[Coco] megaread utility

[Mark Marlette]

Dave,
Thanks for the numbers, it is moving along!
I understand the DW serial interface bandwidth limitiations. :) It is GREAT for what it is though!!!
Regards, Mark Marlette
 http://www.cloud9tech.com
mark at cloud9tech.com
mark at gamecamaddict.com

      From: Dave Philipsen <dave at davebiz.com>
 To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com> 
 Sent: Sunday, March 26, 2017 8:16 PM
 Subject: Re: [Coco] megaread utility
   
I should be more precise in what I typed on the command line.  It was 
'date -t;megaread </r0@;date -t' and I preloaded megaread and date into 
memory first.

So with the ramdisk it takes about 10 seconds to do a 'megaread 8192' 
which yields about 1.25 seconds for a normal megaread (1024).  With my 
off-brand (Edge ProShot 16GB Class 10) SD card it takes about 30 seconds 
to do a 'megaread 8192' which yields about 3.75 seconds for a normal 
megaread (1024).  Just for the sake of comparison, I run Drivewire at 
115,200 bps and I don't even try the 'megaread 8192' because I don't 
want to wait that long.  It takes about 88 seconds to do a normal 
megaread (1024).  And that's certainly not to knock Drivewire because 
there's only so much data you can squeeze through a 115,200 bps pipe.  
The serial data rate is the bottleneck.


Dave


On 3/26/2017 7:23 PM, Mark Marlette wrote:
> Dave,
> How about from the SD card?
> Regards, Mark Marlette
>  http://www.cloud9tech.com
> mark at cloud9tech.com
> mark at gamecamaddict.com
>
>        From: Dave Philipsen <dave at davebiz.com>
>  To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com>
>  Sent: Sunday, March 26, 2017 7:11 PM
>  Subject: Re: [Coco] megaread utility
>    
> Ok, not a big deal.  Just curious. I was using 'date -t; megaread; date -t' to get an idea of how much time the megaread takes.  Well, my times were about one second but, of course, the precision that 'date -t' gives is only in one second increments so I thought I'd do ten times the amount of data and then divide the result by ten to get a fairly accurate approximation of how many seconds is required for a megaread.  Anyway, I suppose I got close enough using 8192 1K blocks.
>
> For the 25 MHz CoCo3FPGA reading from a ramdisk it turned out to be about 1.25 seconds for a standard megaread of 1024 1K blocks.
>
>
> Dave
>
>> On Mar 26, 2017, at 2:23 PM, Robert Gault <robert.gault at att.net> wrote:
>>
>> Dave Philipsen wrote:
>>> I haven't looked at the source for the NitrOS9 'megaread' utility but it seems that perhaps its
>>> option of specifying the number of 1K blocks to read is limited to a 5-digit number.  Can anyone
>>> confirm that?
>>>
>>> Dave
>> In the comments,
>> *  01/02  2009/03/14  Bob Devries
>> * Added functionality to read a number of 1K blocks as specified on the command line.
>> * Command line is now: megaread #####
>> * where ##### is the number of 1K blocks to read; default 1024
>>
>> In the code,
>> start    clra
>>          clrb
>>          bsr  dec2bin    read a character from command line and convert to binary
>>          bsr  dec2bin
>>          bsr  dec2bin
>>          bsr  dec2bin
>>          bsr  dec2bin
>> that's five conversions.
>>
>> The code looks like it could be changed so that $FFFF 1K reads could be made. In that case, you would need six passes through dec2bin and you would need error checking to prevent numbers larger than 65535 being entered.
>>
>> Do you have a problem with a 5-digit limit?
>>
>> Robert
>>
>>
>> -- 
>> Coco mailing list
>> Coco at maltedmedia.com
>> https://pairlist5.pair.net/mailman/listinfo/coco
>


-- 
Coco mailing list
Coco at maltedmedia.com
https://pairlist5.pair.net/mailman/listinfo/coco