[vector] Min/Max support cleanup for x86

[Viswanathan, Sandhya]

Hi John,

Thanks a lot for your inputs. Please find below the webrev with the change:
http://cr.openjdk.java.net/~sviswanathan/vectorIntrinsics/MinMaxCleanup/webrev.01/

Best Regards,
Sandhya

From: John Rose <john.r.rose at oracle.com>
Sent: Friday, February 07, 2020 2:10 PM
To: Viswanathan, Sandhya <sandhya.viswanathan at intel.com>
Cc: panama-dev at openjdk.java.net
Subject: Re: [vector] Min/Max support cleanup for x86

On Feb 3, 2020, at 1:52 PM, Viswanathan, Sandhya <sandhya.viswanathan at intel.com<mailto:sandhya.viswanathan at intel.com>> wrote:


Please find below a webrev which attempts to clean up the Min/Max support in x86.ad file.

http://cr.openjdk.java.net/~sviswanathan/vectorIntrinsics/MinMaxCleanup/webrev.00
Very nice refactoring.  This bit is the best news:

616 lines changed: 69 ins; 481 del; 66 mod; 22624 unchg

Removing lines of code is a fine way to prevent bugs.

I like the API shape of MacroAssembler::vpminmax,
although it seems a single step from separate min and max
and several steps from a macro that handles a larger number
of ops.  I wonder where the sweet spot is.  (No action required
here.)

The macro assembler seems to say both “min_max” and “minimal”.
I suggest picking just convention and sticking with it.
FWIW I think “minmax” is easier to read

(That may be because I’m dreaming here of replacing “minmax” by
“lanewise” with a richer operation parameter.  :-)

— John