On Thu, Mar 13, 2014 at 2:11 PM, Rodolph Perfetta < [email protected]> wrote:
> Also, some of the memcpy uses in the CL are completely optimised away. >>> In those cases, there's no good reason to use a reinterpret_cast so they >>> should probably stay. >>> >> >> For something as performance-sensitive as a simulator, I think some hacks >> are worth the trouble. The CL shaves away more than 1min of developer time >> per quickcheck run and more than half an hour on the bots. >> > > I tried your patch on my machine and see no speed up. We already > established with Jochen that using gcc 4.8 or clang gives a healthy speed > up compared to gcc 4.6. Which compiler are you using? Presumably the newer > compiler are better at optimising away the memcpy. > Discussing compiler versions is a moot point IMHO: Forcing *all* developers and bots to upgrade just for some hypothetical problem is not nice. For the record: I use the standard gcc on our goobuntu machines, which 4.6.3. I agree, the Instr/Instruction distinction shoudl disappear, it is in a64 > because it was in ARM (not a good reason I know) > I'm currently trying to remove this, let's see if the change is small enough... -- -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
