03-Mar-2013 12:03, Russel Winder пишет:
On Sat, 2013-03-02 at 12:52 -0800, Walter Bright wrote:
On 3/2/2013 12:08 PM, H. S. Teoh wrote:
On Sat, Mar 02, 2013 at 12:02:08PM -0800, Walter Bright wrote:
On 3/2/2013 11:42 AM, Russel Winder wrote:
On Sat, 2013-03-02 at 11:30 -0800, Walter Bright wrote:
On 3/2/2013 7:43 AM, Russel Winder wrote:
For writing interpreters, RPython spanks C.
What's RPython doing that makes it faster?
Allowing PyPy to have a good JIT compiler.
I don't understand. Does that JIT generate faster code than a C
compiler would generate?
I don't know, but my wild guess is that a JIT optimizes the *right*
hotspots based on real-time performance measurements, whereas a lot of C
programmers are obsessed with optimizing what they *think* are the
hotspots, but which really aren't.
I meant what the C *compiler* generates.
Yes because the C/C++/D/etc. compilers are attempting to predict the
control flow of the program in execution and optimize all cases for all
possibilities. JIT's are just focussing on the runtime bottlenecks with
the actual data as being used. This allows for more focussed code
generation in the actual context. I would suspect that in many cases the
generated code is effectively the same but JITs can often do unexpected
and faster codes because they have more data to optimize with and less
optimization to do.
Only one thing missing here is that JITs typically has much less time to
do all and any of this.
To say more would require actual comparative data and I suspect no-one
on this list will do that.
Indeed unlikely.
--
Dmitry Olshansky
