Index
> Main > movcc |
| Author |
|
|
bitshifter 28 Sep 2010, 08:14
I would use it, and also define a minimum processor spec in the readme.txt file.
|
|||
28 Sep 2010, 08:14 |
|
|
b1528932 28 Sep 2010, 08:19
no way.
i remember it has a cpuid code to determine if its supported. |
|||
|
28 Sep 2010, 08:19 |
|
|
revolution 28 Sep 2010, 11:19
I use a macro to simulate it in fasmarm. For most practical situations the user will never know the difference.
[edit] I guess it would be helpful to post the macro: Code: irp cc,a,ae,b,be,c,e,g,ge,l,le,na,nae,nb,nbe,nc,ne,ng,nge,nl,nle,no,np,ns,nz,o,p,pe,po,s,z { local i macro cmov#cc [args] \{ \common \local .x j#cc .x load i from $-2 store (i xor 1) at $-2 mov args .x: \} } |
|||
|
28 Sep 2010, 11:19 |
|
|
b1528932 28 Sep 2010, 12:43
jcc will generate exception even if condition is not met.
so this macroand this macro wont. there is a diffrence. |
|||
|
28 Sep 2010, 12:43 |
|
|
revolution 28 Sep 2010, 12:56
b1528932 wrote: jcc will generate exception even if condition is not met. [edit] Oh, I just realised what you mean. A bad memory address will raise an exception. Yeah, okay, it will. So I guess it could fail for something like stack touching routines. |
|||
|
28 Sep 2010, 12:56 |
|
|
ouadji 28 Sep 2010, 16:54
If I understand correctly, even today, it's better not to use this instruction If one wishes that a program can run on most PCs. So, "in practice", this instruction is unusable. _________________ I am not young enough to know everything (Oscar Wilde)-  |
|||
|
28 Sep 2010, 16:54 |
|
|
Fanael 28 Sep 2010, 17:27
Quite frankly, have anyone heard of anybody running modern operating system on x86 CPU that does not support conditional moves (i.e. early PPros and older processors)?
|
|||
|
28 Sep 2010, 17:27 |
|
|
bitRAKE 28 Sep 2010, 17:38
Terje Mathisen wrote: CMOVcc was especially designed for those few short branches which really cannot be predicted! |
|||
|
28 Sep 2010, 17:38 |
|
< Last Thread | Next Thread > |
Forum Rules:
|