[Coco] CoCoSDC performance issue - can someone try this from OS-9?

[Nick Marentes]

On 1/02/2015 5:41 PM, Allen Huffman wrote:
>> On Feb 1, 2015, at 1:34 AM, Nick Marentes <nickma2 at optusnet.com.au> wrote:
>>
>> The SDC reads a FAT32 file system on the SD card. I wonder if this is creating the problem with a loaded SD card?
> That was what I originally wondered about — if the huge 128MB files were fragmented, causing extra seeking around. I expect the low-level FAT firmware isn’t as elegant and speedy as what runs on our desktops.
>
> When I copied all my files off, then reformatted and moved them back, I would expect zero fragmentation of my .DSK files. I have 1.9GB out of 8GB available, so that may not have anything to do with it.
>
> At this point, I do not think it is CoCoSDC. I think it was me not realizing my 128MB drive only had 1-2MB of free space, so RBF is probably having to climb through a bunch of stuff to find spots for even tiny files. Maybe they are only at the end, and it has to read in 64K of bitmap every time looking for the empty bits at the end of the disk.
>
> Nick: There is up to 64K set aside as a bitmap marking which sectors are free. Each bit represents a sector, 0=free/1=used. Thus, if I want to write a tiny file that takes 256 bytes, it opens the bit map and has to parse through it looking for bytes with 0s in it to use.
>
> 		— A
>

You realize that defrag does little speed-wise on any memory based drive?

There is no head seek on these, they just access the next block 
instantly just like normal memory.

Nick