[aarch64-port-dev ] Review, please

[Andrew Haley]

On 14/04/15 19:13, Johnson, E Andrew wrote:

> In running some experiments with different benchmarks, turning on
> and turning off tiered compilation, and turning on "compile all",
> the minimum value to give the best performance was 1200.  I am
> recommending 1500 because that is the default setting for PPC and
> Sparc.

Hmm.  This is the size of compiled code, right?  x86_64 has better
code density than SPARC and PPC, so I'd expect it to be a bit smaller.
AArch64 code density is better than either of those but (as far as I
can see) not as good as x86_64.  I haven't seen any real measurements
for either, though, so all this should not be taken very seriously.

> If there was a separate set of globals for x86_64 (aka amd64), it
> would probably be 1500 there as well.
> 
> I find it interesting that the default inlining threshold value for
> both X86 and Sparc (both architectures supported by Oracle) is
> increased by 1000 when Tiered Compilation is enabled.

So do I.  I don't know why tiered compilation makes a difference.  It
may simply be that C2-compiled code is much denser than C1-compiled,
but that would be a C2 vs. C1 thing rather than a tiered versus
non-tiered.

Andrew.