llvm dev - Sep 2010 - [LLVMdev] Is it possible to have multiple stacks in the backend?

Hi all,

I'm developing a backend for a custom architecture.
The target architecture has different memory for different data types(vector
and scalar).
So is it possible to maintain two different stacks, one for vector and the
other for scalar?

----
Regards,

Dongrui She
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20100920/8332f22e/attachment.html>

On Sep 20, 2010, at 8:09 AM, Dongrui She wrote:
> Hi all,
> 
> I'm developing a backend for a custom architecture. 
> The target architecture has different memory for different data
types(vector and scalar).
> So is it possible to maintain two different stacks, one for vector and the
other for scalar?
> 
I don't see why not. The target independent bits that interface with that
sort of code should have sufficient target hooks in place to be able to handle
any weirdness that results. In particular, I'm thinking of the
Prologue/Epilogue inserter and eliminateFrameIndex() stuff. I don't know of
any targets that are doing this sort of thing currently, though, so it's
very possible there are unintended assumptions here and there that would need to
be fixed.

-Jim