Path: csiph.com!usenet.pasdenom.info!weretis.net!feeder4.news.weretis.net!ecngs!feeder2.ecngs.de!newsfeed.freenet.ag!news2.euro.net!postnews2.euro.net!news.wanadoo.nl!not-for-mail
From: mhx@iae.nl (Marcel Hendrix)
Subject: Re: GA144 polyForth
Newsgroups: comp.lang.forth
Message-ID: <70899297938435@frunobulax.edu>
Date: Sat, 6 Oct 2012 11:14:33 +0200
References: <k4oq5o$vbr$1@dont-email.me>
X-Newsreader: iForth 2.0 console (October 21, 2006)
Lines: 28
Organization: Wanadoo
NNTP-Posting-Date: 06 Oct 2012 09:14:02 GMT
NNTP-Posting-Host: s529d937f.adsl.online.nl
X-Trace: 1349514842 dr4.euro.net 9208 82.157.147.127:58855
X-Complaints-To: abuse@wanadoo.nl
Xref: csiph.com comp.lang.forth:15969

Mikael Nordman <oh2aun@invalid.com> writes Re: GA144 polyForth

> On 28.9.2012 0:16, Howerd wrote:
> > I also ran a speed test :
> > : asd  1000 for 1000 for 0 drop next next ;
> > takes about 3 seconds.
> > All in all it could compete with an MSP430, 8051 or a PIC except that 
> > only with the GA144 do you get so many fast cores with fast I/O to > 
> > play with.

> Out of curiosity I ran the same speed test with FlashForth on PIC18, 
> PIC24 and AVR Atmega.

> 181 ms on PIC24 @ 54 MHz
> 750 ms on PIC18 @ 48 MHz
> 4154 ms on Atmega 328 @ 8 Mhz

That's remarkably good. On a 2.667 GHz i7 desktop machine, iForth (64bit)
takes 2.22 ms. Adjusting for clockspeed, 2.22 ms * 2667 / 54 = 110 ms, as 
compared to PIC24's 181 ms.

The speed difference between PIC24 and PIC18 is much higher than the 
expected factor of 2. I guess FlashForth for PIC24 is radically different 
and uses more of the available registers?

-marcel