Vector type LOAD/STORE with post-increment.

I am trying to implement vector type load/store with post-increment for an out of tree backend.
I see that that ARM NEON support such load/store so I am using ARM NEON as an example of what to do.

The problem is I can’t get any C or C++ code example to actually generate vector load/store with post increment.

I am talking about something like this:
vldr d16, [sp, #8]

Does anybody know any C/C++ code example that will generate such code (especially loop)? Is this supported by the auto-vectorizer?

Thanks.

I am talking about something like this:
     vldr d16, [sp, #8]

Hi Francois,

This is just using the offset, not updating the register (see ARM ARM
A8.5). Post-increment only has meaning if you write-back the new value to
the register like:

  vldr d16, [sp], #8

Did you mean write-back? or just offset?

Does anybody know any C/C++ code example that will generate such code

(especially loop)? Is this supported by the auto-vectorizer?

It's not simple to go from a loop in C++ to an instruction in machine code,
especially when you're considering the vectorizer. Today you can generate a
post-indexing load, tomorrow a pre-indexing load, and the next day a simple
offset. All depending on how the IR is constructed, changed and lowered,
all of which change daily.

The quickest and surest way to generate NEON instructions is with NEON
intrinsics, but even so, LLVM is allowed to twist your code to generate
better instructions than you have thought possible. You can try to create
an IR that can generate post-indexed VLDRs on ARM, but that will not
guarantee it'll generate the same on any other backend.

A code that will generate post-indexed loads in ARM, might generate a
completely different instruction on Intel, or on your backend. What you
have to do is to understand the patterns that vectorized code has in IR,
possibly copy what's been done by other backends to lower such IR, and make
sure your backend uses the post-indexed load for the cases you care about.

Makes sense?

cheers,
--renato

I am talking about something like this:
     vldr d16, [sp, #8]

Hi Francois,

This is just using the offset, not updating the register (see ARM ARM
A8.5). Post-increment only has meaning if you write-back the new value to
the register like:

  vldr d16, [sp], #8

Did you mean write-back? or just offset?

yes I mean just offset like :
vldr d16, [sp], #8

Does anybody know any C/C++ code example that will generate such code

(especially loop)? Is this supported by the auto-vectorizer?

It's not simple to go from a loop in C++ to an instruction in machine
code, especially when you're considering the vectorizer. Today you can
generate a post-indexing load, tomorrow a pre-indexing load, and the next
day a simple offset. All depending on how the IR is constructed, changed
and lowered, all of which change daily.

The quickest and surest way to generate NEON instructions is with NEON
intrinsics, but even so, LLVM is allowed to twist your code to generate
better instructions than you have thought possible. You can try to create
an IR that can generate post-indexed VLDRs on ARM, but that will not
guarantee it'll generate the same on any other backend.

A code that will generate post-indexed loads in ARM, might generate a
completely different instruction on Intel, or on your backend. What you
have to do is to understand the patterns that vectorized code has in IR,
possibly copy what's been done by other backends to lower such IR, and make
sure your backend uses the post-indexed load for the cases you care about.

Makes sense?

yes... makes sense.. it is just not trivial to test if the post increment
will work when generated via C/C++ code. I was just wondering if there was
some simple C/C++ code snippets that would normally generate such post
increment load for vector. Apparently not.

I am trying to implement vector type load/store with post-increment
for an out of tree backend.
I see that that ARM NEON support such load/store so I am using ARM
NEON as an example of what to do.

The problem is I can't get any C or C++ code example to actually
generate vector load/store with post increment.

I am talking about something like this:
vldr d16, [sp, #8]

Does anybody know any C/C++ code example that will generate such code
(especially loop)? Is this supported by the auto-vectorizer?

To add slightly to Renato's answer...

Auto-vectorization, which is an IR-level pass, happens well before Pre/Post-increment formation (which happens during DAGCombine). Whether or not vector loads/stores are eligible for post-increment formation will depend on calling setIndexedLoadAction(ISD::POST_INC, ...) and how you've implemented <Target>TargetLowering::getPreIndexedAddressParts.

-Hal