Raymond Hettinger wrote: > If there are no objections, I would like to add an opcode for calling > PyList_Append(). This simplifies the generated code for list > comprehensions and reduces the overhead on each pass, improving the > timings by about 35% on [i for i in itertools.repeat(None, 500)]. > > The patch is amazingly brief and clear: > www.python.org/sf/910929 > Ah, nuts! That was going to be part of my thesis! =) I was going to along the lines of "Type-specific Opcodes for Python Based on Static Type Inferencing of Local Variables". At this rate the the first four words won't be needed. =) Joking aside, I am +1 since I was planning on adding that myself as long as my thesis pans out and actually is a gain. Oh, and beware the evil cache monkey and his notorious way of making the eval loop not fit in the CPU cache when a single thing is added to it (or at least I remember a comment along those lines last time people tried fiddling with the eval loop). -Brett
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4