1. Eu 2.5 verryy sloww on Win XP ?

I have downloaded the alpha Eu 2.5, and it seems odd.
On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2% 
of the speed of Eu 2.4. I installed it clean (renamed previous Euphoria
directory) and tried WinWire demo program - you can follow the letter E
quite easily, whereas at home it is all a blur and hard to see at all.
I tried Judith's IDE (I have a bound version which is reliable) and it
took over a minute just for the splash screen to appear.

So - does anyone have any suggestions? it sounds decidedly weird to me.

Andy

new topic     » topic index » view message » categorize

2. Re: Eu 2.5 verryy sloww on Win XP ?

I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes slow.
Daniel


On Thu, 18 Nov 2004 02:36:31 -0800, Andy Drummond
<guest at rapideuphoria.com> wrote:
> 
> posted by: Andy Drummond <andy at kestreltele.com>
> 
> I have downloaded the alpha Eu 2.5, and it seems odd.
> On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
> On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2%
> of the speed of Eu 2.4. I installed it clean (renamed previous Euphoria
> directory) and tried WinWire demo program - you can follow the letter E
> quite easily, whereas at home it is all a blur and hard to see at all.
> I tried Judith's IDE (I have a bound version which is reliable) and it
> took over a minute just for the splash screen to appear.
> 
> So - does anyone have any suggestions? it sounds decidedly weird to me.
> 
> Andy
> 
> 
> 
>

new topic     » goto parent     » topic index » view message » categorize

3. Re: Eu 2.5 verryy sloww on Win XP ?

codepilot Gmail Account wrote:
> 
> I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes slow.
> Daniel
> 

No problems here on my AMD64 3500+.  Can't tell the difference between 2.4 &
2.5.  It even runs pretty fast on the PD source code version...

new topic     » goto parent     » topic index » view message » categorize

4. Re: Eu 2.5 verryy sloww on Win XP ?

Yes, I found this to be the case on Win XP as well. 
I tested 2.5 with a Windows program, and you could tell right away by
how long it took for the program window to come up compared to 2.4. 

'Debo

codepilot Gmail Account wrote:
> 
> I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes slow.
> Daniel
> 
> 
> On Thu, 18 Nov 2004 02:36:31 -0800, Andy Drummond
> <guest at rapideuphoria.com> wrote:
> > 
> > posted by: Andy Drummond <andy at kestreltele.com>
> > 
> > I have downloaded the alpha Eu 2.5, and it seems odd.
> > On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
> > On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2%
> > of the speed of Eu 2.4. I installed it clean (renamed previous Euphoria
> > directory) and tried WinWire demo program - you can follow the letter E
> > quite easily, whereas at home it is all a blur and hard to see at all.
> > I tried Judith's IDE (I have a bound version which is reliable) and it
> > took over a minute just for the splash screen to appear.
> > 
> > So - does anyone have any suggestions? it sounds decidedly weird to me.
> > 
> > Andy
> > 
> >

new topic     » goto parent     » topic index » view message » categorize

5. Re: Eu 2.5 verryy sloww on Win XP ?

Gbadebo Oladosu wrote:
> Yes, I found this to be the case on Win XP as well. 
> I tested 2.5 with a Windows program, and you could tell right away by
> how long it took for the program window to come up compared to 2.4. 
> 
> 'Debo
> 
> codepilot Gmail Account wrote:
> > 
> > I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes slow.
> > Daniel
> > 
> > On Thu, 18 Nov 2004 02:36:31 -0800, Andy Drummond
> > <guest at rapideuphoria.com> wrote:
> > > 
> > > posted by: Andy Drummond <andy at kestreltele.com>
> > > 
> > > I have downloaded the alpha Eu 2.5, and it seems odd.
> > > On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
> > > On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2%
> > > of the speed of Eu 2.4. I installed it clean (renamed previous Euphoria
> > > directory) and tried WinWire demo program - you can follow the letter E
> > > quite easily, whereas at home it is all a blur and hard to see at all.
> > > I tried Judith's IDE (I have a bound version which is reliable) and it
> > > took over a minute just for the splash screen to appear.
> > > 
> > > So - does anyone have any suggestions? it sounds decidedly weird to me.

winwire looks pretty fast to me. 2.5 alpha, XP, Pentium 4 1.8 GHz.

