Date: prev next · Thread: first prev next last
2013 Archives by date, by thread · List index


Hi :)
+1

That is why i dont think it's worth us spending our time doing benchmarking.  Another  reason is 
that we are obviously a fairly biased bunch with only a few individuals that may be biased the 
other way or be "on the fence".  Anyone else looking at the results of a serious bench-marking 
study done by us will assume we have too much bias.  

Studies into this sort of thing are usually paid for by MS but they are usually able to hide the MS 
involvement sufficiently that people get the impression it's an independent bit of research.  

So, a quick bit of bench-marking by a few individuals has been great.  It's been really good to see 
and compare some of the different stats people have been giving but its not worth us spending much 
time over getting more than a "rule of thumb" or general idea.  

Thanks all for their work though!
Regards from 
Tom :)  







________________________________
From: Andrew Brown <andrewbr@icon.co.za>
To: 
Cc: users@global.libreoffice.org 
Sent: Thursday, 8 August 2013, 10:30
Subject: Re: [libreoffice-users] Re: start up speed


This is to what I have noticed by many who become anal retentive in this 
matter, and obsessive compulsive, over their possessions for want of a 
better word. I see it in the way they buy their worldly possessions from 
cars, to HiFi, to mobile phones, to their homes and it's contents. All 
of it must have a spec sheet a mile long to "PROVE" it's the best out 
there and better than yours and mine.

We are losing site of reality, as you covered, over .1 or .2 of a 
second. There's a real world we live in going to pot and soon we will 
battle for clean water and wholesome fresh produce, never mind the rapid 
loss of natural flora and fauna, over how many seconds a piece of 
silicon and soft, as in the real sense of what soft means, code is running.

My original purpose of starting this post was to simply show, in a real 
world use that LO is not slow by any means to any competitive product 
and does the job, for the majority of users, who are far in reality from 
their self proclaimed power users status, as equally good as any 
competitive product, but again my first paragraph observation is surfacing.

The bottom line there are millions of users of LO, and other users of 
non-MS products, working perfectly fine with it and could not really 
give a hoot of it's millisecond or second performance.

Andrew Brown

On 07/08/2013 04:29 PM, Kracked_P_P---webmaster wrote:

Yes, there are a lot of people who can give others "proof" of what 
their system can do.  Either by an active demonstration or via 
benchmark packages.

I love doing active presentations to non-believers.

I have taken and proved that LO can do things that I claim, by 
bringing my laptop[s] to these people and run LO through some tests.  
Then I hand them a USB drive and ask them to place some "sample" 
documents they use on a daily/weakly basis and then open them up with 
LO and show them that "yes" LO can work with your files "easily".

I once had them click on their Word icon on their multi-core Windows 
desktop and the same time as I click on the LO icon on my Windows boot 
partition of my Windows/Ubuntu dual booting laptop[s].  Now that we 
have Win7 as the "new standard" for business computers, I can use 
either my Win7 Home Premium laptop or my Win7 Professional one.  Both 
are dual core laptops, but the "Professional" install is on the slower 
system.  One day I will take the "Home" laptop and make it 
"Professional" to solve some "back port" issues where "Home" might not 
allow certain XP and Vista era packages to be installed while 
"Professional" has not problem installing those packages.

So having a live demonstration on what LO can do and how fast it can 
do those things is a good "marketing tool".

Having a "benchmark" style of information sheet tends to make many 
manager's eyes "cloud over and ears stop hearing you" as you discuss 
the benchmark results.

Yes, there can be some "guesswork" for some things, and some 
subjective issues, but it does not mean that those "guesses" are wrong.

The seconds count or timed with a stopwatch is not very accurate if it 
is the difference of a second plus/minus.  But most people can not 
tell the difference between 3 and 4 seconds, or 3.4 and 4.1 seconds.  
We are not built that way, or most of us are not build that way.



On 08/07/2013 08:05 AM, Sina Momken wrote:
On 08/07/2013 03:05 PM, Kracked_P_P---webmaster wrote:
YES my point exactly........

Unless we do such a large data gathering project taking into account 
all
of the different options, EVERYTHING is just guesswork or personally
view performance.
They are more than some guesswork which I may say from myself. They are
not only my opinions. Actually I could prove my claims to some degree
using some simple experiments.
Yeah, my experiments were not comprehensive enough to certainly conclude
from, but they can prove my claims with good probability (at least in my
case).
Proving by experimentation is not like in the math which can 100%
confirm a lemma. One must limit his experiments based on his time and
efforts; We can not put a ball down to ground on each planet to test the
gravity theory of Newton!
All what I want to say is that I believe based on my (not comprehensive)
experiments, my claims about LO performance is more than just a
GUESSWORK or OPINION!

