Re: VIRT_ADDR_VARIES

[Paul Eggert <eggert@cs.ucla.edu>]

On 11/07/11 00:38, Andreas Schwab wrote:
> Paul Eggert <eggert@cs.ucla.edu> writes:
>> I expect the main reason is performance.
> 
> How much of it matters?

Compiling without VIRT_ADDR_VARIES speeds up aset by 44% on my host
(Fedora 15 x86-64, GCC 4.6.2, AMD Phenom II X4 910e).
Benchmark code is shown below; I ran (benchmark 100000000)
five times on the byte-compiled version of this code,
and took the median.

>> for Emacs it shouldn't matter whether PURE_P checks
>> accurately or loosely -- either way Emacs should operate correctly.
> 
> Which it doesn't.

Sorry, I don't know the context.  What's the bug here?
If Emacs doesn't define VIRT_ADDR_VARIES when it should,
then that can cause bugs -- is that the issue?

Here's the benchmark code.

(defun benchmark-with-aset (n)
  (let ((start (float-time (get-internal-run-time)))
	(v (make-vector 1 0))
	(i 0))
    (while (< i n)
      (aset v 0 1)
      (setq i (1+ i)))
    (- (float-time (get-internal-run-time)) start)))

(defun benchmark-without-aset (n)
  (let ((start (float-time (get-internal-run-time)))
	(v (make-vector 1 0))
	(i 0))
    (while (< i n)
      (setq i (1+ i)))
    (- (float-time (get-internal-run-time)) start)))

(defun benchmark (n)
  (- (benchmark-with-aset n)
     (benchmark-without-aset n)))