Re: Pros and cons of various mailbox formats

2002-09-18 Thread Will Yardley

jz wrote:
> * Will Yardley <[EMAIL PROTECTED]> [2002-09-15 21:16]:
> > jz wrote:
> > > In my experience it is not so slow. Deleting a message from a
> > > 3-message maildir takes 3 to 5 seconds at most on my 4 yrs old
> > > celeron.

> > It's not the processor speed, but the filesystem that's the main issue
> > here. You don't mention what filesystem (or OS) you're using here.
 
> A little primitive test if anyone cares to read it. The machine
> was celeron 540 w/ 256M memory with a relatively new IDE and a
> SCSI oldie running FreeBSD 4.2-RC2. Task was opening a folder
> with 30300 messages, roughly a bit over 120M in mbox format.
> 
> 40G IBM IDE running ffs w/ softupdates enabled
>   maildir: 57 sec
>   mbox: 27 sec
> old 2G barracuda (SCSI) w/ same fs parameters
>   maildir: 69 sec
>   mbox: 17 sec
> 
> Machine was unloaded, though I think the numbers could vary for
> several seconds in both directions if I repeated it several
> times. Please keep in mind that this test is by no means
> representative, it just gives a very rough comparison between
> formats. I have no idea whatsoever what the result would be on a
> decent SCSI based machine running Solaris with logging on or
> Linux async mounted ext2 or whatever. YMMV.
> 
> The beauty of maildir is faster deleting, updating, writing and
> higher reliability.
> 
> Maybe mutt power users or developers could shed more light on
> the subject.

The header caching patch seems to help a little for Maildirs, although
it still seems to be a bit buggy.

I'd say that the same folder would take a lot longer to open on ext2.

-- 
Will Yardley
input: william < @ hq . newdream . net . >




Re: Pros and cons of various mailbox formats

2002-09-18 Thread PeterKorman



I just did some trivial tests.

I'm working on something with 34924 files in a single directory
on linux ext3. I timed a simple file remove it took 0.044 seconds.
Most people I know dont call this a low end system. Based on this 
data I think that this particular machine can tolerate any overhead
introduced by Maildir. I think this will remain true for a week
or two. Your milage may vary.

---
$ time rm x22407~

real0m0.044s
user0m0.000s
sys 0m0.000s
--
$ ls -1 | wc -l
   34924

--
processor   : 0
vendor_id   : AuthenticAMD
cpu family  : 6
model   : 6
model name  : AMD Athlon(TM) XP 1900+
stepping: 2
cpu MHz : 1600.095
cache size  : 256 KB
fdiv_bug: no
hlt_bug : no
f00f_bug: no
coma_bug: no
fpu : yes
fpu_exception   : yes
cpuid level : 1
wp  : yes
flags   : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca
cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow
bogomips: 3191.60
--

Disk is Maxtor 96147H6 IDE running off the ASUS A7M266 motherboard controller.



On Wed, Sep 18, 2002 at 04:10:46PM +0200, jz wrote:
> * Will Yardley <[EMAIL PROTECTED]> [2002-09-15 21:16]:
> > jz wrote:
> > > In my experience it is not so slow. Deleting a message from a
> > > 3-message maildir takes 3 to 5 seconds at most on my 4 yrs old
> > > celeron.
> > 
> > It's not the processor speed, but the filesystem that's the main issue
> > here. You don't mention what filesystem (or OS) you're using here.
> 
> A little primitive test if anyone cares to read it. The machine
> was celeron 540 w/ 256M memory with a relatively new IDE and a
> SCSI oldie running FreeBSD 4.2-RC2. Task was opening a folder
> with 30300 messages, roughly a bit over 120M in mbox format.
> 
> 40G IBM IDE running ffs w/ softupdates enabled
>   maildir: 57 sec
>   mbox: 27 sec
> old 2G barracuda (SCSI) w/ same fs parameters
>   maildir: 69 sec
>   mbox: 17 sec
> 
> Machine was unloaded, though I think the numbers could vary for
> several seconds in both directions if I repeated it several
> times. Please keep in mind that this test is by no means
> representative, it just gives a very rough comparison between
> formats. I have no idea whatsoever what the result would be on a
> decent SCSI based machine running Solaris with logging on or
> Linux async mounted ext2 or whatever. YMMV.
> 
> The beauty of maildir is faster deleting, updating, writing and
> higher reliability.
> 
> Maybe mutt power users or developers could shed more light on
> the subject.
> 
> jz

-- 
GnuPG: ECBA EA08 C3C1 251E 5FB5  D196 F8C8 F8B7 AB60 234D



msg31029/pgp0.pgp
Description: PGP signature


Re: Pros and cons of various mailbox formats

2002-09-18 Thread jz

* Will Yardley <[EMAIL PROTECTED]> [2002-09-15 21:16]:
> jz wrote:
> > In my experience it is not so slow. Deleting a message from a
> > 3-message maildir takes 3 to 5 seconds at most on my 4 yrs old
> > celeron.
> 
> It's not the processor speed, but the filesystem that's the main issue
> here. You don't mention what filesystem (or OS) you're using here.

A little primitive test if anyone cares to read it. The machine
was celeron 540 w/ 256M memory with a relatively new IDE and a
SCSI oldie running FreeBSD 4.2-RC2. Task was opening a folder
with 30300 messages, roughly a bit over 120M in mbox format.

40G IBM IDE running ffs w/ softupdates enabled
maildir: 57 sec
mbox: 27 sec
old 2G barracuda (SCSI) w/ same fs parameters
maildir: 69 sec
mbox: 17 sec

Machine was unloaded, though I think the numbers could vary for
several seconds in both directions if I repeated it several
times. Please keep in mind that this test is by no means
representative, it just gives a very rough comparison between
formats. I have no idea whatsoever what the result would be on a
decent SCSI based machine running Solaris with logging on or
Linux async mounted ext2 or whatever. YMMV.

The beauty of maildir is faster deleting, updating, writing and
higher reliability.

Maybe mutt power users or developers could shed more light on
the subject.

jz