hi
just want to know if there are some intentions to add more optimizations for i386 arch
because in fpc releases they didn't speak much about optimization ,

the actual fpc code optimization is good (thanks to fpc team) iam just curious if we will see in the future fpc produce faster code like gcc or vcpp

thanks

FPC doesn't do architecture specific optimization, at least that's what Jonas told me a while ago. But the architecture specific code generator picks up best instruction for the optimized generic tree nodes. If you look at http://wiki.lazarus.freepascal.org/FPC_New_Features_Trunk, you'll see some more experimental optimizations have been added (they're still causing problems in certain cases last time I tried, though).

"leledumbo" wrote

FPC doesn't do architecture specific optimization, at least that's what

Jonas

told me a while ago.

That's not entirely true. For example on the instruction set level FPC can and indeed does for some platforms. See the work on the ARM optimizers for example.
Other platform specific optimizations might be possible with FPC's current architecture as well, but I've not enough insight into that topic yet. In the end it's a matter of wether there is someone who implements it.

Regards,
Sven

That's not entirely true. For example on the instruction set level FPC can

and indeed does for some platforms. See the work on the ARM optimizers for example.

That's good to hear :)