the potential performance benefits of llvm jit and/or 
post-link optimizer could be enormous. 

phoronix recently did a head to head test of gcc vs llvm and clang 
and recent gcc has pulled ahead of llvm, however the test did not 
specify any of the post-link advantages were in play from the 
llvm-gcc option or from clang.  

LLVM produces an intermediary code before emitting final object 
code.  this IR can be run in a re-linking jit compiler or optimized
 as a static binary for a single architecture.  this IR is specific
 to the compiler hosting the libraries and headers, so it's not as 
portable as java class files but managed configurations are 
attainable, such as apple's X-code demonstrates.

_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to