PSN-L Email List Message

Subject: Re: Basic Programming Help desired?
From: ian ian@...........
Date: Fri, 09 May 2008 16:04:25 +0100


Hi,

a good compiler might well optimise it out but developing fast code 
involves a bit of trial and error, which is why I suggested trying it.  
If it doesn't give a useful gain, then just take it out again.  
Compilers are mysterious beasts.  I suspect that Basic compilers will be 
amongst the most bestial.  :-)

I can understand the initial deterrent of migrating to another system 
but if you use the right tools then you'll get the best results.  When 
Linux first came out, I ran it on an old 486 machine.  These days, the 
slowest Linux machine I'm running is a 586 220 MHz - it just sits there, 
month after month, doing its job!

Ian

Kevin Brunt wrote:
> Hi,
>
> While I am not going to argue against the "unrolling" of the code within a loop, my take on it is that it is something that a good optimising compiler will do for you. In your example, I suspect that the cost of the repeated "multiply-by-sixteen" operations would lose a lot of the time saved by reducing the amount of "iteration overhead".
>
> I don't think that a "Linux plus C" approach will serve Geoff's purpose. Apart from the fact that he obviously would rather get on with the seismology, rather than learning another programming language, he's wanting to run his data capture system on a low-spec machine (retired from other uses, I suspect).
>
> Kevin
>
> ---- Original message ----
>   
>> Date: Fri, 09 May 2008 07:42:12 +0100
>> From: ian ian@...
>>     
>
>   
>>   I used the inline code example (I posted earlier) on
>>   PDP 11 computers.
>>     
>
>   
>>   I wrote the time critical parts in assembler and the
>>   rest in Pascal.  As for the current PSN problem, I
>>   would suggest scrapping basic and windows and going
>>   for a Linux machine and program in C.
>>     
> __________________________________________________________
>
> Public Seismic Network Mailing List (PSN-L)
>
> To leave this list email PSN-L-REQUEST@.............. with 
> the body of the message (first line only): unsubscribe
> See http://www.seismicnet.com/maillist.html for more information.
>
>
>   



  


Hi,

a good compiler might well optimise it out but developing fast code involves a bit of trial and error, which is why I suggested trying it.  If it doesn't give a useful gain, then just take it out again.  Compilers are mysterious beasts.  I suspect that Basic compilers will be amongst the most bestial.  :-)

I can understand the initial deterrent of migrating to another system but if you use the right tools then you'll get the best results.  When Linux first came out, I ran it on an old 486 machine.  These days, the slowest Linux machine I'm running is a 586 220 MHz - it just sits there, month after month, doing its job!

Ian

Kevin Brunt wrote:
Hi,

While I am not going to argue against the "unrolling" of the code within a loop, my take on it is that it is something that a good optimising compiler will do for you. In your example, I suspect that the cost of the repeated "multiply-by-sixteen" operations would lose a lot of the time saved by reducing the amount of "iteration overhead".

I don't think that a "Linux plus C" approach will serve Geoff's purpose. Apart from the fact that he obviously would rather get on with the seismology, rather than learning another programming language, he's wanting to run his data capture system on a low-spec machine (retired from other uses, I suspect).

Kevin

---- Original message ----
  
Date: Fri, 09 May 2008 07:42:12 +0100
From: ian ian@...
    

  
  I used the inline code example (I posted earlier) on
  PDP 11 computers.
    

  
  I wrote the time critical parts in assembler and the
  rest in Pascal.  As for the current PSN problem, I
  would suggest scrapping basic and windows and going
  for a Linux machine and program in C.
    
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)

To leave this list email PSN-L-REQUEST@.............. with 
the body of the message (first line only): unsubscribe
See http://www.seismicnet.com/maillist.html for more information.


  

[ Top ] [ Back ] [ Home Page ]