Hi Vivek,
459 public static int mismatch(float[] a, int aFromIndex,
460 float[] b, int bFromIndex,
461 int length) {
462 int i = 0;
463 if (length > 1) {
464 if (a[aFromIndex] != b[bFromIndex]) {
465 i = 0;
466 }
467 if (a[aFromIndex] == b[bFromIndex]) {
Sorry, i don’t think i was clear. You can reduce down to a single equality
check, so Lines #464-466 are redundant. This does result in taking the slow
path if both values are NaN, which i think was the same for your previous
patch. If that is important we can mitigate that by performing the negation of
the same checks in the slow loop.
WDYT?
Paul.
> On Jun 20, 2018, at 5:05 PM, Deshpande, Vivek R <[email protected]>
> wrote:
>
> Hi Paul
>
> I have made the change you have suggested.
> Please find the updated webrev at this location:
> http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.02/
> <http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.02/>
>
> Regards,
> Vivek
> From: Paul Sandoz [mailto:[email protected]
> <mailto:[email protected]>]
> Sent: Wednesday, June 20, 2018 2:30 PM
> To: Deshpande, Vivek R <[email protected]
> <mailto:[email protected]>>
> Cc: [email protected] <mailto:[email protected]>; David Holmes
> <[email protected] <mailto:[email protected]>>;
> [email protected] <mailto:[email protected]>;
> Viswanathan, Sandhya <[email protected]
> <mailto:[email protected]>>
> Subject: Re: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> 459 public static int mismatch(float[] a, int aFromIndex,
> 460 float[] b, int bFromIndex,
> 461 int length) {
> 462 int i = 0;
> 463 if (length > 1) {
> 464 if (a[aFromIndex] != b[bFromIndex]) {
> 465 i = 0;
> 466 } else {
>
> You fold the if/else to;
>
> if (a[aFromIndex] == b[bFromIndex]) {
> …
> }
>
> same applies in other cases in both source files.
>
>
>
> On Jun 20, 2018, at 1:21 PM, Deshpande, Vivek R <[email protected]
> <mailto:[email protected]>> wrote:
>
> Hi All
>
> I have updated the webrev with all the suggestions, which passes
> ArraysEqCmpTest.java.
> Earlier webrev failed the same test.
>
> Thanks for confirming.
>
>
> This webrev also modifies the BufferMismatch.java in similar way.
>
> Great!
>
> Paul.
>
>
> Please take a look and let me know what you think.
>
> Updated webrev is here:
> http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.01/
> <http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.01/>
> I have also modified the bug with updated webrev.
> Regards,
> Vivek
>
> From: Deshpande, Vivek R
> Sent: Tuesday, June 19, 2018 10:57 AM
> To: Paul Sandoz <[email protected] <mailto:[email protected]>>;
> [email protected] <mailto:[email protected]>
> Cc: David Holmes <[email protected] <mailto:[email protected]>>;
> [email protected] <mailto:[email protected]>;
> Viswanathan, Sandhya <[email protected]
> <mailto:[email protected]>>
> Subject: RE: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> Hi Roger
>
> I will also test with zero length arrays and let you know.
> Thanks for the input.
>
> Regards,
> Vivek
>
> From: Deshpande, Vivek R
> Sent: Tuesday, June 19, 2018 10:17 AM
> To: 'Paul Sandoz' <[email protected] <mailto:[email protected]>>
> Cc: David Holmes <[email protected] <mailto:[email protected]>>;
> [email protected] <mailto:[email protected]>;
> Viswanathan, Sandhya <[email protected]
> <mailto:[email protected]>>
> Subject: RE: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> Thanks Paul for quick review. I will work on the things you have mentioned
> and get back soon.
> I will also test this with test/jdk/java/util/Arrays/ArraysEqCmpTest.java.
>
> Regards,
> Vivek
>
> From: Paul Sandoz [mailto:[email protected]
> <mailto:[email protected]>]
> Sent: Tuesday, June 19, 2018 9:55 AM
> To: Deshpande, Vivek R <[email protected]
> <mailto:[email protected]>>
> Cc: David Holmes <[email protected] <mailto:[email protected]>>;
> [email protected] <mailto:[email protected]>;
> Viswanathan, Sandhya <[email protected]
> <mailto:[email protected]>>
> Subject: Re: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> Hi Vivek,
>
> Thanks for investigating this.
>
> 164 public static int mismatch(boolean[] a,
> 165 boolean[] b,
> 166 int length) {
> 167 int i = 0;
> 168 if (a[i] != b[i])
> 169 return i;
> You might as well replace the use if i with 0 and i think you can move this
> to be performed when the length is greater than the threshold, that way you
> don’t impact small arrays below the threshold.
>
> 186 public static int mismatch(boolean[] a, int aFromIndex,
> 187 boolean[] b, int bFromIndex,
> 188 int length) {
> 189 int i = 0;
> 190 if (a[i] != b[i])
> 191 return i;
> This is incorrect. You need to use aFromIndex and bFromIndex.
>
> Do you run the test test/jdk/java/util/Arrays/ArraysEqCmpTest.java? If that
> passed then we need to strengthen for the case of a mismatch on the first
> relative element in each array.
>
> Paul.
>
>
> On Jun 19, 2018, at 9:36 AM, Deshpande, Vivek R <[email protected]
> <mailto:[email protected]>> wrote:
>
> Thanks David.
> Sending it to core-libs-dev.
>
> I would like to contribute a patch which improves the array comparison when
> there is a mismatch for the first element.
> This avoids call to vectorizedMismatch method and gives ~80x speed up.
> Could you please review and sponsor the patch.
> Link to bug:
> https://bugs.openjdk.java.net/browse/JDK-8205194
> <https://bugs.openjdk.java.net/browse/JDK-8205194>
> webrev:
> http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.00/
> <http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.00/>
>
> Regards,
> Vivek
>
> -----Original Message-----
> From: David Holmes [mailto:[email protected]
> <mailto:[email protected]>]
> Sent: Monday, June 18, 2018 10:32 PM
> To: Deshpande, Vivek R <[email protected]
> <mailto:[email protected]>>; [email protected]
> <mailto:[email protected]>
> Subject: Re: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> Hi Vivek,
>
> Reviews should take place on the appropriate mailing list for the code being
> changed, not on the jdk-dev list. Please takes this to core-libs-dev.
>
> Thanks,
> David
>
> On 19/06/2018 9:52 AM, Deshpande, Vivek R wrote:
>
> Hi All
>
> Forgot to add the links:
> https://bugs.openjdk.java.net/browse/JDK-8205194
> <https://bugs.openjdk.java.net/browse/JDK-8205194>
> webrev:
> http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.0
> <http://cr.openjdk.java.net/~vdeshpande/vectorizedMismatch_jdk/webrev.0>
> 0/
>
> Regards.
> Vivek
>
> From: Deshpande, Vivek R
> Sent: Monday, June 18, 2018 4:50 PM
> To: '[email protected] <mailto:[email protected]>'
> <[email protected] <mailto:[email protected]>>
> Cc: 'Paul Sandoz' <[email protected] <mailto:[email protected]>>;
> Viswanathan, Sandhya
> <[email protected] <mailto:[email protected]>>
> Subject: RFR(S): 8205194: Improve the Array Comparison when there is a
> mismatch at first element.
>
> Hi All
>
> I would like to contribute a patch which improves the array comparison when
> there is a mismatch for the first element.
> This avoids call to vectorizedMismatch method and gives ~80x speed up.
> Please review and sponsor the patch.
>
> Regards,
> Vivek
>