Regards

Some faster systems, for whatever reasons, load and run LO slower than
the "slower CPU" called slower due to number of cores or the speed at
which it is running at.

So all it opinion until someone decides to "prove" those opinions and
results on an individual basis.



On 08/06/2013 10:09 PM, Sina Momken wrote:
On 08/07/2013 05:43 AM, Kracked_P_P---webmaster wrote:
I would expect that .doc would load slower in Writer and .odt 
would load
slower in Word.

The question really is how well does Writer load both. How well it 
load
the 10 page documents vs. the 50 page ones.  Both with the same 
average
number of graphics per page.

Then look at the simple 20 or 50 page documents vs. the very complex
ones.

Get an over all load times for the same documents on Writer and 
Word on
various Windows systems and various version of Windows [Win7 - 
Home/H.
Premium/Professional - 64-bit and 32-bit.  Vista versions in both 
32 and
64 bit.]  Then look into the same documents with Writer run on 
some of
the different version of Linux [32-bit and 64-bit OS] such as Ubuntu,
Fedora, Mint, Mageia, Arch, etc., etc..

Then with all that data make a chart and add to it every time someone
tries the "standard" documents on different systems and 
specifications.

Then we would have a chart that will tell us how much different 
systems
and specifications effect the load and run speeds of LO, Writer
specifically, and Word specifically.

Does more RAM or more CPU power influence it most.  How does 4.0.4 vs
4.1.0 compare on the same system/specs.  How much faster a 64-bit
install is over the same distro's 32-bit version.
What you're requesting here is an exact benchmark with will take so 
much
time and effort. Besides different file formats, size and heaviness of
the file, different OSes and different HW Architectures, the exact
conditions of the system during experiment (like the software and
processes running in the background, etc.) and the number of 
repetitions
for each experiment must also be specified. Ideally no other excessive
processes must be run and each experiment must run more than 10 times.
It's accurate to write a test program to automatically test these
factors with any repetition desired.

But doing all these is a major job and takes much time and effort. If
I'd done this before, I've published this on my website or other major
website, not on this mailing list which doesn't have many visitors.

I only wanted to show you a rule of thumb about LO Writer dealing with
heavy files.

Without these types of data charted, we could just say what we 
"think"
is true or want works better for you.

To be honest, when I was using it and it worked well, my AMD64 CPU
laptop worked better than my Intel dual core laptop.  When I asked 
why
my older slower AMD laptop worked faster creating the .iso file using
DeVeDe .avi/.mp4 file to DVD-movie disc conversion tool, I was 
told that
the faster dual core laptop was not powerful enough to do the work 
even
though my older slower AMD64 laptop could do it just fine.

So, no matter how I think it should not be true, sometimes newer 
faster
systems that we think is more powerful and faster might now be a 
good as
we think and the older slower less powerful systems might actually 
work
better at some job or package.  Slower single core laptop working 
better
than a faster speed dual core laptop, does not make sense, but in
practice it works that way.
I doesn't say that. Actually I exactly said opposite of that. I have a
single core pentium4 @2.8GHz desktop which runs LO Writer faster 
than my
dual core core2due @2.2GHz laptop. Maybe power of both cores of my
laptop be more than power of cpu of my desktop, but power of a single
core of my laptop is surely less than power of a single core of my
desktop and because LO only uses 1 core, my older desktop PC wins.

So, maybe someone should collect some data and let us know how it 
worked
out.  Maybe we could be surprised on what we find.
Making a precise benchmark is always a valuable and highly regarded
work, can practically assess a software and help to make it better.

I sure was running DeVeDe on 2 different laptops, both as XP/Vista 
and
Ubuntu 10.04/ U. 10.04 systems.
Regards,
      Sina Momken

<snip>



-- 
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and cannot be deleted



-- 
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and cannot be deleted

Context


Privacy Policy | Impressum (Legal Info) | Copyright information: Unless otherwise specified, all text and images on this website are licensed under the Creative Commons Attribution-Share Alike 3.0 License. This does not include the source code of LibreOffice, which is licensed under the Mozilla Public License (MPLv2). "LibreOffice" and "The Document Foundation" are registered trademarks of their corresponding registered owners or are in actual use as trademarks in one or more countries. Their respective logos and icons are also subject to international copyright laws. Use thereof is explained in our trademark policy.