i < 1000000; i++)
+				t = System.currentTimeMillis();
+			System.out.println(t - t0);
+		}
+
+		{
+			long t = System.nanoTime(), t0 = System.nanoTime();
+			for (int i = 0; i < 1000000; i++)
+				t = System.nanoTime();
+			System.out.println(t - t0);
+		}
+
+	}
+
+}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.