Judith's IDE is a special case.
It has code at the very beginning of the source 
to display a splash screen. Under 2.4 this splash
screen appears almost immediately, then you wait for parsing
to complete. That's because 2.4 will execute code before it 
has finished parsing the program. 2.5 parses the *whole* program before
executing anything. Including Win32Lib etc., the IDE is
100,000 lines of Euphoria code. That's a lot
of parsing to do, and 2.5 has a Euphoria-coded parser.
2.4 has a C-coded parser. So you might consider 
the IDE to be the "worst-case" example. It's the biggest
Euphoria program I know of.
If you bind the IDE using 2.5, or translate/compile it, 
it will start up much faster, because no parsing need be done. 
(Bound programs under 2.4 must be parsed).
Other very large (tens of thousands of lines) programs may also 
appear to start up a tad slower using the 2.5 interpreter, 
but we're only talking about a second or two (unless you have an ancient
machine). I think with the vast majority of programs, 
you'd hardly notice any difference. As computers get faster, 
this small difference will get even smaller.

Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com

new topic     » goto parent     » topic index » view message » categorize

6. Re: Eu 2.5 verryy sloww on Win XP ?

I just timed:
   exw ide 
(ide-18), and it took 6 seconds to completely start up.

Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com

new topic     » goto parent     » topic index » view message » categorize

7. Re: Eu 2.5 verryy sloww on Win XP ?

With the PD Euphoria source:
   exw \euphoria\source\eu.ex ide.exw
I get:
    34 seconds total
    14 seconds with WinMain() and openWindow(splash... removed.

With:
   exw ide.exw
I get:
    6 seconds total
  4.5 seconds with WinMain() and openWindow(splash... removed

So there's more than just parsing. The IDE has to initialize
a bunch of stuff as well.

Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com

new topic     » goto parent     » topic index » view message » categorize

8. Re: Eu 2.5 verryy sloww on Win XP ?

Robert Craig wrote:

> Robert Craig wrote:
>> Maybe you've left "with trace(3)" in one
>> of your files. That will kill you.
>
> Sorry, I meant:
>     with trace
>
>     ...
>
>     trace(3)
>
> Look for "ctrace.out" in the current directory.
> Even "with trace" by itself will waste about
> 10% or more of your performance.

You mean only the statement
   with trace

without any call of
   trace(1),
   trace(2) or
   trace(3)

in the program? Huuu ...
I think that is worth to be mentioned in the documentation
(couldn't find this info in the docs).

On this occasion I read in 'lib_s_t.htm#trace':
"All forms of trace() are supported in the Public Domain interpreter,
but only for programs up to 300 statements in size. For larger programs
you'll need the Complete Edition interpreter."

But in 'relnotes.htm' it reads:
"The free Public Domain Interpreter (ex.exe, exw.exe, exu), now includes
full support for trace() beyond 300 statements, and for profiling. Thus
there is no longer a registered 'Complete Edition' interpreter."

Regards,
   Juergen

new topic     » goto parent     » topic index » view message » categorize

9. Re: Eu 2.5 verryy sloww on Win XP ?

Juergen Luethje wrote:
> You mean only the statement
>    with trace
> 
> without any call of
>    trace(1),
>    trace(2) or
>    trace(3)
> 
> in the program? Huuu ...
> I think that is worth to be mentioned in the documentation
> (couldn't find this info in the docs).

It's in perform.htm near the top.
In the PD source you can see that extra IL ops
are emitted wherever "with trace" is in effect.

> On this occasion I read in 'lib_s_t.htm#trace':
> "All forms of trace() are supported in the Public Domain interpreter,
> but only for programs up to 300 statements in size. For larger programs
> you'll need the Complete Edition interpreter."

Whoops. I missed that, plus a couple of other lingering references
to "Complete Edition". 

There is now only one, fully-enabled, interpreter.

Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com

new topic     » goto parent     » topic index » view message » categorize

10. Re: Eu 2.5 verryy sloww on Win XP ?

>>> guest at RapidEuphoria.com 11/18/2004 3:00:09 PM >>>
There is now only one, fully-enabled, interpreter.

Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com 
-----------------------------------------------------

Then why did trace(1) in an exu program still come back with an
error?!?
Better make that 2 things to fix, eh?

Mike

new topic     » goto parent     » topic index » view message » categorize

11. Re: Eu 2.5 verryy sloww on Win XP ?

Robert Craig wrote:
> 
> Gbadebo Oladosu wrote:
> > Yes, I found this to be the case on Win XP as well. 
> > I tested 2.5 with a Windows program, and you could tell right away by
> > how long it took for the program window to come up compared to 2.4. 
> > 
> > 'Debo
> > 
> > codepilot Gmail Account wrote:
> > > 
> > > I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes slow.
> > > Daniel
> > > 
> > > On Thu, 18 Nov 2004 02:36:31 -0800, Andy Drummond
> > > <guest at rapideuphoria.com> wrote:
> > > > 
> > > > posted by: Andy Drummond <andy at kestreltele.com>
> > > > 
> > > > I have downloaded the alpha Eu 2.5, and it seems odd.
> > > > On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
> > > > On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2%
> > > > of the speed of Eu 2.4. I installed it clean (renamed previous Euphoria
> > > > directory) and tried WinWire demo program - you can follow the letter E
> > > > quite easily, whereas at home it is all a blur and hard to see at all.
> > > > I tried Judith's IDE (I have a bound version which is reliable) and it
> > > > took over a minute just for the splash screen to appear.
> > > > 
> > > > So - does anyone have any suggestions? it sounds decidedly weird to me.
> 
> winwire looks pretty fast to me. 2.5 alpha, XP, Pentium 4 1.8 GHz.
Using a Athlon 64 2.8Ghz with 2GB RAM NVIDIA GeForce4 440
Go 64MB Gfx card, Win XP.
2.4: looks like a spinning E
2.5A: looks like colored static(that's quite fast!)

