Previously, every expression/statement type was responsible for
pruning or clearing the store-cache when branches were involved. This
was difficult to reason about and easy to get wrong, particularly if
the details are not fresh in your mind.
Now, `writeLabel` takes care of the details for you. Pass in the
location of the branch(es) which use the label, and the proper cache
updating behavior will occur automatically.
Some of the label enum types are not strictly necessary and exist for
the benefit of a reader. Specifically:
- `kBranchlessBlock` and `kBranchIsOnPreviousLine` are synonyms
- `kBranchIsBelow` and `kBranchesOnBothSides` are also synonyms
The hope is that extra enum names will be easier for a reader to
follow, versus fewer but very-verbose enum names (like
`kBranchIsBelowOrOnBothSides`).
This change earned some very minor switch-related dividends. Previously,
every label in a switch was treated as a forward-branch, but in fact,
the very first label in a switch is privileged. This is because we are
branching from the previous line, and the store cache is trustworthy in
this case. (Versus "branching from above," where the store cache needs
to be pruned before it can be trusted.)
Change-Id: I38b539069c22be9f0777b632f60f0eab2409d687
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/531540
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
We now have two functions `writeOpLoad` and `writeOpStore` which are
in charge of writing SpvOpLoad and SpvOpStore instructions.
`writeOpStore` also keeps track of pointer stores in a "store cache."
Subsequent loads from that same pointer will be found in the cache and
will return the value stored in that pointer instead.
Such a cache definitely cannot work in the face of control flow, so we
make the following concessions:
- `pruneReachableOps` is now `pruneConditionalOps`. Any pointers that
are altered inside a potentially-unreachable block are cleared from
the cache entirely.
- The entire store cache is cleared at all OpLabels within a loop.
The cache also cannot work in the presence of swizzled stores, so we
make another significant concession:
- The entire store cache is cleared whenever we store into a non-memory
pointer (e.g., assigning into a swizzled LValue, such as `foo.xz`).
Despite these significant limitations, this manages to dramatically
shrink many real-world examples.
Change-Id: I0981a0cf7b45b064e153e9ada271494c8e00cad5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/530054
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Previously, we stringized the types and put them into fTypeMap. Using
the op cache is a simpler mechanism that should work equally well.
Output diffs are almost all ID reorderings. In a few cases we
managed to deduplicate function types that stringize differently but
come out the same in SPIR-V (e.g. no float/half distinction).
Change-Id: If7de5b2dafa12d05c3c2c497a243e9e3908dfee7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529805
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
If we start with an OpConstantComposite, then we do an
OpCompositeExtract from it, we can look up the result directly from op
cache and avoid doing any work. This helps our matrix code a lot.
Change-Id: Idfbdc0c69676b9c1e91cdc57bf0d6382b9b5d8d4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529339
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Previously, we didn't usually generate OpConstantComposite ops for
matrices. Now, a matrix assembled from constants should come out as a
constant.
Change-Id: I458718901686dffb84e4079a81017d61195420d3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529338
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
The output changes here are almost entirely a wash, because we already
had support for caching scalars and vectors. Almost all changes are just
inconsequential reorderings of IDs, and the removal of RelaxedPrecision
decorators on constants (which were not meaningful).
Change-Id: I45340c4a240cb504b7c4a934b3db178d2f39ec99
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528709
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Some GPUs (Adrenos in particular) perform noticeably better when we
use OpConstantComposite instead of OpCompositeConstruct. This also gives
us some deduplication of redundant ops.
Change-Id: I53b7a3e1cf61e51647a661a08ff4c7b53ee60f10
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528636
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Adding tests for matrix math and comparison
bug: skia:12681
Change-Id: Ia1537ee2e411383749456fd6ff938b7c9a2e1061
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/493416
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Julia Lavrova <jlavrova@google.com>
