Re: Fast copy routine needed

From: silverdr_at_wfmh.org.pl
Date: Tue, 25 Aug 2015 00:50:28 +0200
Message-ID: <55DB9FB4.7040002@wfmh.org.pl>
On 2015-08-25 00:43, Justin wrote:

>>> On 2015-08-24, at 22:32, Justin <shadow@darksideresearch.com>
>>> wrote:
>>> 
>>> I have often wondered, but have never pursued, the possibility of
>>> compressing data on the processor in say a 1541 and then
>>> decompressing it on the 64 to get some more throughput.
>> 
>> That's what we've been doing for a long time. Only not "on-the-fly"
>> ;-) And when I started working with DolphinDOS, I started to regret
>> all the meticulous compressing I did to many of my files earlier...
>> :-)

> Yeah I should have been more clear, I meant on the fly. That was why
> I had looked at LZ4 and gotten interested when I saw that 6502
> assembler implementation. I promptly forgot about it, but this thread
> reminded me. Neither the behavior of the 1541 nor the rolling of a
> kernal to do that is really within my domain, so it was just
> curiosity on my part, not something I intended to try to do.

The problem is that no reasonable compression (only a kind of suboptimal
"charpacking", maybe) can be done in real time on the CPUs in question.

-- 
SD!

       Message was sent through the cbm-hackers mailing list
Received on 2015-08-24 23:00:06

Archive generated by hypermail 2.2.0.