rentzsch
Ken Ferry
@rentzsch are they disallowing ISA-specific intrinsics? They're definitely representable in bitcode…
I think bitcode is a huge win. About only thing I’ll miss is cool ISA-specific instructions. Now reliant on OS vendor providing access.
-
-
-
@kongtomorrow I was assuming so. I probably shouldn’t. -
@rentzsch I need to watch the sessions… biggest wins I could think of were (1) AOT LTO with snapshots of system libraries, (2) microarchs. -
@rentzsch you can inline code from the frameworks if you can guarantee a (partial) recompile every time the OS frameworks change. - View other replies
-
@kongtomorrow don’t know if that jives with memory pressure perf costs. Still, it’s very interesting and flexible. - Show more
-
-
-
@rentzsch no more assembly? -
@rudyrichter@rentzsch That’s the other one that troubles me. /cc@landonfuller
-
-
-
@rentzsch Bitcode isn't actually ISA-independent, though (and if it was, we'd need a much more high-level VM definitions to work with).
-
-
@rentzsch huge flexibility for Apple to innovate in moving optimised routines to silicon in Sx chips, eg. Crypto, AV, etc optimised inst
Rudy Richter
Gwynne Raskind
Landon Fuller
Craig Morgan
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.