Re: Indentation and gc

[Ergus <spacibba@aol.com>]

On Sat, Mar 11, 2023 at 10:54:24AM +0000, Ihor Radchenko wrote:
>Eli Zaretskii <eliz@gnu.org> writes:
>
>>> ;; Make gc pauses faster by decreasing the threshold again (from the increased initial).
>>> (setq gc-cons-threshold (* 20 1024 1024))
>>> ;; original value: 800 000
>>>
>>> ;; speed up reading from external processes
>>> (setq read-process-output-max (* 1024 1024)) ;; 1mb
>>
>> This can only be done around specific portions of code known in
>> advance to be long and GC-intensive.  I don't think this kind of
>> technique can be used in the situation described by the OP.
>
>May it be done when loading init.el and early-init.el?
>Init files are commonly known to be resource-intensive. They also tend
>to trigger GC more because the heap size is not yet very large and thus
>`gc-cons-percentage' does not yet take precedence over `gc-cons-threshold'.
>
>As for "known in advance", may Emacs keep track of how many GCs are
>triggered by user commands and then adjust GC dynamically?
>

Hi Ihor:

I have this in my early-init since a few years now.


```
(defconst my/file-name-handler-alist file-name-handler-alist)
(defconst my/gc-cons-threshold (* 2 gc-cons-threshold))

(setq-default file-name-handler-alist nil
               gc-cons-threshold most-positive-fixnum   ;; Defer Garbage collection
               gc-cons-percentage 1.0
               message-log-max 16384)

(add-hook 'window-setup-hook
           (lambda ()
             (setq file-name-handler-alist my/file-name-handler-alist
                   gc-cons-threshold my/gc-cons-threshold
                   gc-cons-percentage 0.1)
             (let ((curtime (current-time)))
               (message "Times: init:%.06f total:%.06f gc-done:%d"
                        (float-time (time-subtract after-init-time before-init-time))
                        (float-time (time-subtract curtime before-init-time))
                        gcs-done)))
           90)
```

This reduced my load time by 50% on GNU/Linux (a bit more indeed)... on
MS-Windows I am still above the 7 seconds with exactly same config (so
more than 14x slower than GNU/Linux... but I guess the problem is maybe
outside GNU Linux... antivirus virus and MS process creation slowness.)


>-- 
>Ihor Radchenko // yantar92,
>Org mode contributor,
>Learn more about Org mode at <https://orgmode.org/>.
>Support Org development at <https://liberapay.com/org-mode>,
>or support my work at <https://liberapay.com/yantar92>
>