speed of simple loops: comparison to other language

[Osvaldo Doederlein]

Hi Tom,

2010/7/1 Tom Rodriguez <tom.rodriguez at oracle.com>

>
> > Well, I guess just need to wait a few more builds until C1 comes to
> Win64?
>
> We'll never ship a stand alone 64 bit client vm.  64 C1 is going to be used
> as part of a tiered system with both client and server in the same binary so
> you'll get good startup and good peak performance.
>

Great news. I've been tracking the tiered compiler effort, indeed I've
noticed some time ago that it's already functional (but doesn't have
expected performance traits, at least in the 32-bit VM last time I checked -
I guess due to heuristics not good enough yet). If the tiered compiler is a
release driver for JDK7, that will be an even-more kick-ass release.

A+
Osvaldo


>
> > Hopefully that will also have UseCompressedOops on by default when heap
> size is sufficiently "small", otherwise 64-bit applets or JAWS apps will
> remain overweight...
>
> C1 doesn't currently support compressed oops but it will eventually in the
> tiered system.
>
> tom
>
> >
> > While we're at this, any idea if Oracle plans to solve the problem of
> double isolation of plugin2 in new browsers? (Latest versions of most
> browsers already isolate native plugins in separate processes, so the extra
> isolation created by plugin2 has become redundant and just adds overhead.)
> >
> > A+
> > Osvaldo
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20100701/05188b0e/attachment.html