If we are talking of intrinsics in general, the lower level factoring has happened to some extent via the vmmagics, for api's the only way I can think of is pick out frequently used operations from perf runs etc. as Mikhail proposes. Some of the helper fastpaths could themselves be candidates in jitrino.opt, several JIT's offer an intrinsic for fast alocation, for example. We could leave the fastpaths in place for down level and new jit compilers.
On 12/18/06, Xiao-Feng Li <[EMAIL PROTECTED]> wrote:
>I didn't mean to be against the idea of instrinc substitution, just >feel (intruitively, not necessarily reasonably) there are some >approach to make the work systematic, instead of identifying the >method/class one by one. For example, to categorize the intrinsics >domains into some common subroutines. >Well just two cents, don't know if there is better approach than what >Mikhail proposed. >Thanks, >xiaofeng
