[Prev][Next][Index][Thread]

forwarded message from Thomas Davis



Tom asked me to forward this to the list.  Some of you may get two
copies of this -- my apologies.

------- start of forwarded message (RFC 934 encapsulation) -------
From: Thomas Davis <Thomas.Davis@mnscorp.com>
To: nelson@crynwr.com (Russell Nelson)
Subject: Re: qcam on NetBSD
Date: Fri, 12 Jan 1996 10:20:03 -0600 (CST)

Russ Nelson writes:
> 
> Hell, I don't care if this becomes a connectix development list.  The
> more people we have working on it, the better.

Fine with me.  Just no OS wars, ok?  

> Now hmmm... can
> anybody think of a portable way to unroll the loops in qc_scan?  I
> don't think an unrolled C implementation is going to be fast enough
> because of the need to keep the I/O port in the DX register where IN
> and OUT need it.  I think it'll be portable "enough" if we use ``as''
> assembler source, because the DOS and OS/2 people can compile it using
> their respective gcc ports.
>

Then my Alpha/AXP can't run the bloody thing then..  :-)

Keep both, one for non-intel stuff, one for everything else.

I think the code for the bi-directional is efficient enough.  In testing
my new (and unreleased) kernel driver, I can get 24fps out of the bloody
thing (80x60, 6bpp) 17fps at 240x120, 6bpp, and 4fps at 320x240, 6bpp!

Where the code gets slow is the decoding of the 6bpp/uni-directional mode
and the 4bpp/bi-directional mode.  The nybble packing just kills things..
(but makes for a fast transfer..)

I hope to have a something this weekend.  (ioctls, 4bpp bidirectional mode,
and uni-directional modes are left to go)

- -- 
Thomas Davis			| Internet:	Thomas.Davis@mnscorp.com
Systems Consultant		| Snail Mail:	Suite 528
Midwest Network Solutions Corp.	| 		1941 South 42nd Street
(402) 346-7687			|		Omaha, NE 68105-2939
------- end -------