> 
> Judith's IDE is a special case.
Time till splash screen:
2.4: less than 1 second
2.5a: 3 seconds

Time till completely started:
2.4: about 10 seconds
2.5a: about 5 seconds

CJBN Webserver start time:
2.4: 7 seconds (4 till splash)
2.5a: 2 seconds (2 till splash)

The Win32Dib demos(fps):
2.4: 20-800
2.5a: 100-1000

2.5 alpha seems to start the program slower,
but overall it runs much faster.
I care more about program speed than startup
time.


> It has code at the very beginning of the source 
> to display a splash screen. Under 2.4 this splash
> screen appears almost immediately, then you wait for parsing
> to complete. That's because 2.4 will execute code before it 
> has finished parsing the program. 2.5 parses the *whole* program before
> executing anything. Including Win32Lib etc., the IDE is
> 100,000 lines of Euphoria code. That's a lot
> of parsing to do, and 2.5 has a Euphoria-coded parser.
> 2.4 has a C-coded parser. So you might consider 
> the IDE to be the "worst-case" example. It's the biggest
> Euphoria program I know of.
> If you bind the IDE using 2.5, or translate/compile it, 
> it will start up much faster, because no parsing need be done. 
> (Bound programs under 2.4 must be parsed).
> Other very large (tens of thousands of lines) programs may also 
> appear to start up a tad slower using the 2.5 interpreter, 
> but we're only talking about a second or two (unless you have an ancient
> machine). I think with the vast majority of programs, 
> you'd hardly notice any difference. As computers get faster, 
> this small difference will get even smaller.
> 
> Regards,
>    Rob Craig
>    Rapid Deployment Software
>    <a href="http://www.RapidEuphoria.com">http://www.RapidEuphoria.com</a>
>

new topic     » goto parent     » topic index » view message » categorize

12. Re: Eu 2.5 verryy sloww on Win XP ?

We'll I just tried again to make sure, and 2.4 winwire very very
fast>60fps, 2.5 like <20fps I think. 2400+athlonxp win98se.
Dan


On Thu, 18 Nov 2004 19:44:18 -0800, CoJaBo <guest at rapideuphoria.com> wrote:
> 
> posted by: CoJaBo <cojabo at suscom.net>
> 
> 
> Robert Craig wrote:
> >
> > Gbadebo Oladosu wrote:
> > > Yes, I found this to be the case on Win XP as well.
> > > I tested 2.5 with a Windows program, and you could tell right away by
> > > how long it took for the program window to come up compared to 2.4.
> > >
> > > 'Debo
> > >
> > > codepilot Gmail Account wrote:
> > > >
> > > > I have 2600+ AMD XP, win98 euphoria 2.5 alpha, and winwire also goes
> > > > slow.
> > > > Daniel
> > > >
> > > > On Thu, 18 Nov 2004 02:36:31 -0800, Andy Drummond
> > > > <guest at rapideuphoria.com> wrote:
> > > > >
> > > > > posted by: Andy Drummond <andy at kestreltele.com>
> > > > >
> > > > > I have downloaded the alpha Eu 2.5, and it seems odd.
> > > > > On my home PC - Pentium 3 500MHz Win 98SE - it runs fast as you like.
> > > > > On my work PC - Pentium 4 2.8GHz Win XP & SP2 - it runs about 1-2%
> > > > > of the speed of Eu 2.4. I installed it clean (renamed previous
> > > > > Euphoria
> > > > > directory) and tried WinWire demo program - you can follow the letter
> > > > > E
> > > > > quite easily, whereas at home it is all a blur and hard to see at all.
> > > > > I tried Judith's IDE (I have a bound version which is reliable) and it
> > > > > took over a minute just for the splash screen to appear.
> > > > >
> > > > > So - does anyone have any suggestions? it sounds decidedly weird to
> > > > > me.
> >
> > winwire looks pretty fast to me. 2.5 alpha, XP, Pentium 4 1.8 GHz.
> Using a Athlon 64 2.8Ghz with 2GB RAM NVIDIA GeForce4 440
> Go 64MB Gfx card, Win XP.
> 2.4: looks like a spinning E
> 2.5A: looks like colored static(that's quite fast!)
> 
> >
> > Judith's IDE is a special case.
> Time till splash screen:
> 2.4: less than 1 second
> 2.5a: 3 seconds
> 
> Time till completely started:
> 2.4: about 10 seconds
> 2.5a: about 5 seconds
> 
> CJBN Webserver start time:
> 2.4: 7 seconds (4 till splash)
> 2.5a: 2 seconds (2 till splash)
> 
> The Win32Dib demos(fps):
> 2.4: 20-800
> 2.5a: 100-1000
> 
> 2.5 alpha seems to start the program slower,
> but overall it runs much faster.
> I care more about program speed than startup
> time.
> 
> 
> > It has code at the very beginning of the source
> > to display a splash screen. Under 2.4 this splash
> > screen appears almost immediately, then you wait for parsing
> > to complete. That's because 2.4 will execute code before it
> > has finished parsing the program. 2.5 parses the *whole* program before
> > executing anything. Including Win32Lib etc., the IDE is
> > 100,000 lines of Euphoria code. That's a lot
> > of parsing to do, and 2.5 has a Euphoria-coded parser.
> > 2.4 has a C-coded parser. So you might consider
> > the IDE to be the "worst-case" example. It's the biggest
> > Euphoria program I know of.
> > If you bind the IDE using 2.5, or translate/compile it,
> > it will start up much faster, because no parsing need be done.
> > (Bound programs under 2.4 must be parsed).
> > Other very large (tens of thousands of lines) programs may also
> > appear to start up a tad slower using the 2.5 interpreter,
> > but we're only talking about a second or two (unless you have an ancient
> > machine). I think with the vast majority of programs,
> > you'd hardly notice any difference. As computers get faster,
> > this small difference will get even smaller.
> >
> > Regards,
> >    Rob Craig
> >    Rapid Deployment Software
> >    <a href="http://www.RapidEuphoria.com">http://www.RapidEuphoria.com</a>
> 
> 
> 
> 
>

