[aarch64-port-dev ] Changes to Bellsoft/Marvell method of developing intrinsics

[Derek White]

Hi Andrew,

[dropping hotspot-compiler-dev at openjdk.java.net to keep this more aarch64 focused]

> -----Original Message-----
> From: Andrew Haley <aph at redhat.com>
> Sent: Thursday, January 24, 2019 11:52 AM
> Subject: Re: Changes to Bellsoft/Marvell method of developing intrinsics
> 
...
> Thank you for that. I would like to add one thing: before doing anything you
> should openly discuss whether a change should be made at all. We need to
> know the potential gains, the maintenance costs, and what the alternatives
> are.
> 
> For example, it may well be possible to write intrinsics in C++ with a little
> vector code that will perform nearly as well as hand-carved assembly
> language. These will be much cheaper to write, easier to maintain, and more
> reliable, for all the usual reasons to do with high-level languages.
> 
> We may decide that we're not going to do an optimization even though it
> will make some operation 10% faster because it's too risky. It's only worth
> making changes if they really are justified by a significant improvement on
> real-world workloads.

Yes, that makes sense. I've just been talking to Arm on this topic - that we (as the arm porters community) don't currently have a process or habit of talking about what we're:
 - analyzing
 - investigating
 - thinking of doing

JIRAs are used for things that have been decided to be done, and I have that google doc listing JEP features that need investigating or porting, but those don't capture everything.

I think we have pretty good 1-to-1 communication between most of the parties involved, but it's tough for us all to see the whole picture. I don't have solution at hand, but it's something to think about.

 - Derek