Comment #7 on issue 2097 by [email protected]: sha1 compiled with Emscripten runs 10X slower than with SpiderMonkey
http://code.google.com/p/v8/issues/detail?id=2097
@#6: AFAIK single-statement jsperf.com tests effectively disable the optimizing compiler. Micro-benchmarks are not indicative of real-world performance anyway; please try to not base your observations on ultra-small micro-benchmarks.
That said, we know that reading 0x80000000 (or similar values) from a Uint32Array causes us to deopt. See comment #5.
-- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
