Prev: String decoding
Next: A replacement for about-to-be-closed NGs
From: Nobody on 24 May 2010 03:32
"MM" <kylix_is(a)yahoo.co.uk> wrote in message
news:c97kv51emvrouprp464jiofrmc235vt5tc(a)4ax.com...
> On Mon, 24 May 2010 01:44:38 -0400, "Nobody" <nobody(a)nobody.com>
> wrote:
>
>>"MM" <kylix_is(a)yahoo.co.uk> wrote in message
>>news:aaiiv51rijeihe7sd3ud0bbdo9npscskg2(a)4ax.com...
>>> I'm talking about 25
>>> secs on the Win 98SE box and 3 secs on the W2K PC.
>>
>>I think you are seeing the effects of file system cache. You need to
>>restart
>>to get a comparable result.
>
> Eh?
>
> The 25 secs figure on 98SE never varies more than a second either way.
> The W2K figure is consistently a fraction of what it is on 98SE.

I meant the entire MDB file was cached in memory. To make sure that the file
is not cached, do a restart. Only the first test is valid as subsequent
tests may use the cached file in memory. To save restarts, you may want to
make multiple duplicates of the MDB file with different file names.


From: MM on 25 May 2010 01:35
On Mon, 24 May 2010 13:45:32 -0700, Gutless Umbrella Carrying Sissy
<taustinca(a)gmail.com> wrote:

>MM <kylix_is(a)yahoo.co.uk> wrote in
>news:aaiiv51rijeihe7sd3ud0bbdo9npscskg2(a)4ax.com:
>
>> The only
>> thing that is different is that the W2K PC has 1gb of RAM
>> whereas my standard Win 98SE PC has 512mb.
>>
>How can you possibly expect a meaningful comparison with that kind of
>difference in environment?

I didn't expect such a significant increase in performance on W2K.

MM
From: MM on 26 May 2010 01:56
On Tue, 25 May 2010 10:30:15 -0700, Gutless Umbrella Carrying Sissy
<taustinca(a)gmail.com> wrote:

>MM <kylix_is(a)yahoo.co.uk> wrote in
>news:37omv5tl0udbiv4t0m9peghmh9prfjsakm(a)4ax.com:
>
>> On Mon, 24 May 2010 13:45:32 -0700, Gutless Umbrella Carrying
>> Sissy <taustinca(a)gmail.com> wrote:
>>
>>>MM <kylix_is(a)yahoo.co.uk> wrote in
>>>news:aaiiv51rijeihe7sd3ud0bbdo9npscskg2(a)4ax.com:
>>>
>>>> The only
>>>> thing that is different is that the W2K PC has 1gb of RAM
>>>> whereas my standard Win 98SE PC has 512mb.
>>>>
>>>How can you possibly expect a meaningful comparison with that
>>>kind of difference in environment?
>>
>> I didn't expect such a significant increase in performance on
>> W2K.
>
>A comparison between W2K and NT 4 would be at least somewhat
>meaningful, sice the former is a "new and improved" version of the
>latter, but 98 is a completely different beast, built, basically,
>as an app running on DOS.
>
>But even W2K/NT 4 isn't a very good way to try to compare two
>different software packages. The only valid comparison is to run
>both on the same hardware/OS.

I really don't see where you're coming from. Let me reiterate: Some
weeks ago I posted here an enquiry about the slow performance of ADO
on Windows 98. Several respondents suggested I try DAO instead, which
I did, and found it somewhat faster. The question of a different OS
wasn't raised. It's only now that I try the very same app and mdb on a
W2K PC that I realise that there is a considerable improvement in
performance simply by running the app on W2K. Therefore, the 'problem'
has simply gone away.

Other suggestions made at the time when I originally raised this issue
were to try a different database, e.g. SQLite, which is considerably
faster. I am still looking into this. So this all has nothing to do
with *comparing* performance between 98SE and more recent op systems;
it is merely an observation of the increased performance that can be
had from ADO 'out of the box' with these later OSs.

MM
From: ralph on 26 May 2010 06:43
On Wed, 26 May 2010 06:56:01 +0100, MM <kylix_is(a)yahoo.co.uk> wrote:

> ... So this all has nothing to do
>with *comparing* performance between 98SE and more recent op systems;
>it is merely an observation of the increased performance that can be
>had from ADO 'out of the box' with these later OSs.
>

In a way it does, as the answer is Yes - Windows applications that use
'System' services such as VMM, base file I/O, and OLE will run
'faster' on Windows 2000 compared to Win98SE.

Don't forget you increased the physical RAM. You would also see
improved performance by increasing your RAM from 1G to 2G. (Also Win98
perfomance would also improve with an equal increase from 512kb to
1gb.)

-ralph
From: David Kerber on 26 May 2010 07:53
In article <uedpv599pduc2e89m4dvd8hlurk9t0tg8n(a)4ax.com>,
kylix_is(a)yahoo.co.uk says...

....

> >A comparison between W2K and NT 4 would be at least somewhat
> >meaningful, sice the former is a "new and improved" version of the
> >latter, but 98 is a completely different beast, built, basically,
> >as an app running on DOS.
> >
> >But even W2K/NT 4 isn't a very good way to try to compare two
> >different software packages. The only valid comparison is to run
> >both on the same hardware/OS.
>
> I really don't see where you're coming from. Let me reiterate: Some
> weeks ago I posted here an enquiry about the slow performance of ADO
> on Windows 98. Several respondents suggested I try DAO instead, which
> I did, and found it somewhat faster. The question of a different OS
> wasn't raised. It's only now that I try the very same app and mdb on a
> W2K PC that I realise that there is a considerable improvement in
> performance simply by running the app on W2K. Therefore, the 'problem'
> has simply gone away.

I did some testing early on of Win95 and 98 against the pure 32-bit
operating systems, both Windows NT, and Windows 2000. My experience was
that ALL 32-bit apps were faster under both NT and 2k than they were
under 9x, often dramatically so. The difference between NT and 2k was
minimal on equivalent hardware.

D
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: String decoding
Next: A replacement for about-to-be-closed NGs