Yes,
I think you can squeeze out a 10% improvement or so on some combinations of
shortness of String and lowness of repeat count, but at the cost of adding
another branch, risking costs due to profile pollution, decreasing
chance the
method gets inlined etc.. it didn't seem worth the trouble for something
that
already is a much larger performance improvement than so over existing
alternatives.
Thanks!
/Claes
On 2018-03-01 15:15, Jim Laskey wrote:
Roger,
Claes did some performance tests against an earlier implementation where I was
doing just that. We determined that the performance difference was lost in the
noise.
Cheers,
— Jim
On Mar 1, 2018, at 10:11 AM, Roger Riggs <roger.ri...@oracle.com> wrote:
Hi Jim,
Perhaps premature optimization...
Have you done any JMH tests on the expected string content and sizes. The
optimization for single 8-bit characters
is good but for short strings like inserting spaces for tab stops, "
".repeat(5), will suffer the overhead of Arraycopy
for very small/short or non-8-bit strings. Perhaps use a regular loop below some
threshold ( limit < 32).
Maybe the intrinsic already handles this efficiently or it is not important
enough to optimize.
I would not expect to see many very long repeats or long strings being repeated.
And I would leave cache-effect optimizations to the implementation of Arraycopy.
Thanks, Roger
On 2/28/2018 9:13 PM, Paul Sandoz wrote:
Hi Jim,
Looks good. I like the power of 2 copying.
2978 * @throws IllegalArgumentException if the {@code count} is
2979 * negative.
2980 */
2981 public String repeat(int count) {
Missing @since11 on the method.
Like Stuart suggests, turn the explanatory text into an api note, perhaps with
a small code sample.
Thanks,
Paul.
On Feb 28, 2018, at 8:31 AM, Jim Laskey <james.las...@oracle.com> wrote:
Introduction of a new instance method String::repeat to allow an efficient and
concise approach for generating repeated character sequences as strings.
Performance information in JBS.
Thank you.
Cheers,
— Jim
JBS: https://bugs.openjdk.java.net/browse/JDK-8197594
<https://bugs.openjdk.java.net/browse/JDK-8197594>
CSR: https://bugs.openjdk.java.net/browse/JDK-8198296
<http://cr.openjdk.java.net/~jlaskey/8197594/String.html>
Webrev: http://cr.openjdk.java.net/~jlaskey/8197594/webrev-02/index.html
<http://cr.openjdk.java.net/~jlaskey/8197594/webrev-02/index.html>
JavaDoc: http://cr.openjdk.java.net/~jlaskey/8197594/String.html
<http://cr.openjdk.java.net/~jlaskey/8197594/String.html>
