 |
 |
| View previous topic :: View next topic |
| Author |
Message |
Hans Wedemeyer
Joined: 15 Sep 2003
Posts: 226
|
| Version 3.19 compared to V3.201 |
 Posted: Wed May 26, 2004 9:55 pm |
 |
|
If V3.201 can optimize the same source and reduce the output by about 4000 bytes! that would be great... if the code still works. !
I think something is wrong somewhere.... !
I never noticed the math error in LST file for V3.190 until tonight.
LST file for same source file:
CCS PCH C Compiler, Version 3.190, 21503
ROM used: 26168 (82%)
Largest free fragment is 26168
RAM used: 1253 (82%) at main() level
1338 (88%) worst case
Stack: 7 worst case (6 in main + 1 for interrupts)
CCS PCH C Compiler, Version 3.201, 21503
ROM used: 22118 bytes (69%)
Largest free fragment is 9814
RAM used: 1253 (82%) at main() level
1338 (88%) worst case
Stack: 7 worst case (6 in main + 1 for interrupts) |
|
 |
C-H Wu
Guest
|
| expect a 3.202 on Friday |
| Posted: Thu May 27, 2004 2:35 am |
|
|
A code size reduction of 12% with #opt10 since 3.187 is typical.
However, read this:
3.201 New infrastructure added in 3.200 to support new chips and families has
3.201 caused a large number of problems, The reported problems are fixed in 3.201
3.201 Use with caution, expect a 3.202 on Friday. 3.191 is stable.
with great caution, I just download 3.191 again and found that the delay_us() is still there !
3.187 is better for me because it does not the BSR setting bug in 3.188 ~ 3.200, nor the sint32 relational comparison bug in 3.190, nor the delay_us() bug in 3.191 ~ 3.200.
I am not going to beta test the 3.201 or 3.202 on weekend. 
Best wishes |
|
|
Guest
|
| Re: expect a 3.202 on Friday |
| Posted: Thu May 27, 2004 6:55 am |
|
|
| C-H Wu wrote: | A code size reduction of 12% with #opt10 since 3.187 is typical.
Best wishes |
Yes I read that warning and only downloaded just to see what it did with the bugs I had reported earlier.
That fact that CCS even issued a warning is a step in the right direction !
I never switch on optimization unless I'm running into a code size problem. So about 4000 byte redution with no Opt. is a lot !
Hans W |
|
|
C-H Wu
Guest
|
| Re: expect a 3.202 on Friday |
| Posted: Thu May 27, 2004 7:16 am |
|
|
| Anonymous wrote: | | C-H Wu wrote: | A code size reduction of 12% with #opt10 since 3.187 is typical.
Best wishes |
Yes I read that warning and only downloaded just to see what it did with the bugs I had reported earlier.
That fact that CCS even issued a warning is a step in the right direction !
I never switch on optimization unless I'm running into a code size problem. So about 4000 byte redution with no Opt. is a lot !
Hans W |
>> with no Opt.
no no, the default optimization level for PCW is 9, i.e. same as #opt 9, which is the same as #opt 10 for 3.188 ~ 3.191. If you really want safety, you should set it back to #opt 5 or #opt 7, which will bring you back to the old days.
C-H Wu |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
