Indeed, it is!srod wrote:Now that's a different issue completely!
I just couldn't resist the chance to add this angle...
Out of cpu registers!
-
dell_jockey
- Enthusiast
- Posts: 767
- Joined: Sat Jan 24, 2004 6:56 pm
Thing is, whilst you are correct and that my given example should only compile to a single MOV instruction, you'd have to say that the original programmer would be at fault here (i.e. me) for conceiving such an appalling piece of code.
I'd say that it would be my responsibility to optimise this particular piece of code.
It's similar to:
only a real banana head would code the above and so, the question is, should we really expect the compiler to look out for such redundancies?
One thing I love about PB is the speed of compilation. Try an optimising c++ compiler - God it takes forever to compile anything!
It's a fine line I reckon.
I'd say that it would be my responsibility to optimise this particular piece of code.
It's similar to:
Code: Select all
x = 10
x = 20One thing I love about PB is the speed of compilation. Try an optimising c++ compiler - God it takes forever to compile anything!
It's a fine line I reckon.
I may look like a mule, but I'm not a complete ass.
And if the optimiser compiler ever gets it wrong, tracking down the bugs can be a nightmare.
With PB I can get the commented ASM and quickly diagnose even compiler errors.
Again, PB keeps it simple. If you want to have the value evaluated at compile time, you just replace x, y, z with #x, #y, #z.
Again, I see this as a benefit of PB. It makes things easy for me, but never takes control away from me.
Thats how I like it, thats why I use PB.
With PB I can get the commented ASM and quickly diagnose even compiler errors.
Again, PB keeps it simple. If you want to have the value evaluated at compile time, you just replace x, y, z with #x, #y, #z.
Again, I see this as a benefit of PB. It makes things easy for me, but never takes control away from me.
Thats how I like it, thats why I use PB.