http://codereview.chromium.org/179059/diff/1/2 File src/heap.cc (right):
http://codereview.chromium.org/179059/diff/1/2#newcode3995 Line 3995: for (int i = 0; i < kCacheSize; i++) { On 2009/09/01 08:40:31, Søren Gjesse wrote: > Would memcpy be faster here? Possibly, but I don't expect this function to be called all that often, and it would look messier. http://codereview.chromium.org/179059/diff/1/4 File src/heap.h (right): http://codereview.chromium.org/179059/diff/1/4#newcode1533 Line 1533: TranscendentalCache* cache = caches_[type]; On 2009/09/01 08:40:31, Søren Gjesse wrote: > The description says "Cache the results of slow math operations on machines that > don't have hardware fpu instructions to execute them.", however it looks as if > the cache is always enabled. Is that intentional? The cache only works in the runtime system. If we have an instruction for the transcendental then we generate it inline and never normally enter the runtime system. http://codereview.chromium.org/179059 --~--~---------~--~----~------------~-------~--~----~ v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev -~----------~----~----~----~------~----~------~--~---
