[all-commits] [llvm/llvm-project] 39db57: [LV][ARM] Inloop reduction cost modelling

[David Green via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 39db5753f993abcc4289dd165e8297a4e28f4b0a
      https://github.com/llvm/llvm-project/commit/39db5753f993abcc4289dd165e8297a4e28f4b0a
  Author: David Green <david.green at arm.com>
  Date:   2021-01-21 (Thu, 21 Jan 2021)

  Changed paths:
    M llvm/include/llvm/Analysis/TargetTransformInfo.h
    M llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
    M llvm/include/llvm/CodeGen/BasicTTIImpl.h
    M llvm/lib/Analysis/TargetTransformInfo.cpp
    M llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
    M llvm/lib/Target/ARM/ARMTargetTransformInfo.h
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    M llvm/test/Transforms/LoopVectorize/ARM/mve-reduction-types.ll
    M llvm/test/Transforms/LoopVectorize/ARM/mve-reductions.ll

  Log Message:
  -----------
  [LV][ARM] Inloop reduction cost modelling

This adds cost modelling for the inloop vectorization added in
745bf6cf4471. Up until now they have been modelled as the original
underlying instruction, usually an add. This happens to works OK for MVE
with instructions that are reducing into the same type as they are
working on. But MVE's instructions can perform the equivalent of an
extended MLA as a single instruction:

  %sa = sext <16 x i8> A to <16 x i32>
  %sb = sext <16 x i8> B to <16 x i32>
  %m = mul <16 x i32> %sa, %sb
  %r = vecreduce.add(%m)
  ->
  R = VMLADAV A, B

There are other instructions for performing add reductions of
v4i32/v8i16/v16i8 into i32 (VADDV), for doing the same with v4i32->i64
(VADDLV) and for performing a v4i32/v8i16 MLA into an i64 (VMLALDAV).
The i64 are particularly interesting as there are no native i64 add/mul
instructions, leading to the i64 add and mul naturally getting very
high costs.

Also worth mentioning, under NEON there is the concept of a sdot/udot
instruction which performs a partial reduction from a v16i8 to a v4i32.
They extend and mul/sum the first four elements from the inputs into the
first element of the output, repeating for each of the four output
lanes. They could possibly be represented in the same way as above in
llvm, so long as a vecreduce.add could perform a partial reduction. The
vectorizer would then produce a combination of in and outer loop
reductions to efficiently use the sdot and udot instructions. Although
this patch does not do that yet, it does suggest that separating the
input reduction type from the produced result type is a useful concept
to model. It also shows that a MLA reduction as a single instruction is
fairly common.

This patch attempt to improve the costmodelling of in-loop reductions
by:
 - Adding some pattern matching in the loop vectorizer cost model to
   match extended reduction patterns that are optionally extended and/or
   MLA patterns. This marks the cost of the reduction instruction correctly
   and the sext/zext/mul leading up to it as free, which is otherwise
   difficult to tell and may get a very high cost. (In the long run this
   can hopefully be replaced by vplan producing a single node and costing
   it correctly, but that is not yet something that vplan can do).
 - getExtendedAddReductionCost is added to query the cost of these
   extended reduction patterns.
 - Expanded the ARM costs to account for these expanded sizes, which is a
   fairly simple change in itself.
 - Some minor alterations to allow inloop reduction larger than the highest
   vector width and i64 MVE reductions.
 - An extra InLoopReductionImmediateChains map was added to the vectorizer
   for it to efficiently detect which instructions are reductions in the
   cost model.
 - The tests have some updates to show what I believe is optimal
   vectorization and where we are now.

Put together this can greatly improve performance for reduction loop
under MVE.

Differential Revision: https://reviews.llvm.org/D93476