new topic     » goto parent     » topic index » view message » categorize

13. Re: Eu 2.5 verryy sloww on Win XP ?

codepilot Gmail Account wrote:
> 
> We'll I just tried again to make sure, and 2.4 winwire very very
> fast>60fps, 2.5 like <20fps I think. 2400+athlonxp win98se.
> Dan
> 

There must be something wrong with your system, maybe with the 
video drivers or something.

I get 233 fps under Eu 2.5 on my lowly VIA C3 733Mhz WinXP Sp2 Desknote.  
The AthlonXP got to be faster than that.

new topic     » goto parent     » topic index » view message » categorize

14. Re: Eu 2.5 verryy sloww on Win XP ?

ok, found the cause, line 558 in winwire, and I added a frame
counter(cause me eyes don't count that well) in 2.4 count=5-fps=90,
500=1100, 2.5 5=90 500=1400, so basically all things
considered(properly) the count change messed me up, but 2.5 is faster
1100vs1400.

Daniel
2.5 is winwire line 558
	for i = 1 to 5 do
	    spin()
	end for
2.4 is winwire line 558
        for i = 1 to 500 do
            spin()
        end for



On Thu, 18 Nov 2004 23:30:31 -0800, Ken Orr <guest at rapideuphoria.com> wrote:
> 
> posted by: Ken Orr <orr_kenneth at yahoo.ca>
> 
> codepilot Gmail Account wrote:
> >
> > We'll I just tried again to make sure, and 2.4 winwire very very
> > fast>60fps, 2.5 like <20fps I think. 2400+athlonxp win98se.
> > Dan
> >
> 
> There must be something wrong with your system, maybe with the
> video drivers or something.
> 
> I get 233 fps under Eu 2.5 on my lowly VIA C3 733Mhz WinXP Sp2 Desknote.
> The AthlonXP got to be faster than that.
> 
> 
> This email was sent to: codepilot at gmail.com
> 
> 
>

new topic     » goto parent     » topic index » view message » categorize

15. Re: Eu 2.5 verryy sloww on Win XP ?

Robert Craig wrote:

> Juergen Luethje wrote:
>> You mean only the statement
>>    with trace
>>
>> without any call of
>>    trace(1),
>>    trace(2) or
>>    trace(3)
>>
>> in the program? Huuu ...
>> I think that is worth to be mentioned in the documentation
>> (couldn't find this info in the docs).
>
> It's in perform.htm near the top.

I see now. smile Thanks.

> In the PD source you can see that extra IL ops
> are emitted wherever "with trace" is in effect.

BTW ... Thanks a lot for the PD source!! I can learn very much by
studying it!

<snip>

Regards,
   Juergen

new topic     » goto parent     » topic index » view message » categorize

Search



Quick Links

User menu

Not signed in.

Misc Menu