Changes between Version 4 and Version 5 of NonUnaryCore

Timestamp:

Nov 18, 2016 2:58:38 PM (3 years ago)

Author:

nomeata

Comment:

--

NonUnaryCore

@@ -73 +73 @@
 10 loops, best of 3: 5.49 sec per loop
 }}}
-and found that the change improves runtime, in this exteme micro-benchmark, in this iteation, by 10%. Which is not great given the form of the benchmark, but at least a number.
+and found that the change improves runtime, in this extreme micro-benchmark, in this iteration, by 10% (and less when running it again ☹). Which is not great given the form of the benchmark, but at least a number.
 
 Also, the real code would still have to check `%rax` for whether it is an evaluated function there, or a thunk that first needs entering.