On Fri, 2011-12-02 at 12:44 -0800, Walter Bright wrote: > On 12/2/2011 11:56 AM, Kagamin wrote: > > To think, LLVM devs complain about LLVM IR being so low-level, and it would > > so nice to have something as high-level as Java bytecode, which is so sweet > > for optimizers and jit. > > The people I know who have written professional Jits for the Java bytecode > don't > think it is very amenable to it. > > The Java bytecode is designed to be amenable for writing a simple interpreter > for it. Not a Jit.
Zero address stack machines do not map easily to two or three address register architectures such as Intel's and old IBMs. It's easier on RISC architectures such as SPARC. The JVM interpreter is close to trivial, at least for evaluation. The real problem comes with class loading. Class loaders can be a real pain. -- Russel. ============================================================================= Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.win...@ekiga.net 41 Buckmaster Road m: +44 7770 465 077 xmpp: rus...@russel.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
signature.asc
Description: This is a digitally signed message part
