To read all comments associated with this story, please click here.
Certainly a JIT approach can be useful if you've got a machine around and you want to run a non-native app. But it's never going to be a first-choice option, it's more like an option you would choose if you're stuck with an incompatible processor or an incompatible app. Native execution is always going to be better, unless you have an app for such an old architecture that, even with the JIT performance penalty, your new machine can still run it faster.
And that might even be the case for your old Solaris/SPARC apps. We've got plenty of old, underpowered SPARC hardware around here.
One place this will definitely be useful is for sales demos. This technology will allow SPARC-based products to be demonstrated using fairly standard laptops. Current solutions are more problematic; porting to Solaris x86 (which may be complicated by third party software availability), transporting a small SPARC server (even a SPARCbook may be difficult if the salesperson also needs a separate laptop), or using a remote server (prospective customer's firewalls can create substantial problems).




Member since:
2005-12-15
This is great progress, but I think the performance hit is a bit too much to tempt me at this stage. There are a few areas I could definately use this, so I'll watch this over time with much interest.
I'll even give it a shot to investigate compatibility.