bug#5099: 23.1; Performance problems in Emacs 23?

bug#5099: 23.1; Performance problems in Emacs 23?

[Uday S Reddy]

Hello, I am a maintainer of the VM mail reader package.

Back in October 2008, one of our users who was trying out a pre-release of Eamcs 23 reported that "VM continually freezes my emacs". "At first I thought this was because it was collecting mail, so I set `vm-auto-get-new-mail' to nil, but still VM will burn my CPU and freeze my emacs for 5-20 seconds at a time."

My response to this was that his mail folders must have gotten big, and so the garbage collection or auto-save pauses must have become significant.

However, having tried Emacs 23 for an hour, I have noticed very much
the same phenomenon. My main mail file is roughly 100MB large.  It
takes Emacs 22 only about 5 seconds to save it, and the auto-save
pause is probably half that long.  (I use a reasonably fast Solid
State Drive.) The CPU utilisation is an A-shaped curve, implying that
there is nothing much for the CPU to do. In contrast, Emacs 23 takes
about more than 20 seconds to save the same mail file and the CPU utilisation is a pretty intensive burn for most of that time (20% on one processor and 60% on a second processor).  The auto-save pauses are about 10-15 seconds long.

Now, VM mail files are raw-text.  Saving them should be an I/O-bound process.  It is not clear why Emacs 23 should be burning CPU cycles to save raw-text files and taking 3-5 times as long as Emacs 22.

Cheers,
Uday Reddy





In GNU Emacs 23.1.1 (i386-mingw-nt5.1.2600)
 of 2009-07-30 on SOFT-MJASON
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (4.4)'

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: ENU
  value of $XMODIFIERS: nil
  locale-coding-system: cp1252
  default-enable-multibyte-characters: t

Major mode: Info

Minor modes in effect:
  savehist-mode: t
  tooltip-mode: t
  mouse-wheel-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-mode: t
  global-auto-composition-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  line-number-mode: t
  transient-mark-mode: t

Recent input:
<help-echo> <down-mouse-1> <mouse-1> C-h i u u u u 
m e m SPC <return> C-s b u g C-s C-a m <return> C-x 
C-f \ <backspace> <backspace> <backspace> <backspace> 
<backspace> <backspace> e t c / p r o SPC <return> 
<next> <next> <next> <next> <next> <next> <next> <prior> 
<next> <next> C-s s f <backspace> l o w C-s C-s C-s 
C-s C-s C-s C-s C-s C-s C-s C-s C-s C-s C-a C-x , C-s 
2 2 C-s C-s C-s C-s C-r C-r C-s C-s C-s C-s C-s C-a 
<down-mouse-1> <mouse-1> <down-mouse-1> <mouse-1> <next> 
<next> <next> <next> <next> <next> <next> <next> <prior> 
<next> <next> <next> <next> <wheel-down> <double-wheel-down> 
<triple-wheel-down> <triple-wheel-down> <down-mouse-1> 
<mouse-1> <down-mouse-1> <mouse-1> <M-left> <M-left> 
<M-left> <M-left> <M-left> <M-left> <M-left> <M-left> 
<M-left> <M-left> <down-mouse-1> <mouse-1> C-x b <return> 
m m <backspace> u n d e r SPC SPC SPC <return> <next> 
n <down-mouse-1> <mouse-1> <wheel-up> <double-wheel-up> 
<down-mouse-1> <mouse-1> <down-mouse-1> <mouse-1> <wheel-down> 
<down-mouse-1> <mouse-1> <down-mouse-1> <mouse-1> <down-mouse-1> 
<mouse-1> <wheel-down> <double-wheel-down> <triple-wheel-down> 
<triple-wheel-up> <triple-wheel-up> <wheel-down> <down-mouse-1> 
<mouse-1> M-x r e p o r t - = e <backspace> <backspace> 
e m a c s - b u g <return> E a m c s SPC <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
E m a c s SPC 2 3 SPC - SPC p e r f o r m a n c e SPC 
p r o b l e m s ? <return> <f1> <up> <up> <up> <up> 
<up> <up> <up> <up> <down> <down> <down> <down> <down> 
C-SPC C-n C-n C-n C-n C-n C-n C-n C-n C-p C-p C-p C-w 
C-o C-o C-v C-v C-x , C-x u C-x u C-x u C-x u C-x k 
<return> M-x r e p o r t = - e <backspace> <backspace> 
<backspace> - e m a c s - b u SPC <return>

Recent messages:
Quit [2 times]
Info-extract-pointer: Node has no Up
Mark saved where search started [2 times]
Mark set
Mark saved where search started
Parsing d:/Home/udr/.mailrc...
Parsing y:/dotfiles/.mailrc... done
Parsing d:/Home/udr/.mailrc... done
Mark set [2 times]
Undo! [4 times]

bug#5099: 23.1; Performance problems in Emacs 23?

[Stefan Monnier]

> However, having tried Emacs 23 for an hour, I have noticed very much
> the same phenomenon. My main mail file is roughly 100MB large.
> It takes Emacs 22 only about 5 seconds to save it, and the auto-save
> pause is probably half that long.  (I use a reasonably fast Solid
> State Drive.) The CPU utilisation is an A-shaped curve, implying that
> there is nothing much for the CPU to do. In contrast, Emacs 23 takes
> about more than 20 seconds to save the same mail file and the CPU
> utilisation is a pretty intensive burn for most of that time (20% on
> one processor and 60% on a second processor).  The auto-save pauses
> are about 10-15 seconds long.

> Now, VM mail files are raw-text.  Saving them should be an I/O-bound
> process.  It is not clear why Emacs 23 should be burning CPU cycles to
> save raw-text files and taking 3-5 times as long as Emacs 22.

That's indeed surprising and I can't think of anything that could cause
such a major slowdown.  Maybe if the VM mail buffers (containing
"raw-text") are left in multibyte mode, we do something inefficient
somehow (but then again, it'd be surprising for those buffers to be in
multibyte mode anyway).


        Stefan

bug#5099: 23.1; Performance problems in Emacs 23?

[Lars Ingebrigtsen]

Uday S Reddy <u.s.reddy@cs.bham.ac.uk> writes:

> Now, VM mail files are raw-text.  Saving them should be an I/O-bound
> process.  It is not clear why Emacs 23 should be burning CPU cycles to
> save raw-text files and taking 3-5 times as long as Emacs 22.

Is this problem still present in Emacs 24?

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#5099: 23.1; Performance problems in Emacs 23?

[Uday S Reddy]

Lars Ingebrigtsen writes:

> Is this problem still present in Emacs 24?

Dear Lars, Thanks for checking.  I haven't noticed these kind of slow downs
for a long time, and no other user has reported it either.  So, I think all
is well for now.

Cheers,
Uday