China DOS Union

Try DJGPP....
After going through many difficulties, I finally successfully compiled TOdisk using DJGPP... It runs normally
This is the initial version of TOdisk32
(Requires more than 50MB of memory)
Welcome to play...

Oh, finally got it done. It must require more than 50MB of memory. Did you use a fixed-size array? It's recommended to use pointers + the malloc function. At least if there's not enough memory, you can give a prompt. Keep improving. This tool is very convenient.

Just had a successful experiment. It's the initial version, so there's still room for improvement.

With such a large memory, arrays definitely won't be used. We'll use pointers + malloc.

Currently, I'm trying to find ways to improve the speed. Now I want to figure out the reasons for GHOST's ultra-fast speed. Is it because of super-fast compression, or the use of UDMA, or using XMS with assembly in real mode?

Hope those who know can share.

(I downloaded the source code of GHOST FOR LINUX, but didn't find relevant functions. There are many libraries, but I don't know if they are there? And I don't know how to call them? The documentation doesn't seem to have it.. I saw a source code dd_rescue, but it doesn't seem to have compression function..)

[ Last edited by GOTOmsdos on 2007-2-28 at 07:22 AM ]

Tried it out, and so far no issues have been found

Support

Download and have a look

Support original creation. In the long run, GOTO has boldly taken the first step to realize the localization of the core part of backup software. If the version is mature, and then using Easy Language as the shell, it will completely be our own independent intellectual property rights in China.

Originally posted by GOTOmsdos at 2007-2-28 07:21 AM:
Currently, I want to find ways to improve speed, and now I want to figure out the reasons for the super-fast speed of GHOST. Is it due to super-fast compression, or the use of UDMA, or using XMS in real mode with assembly?

1) Ghost processes files as units, extracting only useful data

2) Ghost has its own read/write program (not dependent on the operating system)

3) Low compression (if the parameter -z9 is not forced, the strongest in the image screen is -z3)

4) Maybe the Block Transfer method mentioned by Jack Ellis, at least transferring data in blocks of 16K each. Many I/O subroutines attached to C in the market are not efficient enough

ghost is developed with DJGPP and compressed with UPX!

Is this the latest version?

Something similar to ghost? Let's use it.

Learning

Come and take a look, thanks to GOTOmsdos!

Great stuff, thanks for sharing.