This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFC: Handling of libgcc symbols in SH shared libraries


Hi there,

  I sent this patch to the mailing list. But mail was blocked due to improper
content format. I'm resending the patch to the list.

> That's a non sequitur.  To write an optimized assembler function, you have
> to consider the different algorithms available, and for all the likely
> candidates, see how the needed computations and data flow requirements
> constrain the scheduling of the 'hot' paths through the code, and do your
> instruction selection and scheduling around these constraints.
> When you have enough information to definitely say that one version is
> inferior, you can drop it.

Well, ours is one way to emulate [fd]p-bit.c, with some advantages and disadvantages.
You can reuse the code wherever you need. You can evaluate the approaches
used and submit the better one.

> Then what is the meaning of the blank lines interspersed with the instructions?

We have paired the instructions which are issued at one cycle separated by
a blank line. There are some exceptions with TEST and BRANCH instructions.

Best Regards,
Rakesh Kumar

Attachment: double-paranoia-results
Description: double-paranoia-results

Attachment: patch-fp.txt
Description: patch-fp.txt

Attachment: single-paranoia-results
Description: single-paranoia-results


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]