[prev in list] [next in list] [prev in thread] [next in thread] 

List:       nsbasic-palm
Subject:    Re: [nsbasic-palm] Fw: Bench2 and NSBasic 3.0.0d
From:       George Henne <gh () nsbasic ! com>
Date:       2002-10-30 12:52:29
Message-ID: 200210301252.EAA30472 () mars ! he ! net
[Download RAW message or body]

>Forwarded from Laurent (Bench2 author):
>
>> Here are what I got on my Clié N770C :
>> - let's remember v2.1.1d result (with runtime) : 96.05 seconds
>> - new v3.0.0d with runtime : 62.79 seconds
>> - new v3.0.0d without runtime : 62.62 seconds
>> So, the improvement is significative and it is even a little bit better
>with
>> the FAT application !
>>
>> Now, I just need to update my web page with these results :-)
>> Thanks for the help.
>
>See his page:
>http://www.aldweb.com/palmknow/usbench.htm

The trouble with benchmarks like this is that they can be very 
misleading, and have little to do with the real world. Let me give a 
couple of reasons:

1. You can always write a very fast app if you keep all variables in 
Dynamic Store. This works fine in a test, with a limited number of 
variables. For a sophisticated app, the Dynamic Store is never large 
enough, so the much slower regular store has to be used. If you're 
working in something like C, the complexities of deciding where each 
variable should best reside can be handled by the programmer. For NS 
Basic, we didn't want to have programmers worrying about such stuff, so 
NS Basic manages it. There is definite overhead to this convenience, but 
it sure saves a lot of development time.

2. Use of the Palm object APIs: Palm has a very nice, efficient set of 
form and object API functions. Programs that make full use of them (like 
NS Basic) will be pretty quick at displaying and updating forms, which is 
arguably more of a factor in most Palm apps than calculating perfect 
numbers. Once again, NS Basic concentrates on making it easy to use the 
Palm object API's.

The speed of NS Basic is pretty good for most apps. On the new Palm OS 5 
devices, we've seen elapsed time for incrementing a value 20,000 times 
take just 8 seconds. It's rare that that kind of heavy processing is 
needed in normal apps. 

A benchmark like this needs to be taken in context of the total project: 
- How much of the job is CPU bound? 
- What long will it take to develop your app? 
- How good is the look and feel of the resulting project?
- How compatible will it be with various Palm OS devices? 
- How is the support? 
- How easy is it to deploy? 
- How long has the tools vendor been around - is he likely to be there 
for the full life cycle of your product?

We will continue to improve straight line performance of NS Basic, but 
all these other things are very important to us.

George Henne                                     
NS BASIC Corporation
http://www.nsbasic.com 


Community email addresses:
 Post message: nsbasic-palm@yahoogroups.com 
 Subscribe:  nsbasic-palm-subscribe@yahoogroups.com 
 Unsubscribe:  nsbasic-palm-unsubscribe@yahoogroups.com  
 List owner:  nsbasic-palm-owner@yahoogroups.com 

Shortcut URL to this page:
 http://groups.yahoo.com/group/nsbasic-palm  

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 



[prev in list] [next in list] [prev in thread] [next in thread]