| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
|
|
|
|
|
| |
This specializes the fields of structs based on the types written to them. That is,
if a field is of type A but in practice we always write some subtype B to it
then we can change the type of the field to that.
On j2wasm this manages to improve at least one field in 2% of types. Not a
large amount, but this does lead to further benefits in later opts (e.g. about a third
of the improvements are to turn a field non-nullable).
|
| |
|
|
|
|
|
|
|
| |
Just as the --nominal flag forces all types to be parsed as nominal, the
--structural flag forces all types to be parsed as equirecursive. This is the
current default behavior, but a future PR will change the default to parse types
as either structural or nominal according to their syntax or encoding. This new
flag will then be necessary to get the current behavior.
Also take this opportunity to deduplicate more flags in the help tests.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a new pass to perform global type optimization. So far this just
does one thing, to find fields with no struct.set and to turn them
immutable (where possible - sub and supertypes must agree).
To do that, this adds a GlobalTypeRewriter utility which rewrites
all the heap types in the module, allowing changes while doing so.
In this PR, the change is to flip the mutable field. Otherwise, the
utility handles all the boilerplate of creating temp heap types using
a TypeBuilder, and it handles replacing the types in every place
they are used in the module.
This is not enabled by default yet as I don't see enough of a benefit
on j2cl. This PR is basically the simplest thing to do in the space of
global type optimization, and the simplest way I can think of to
fully test the GlobalTypeRewriter (which can't be done as a unit
test, really, since we want to emit a full module and validate it etc.).
This PR builds the foundation for more complicated things like
removing unused fields, subtyping fields, and more.
|
| |
|
|
|
| |
Locally I saw a 10% speedup on j2cl but reports of regressions have
arrived, so let's disable it for now pending investigation. The option added
here should make it easy to experiment.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Previously the set of functions to keep was initially empty, then the profile
added new functions to keep, then the --keep-funcs functions were added, then
the --split-funcs functions were removed. This method of composing these
different options was arbitrary and not necessarily intuitive, and it prevented
reasonable workflows from working. For example, providing only a --split-funcs
list would result in all functions being split out not matter which functions
were listed.
To make the behavior of these options, and --split-funcs in particular, more
intuitive, disallow mixing them and when --split-funcs is used, split out only
the listed functions.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
An "intrinsic" is modeled as a call to an import. We could also add new
IR things for them, but that would take more work and lead to less
clear errors in other tools if they try to read a binary using such a
nonstandard extension.
A first intrinsic is added here, call.without.effects This is basically the same
as call_ref except that the optimizer is free to assume the call has no
side effects. Consequently, if the result is not used then it can be optimized
out (as even if it is not used then side effects could have kept it around).
Likewise, the lack of side effects allows more reordering and other
things.
A lowering pass for intrinsics is provided. Rather than automatically
lower them to normal wasm at the end of optimizations, the user must
call that pass explicitly. A typical workflow might be
-O --intrinsic-lowering -O
That optimizes with the intrinsic present - perhaps removing calls
thanks to it - then lowers it into normal wasm - it turns into a call_ref -
and then optimizes further, which would turns the call_ref into a
direct call, potentially inline, etc.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To avoid requiring a static memory allocation, wasm-split's instrumentation
defaults to recording profile data in Wasm globals. This causes problems for
multithreaded applications because the globals are thread-local, but it is not
always feasible to arrange for a separate profile to be dumped on each thread.
To simplify the profiling of such multithreaded applications, add a new
instrumentation mode that stores the profiling data in shared memory instead of
in globals. This allows a single profile to be written that correctly reflects
the called functions on all threads.
This new mode is not on by default because it requires users to ensure that the
program will not trample the in-memory profiling data. The data is stored
beginning at address zero and occupies one byte per declared function in the
instrumented module. Emscripten can be told to leave this memory free using the
GLOBAL_BASE option.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some functions run only once with this pattern:
function foo() {
if (foo$ran) return;
foo$ran = 1;
...
}
If that global is not ever set to 0, then the function's payload (after the
initial if and return) will never execute more than once. That means we
can optimize away dominated calls:
foo();
foo(); // we can remove this
To do this, we find which globals are "once", which means they can
fit in that pattern, as they are never set to 0. If a function looks like the
above pattern, and it's global is "once", then the function is "once" as
well, and we can perform this optimization.
This removes over 8% of static calls in j2cl.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The goal of this mode is to remove obviously-unneeded code like
(drop
(i32.load
(local.get $x)))
In general we can't remove it, as the load might trap - we'd be removing
a side effect. This is fairly rare in general, but actually becomes quite
annoying with wasm GC code where such patterns are more common,
and we really need to remove them.
Historically the IgnoreImplicitTraps option was meant to help here. However,
in practice it did not quite work well enough for most production code, as
mentioned e.g. in #3934 . TrapsNeverHappen mode is an attempt to fix that,
based on feedback from @askeksa in that issue, and also I believe this
implements an idea that @fitzgen mentioned a while ago (sorry, I can't
remember where exactly...). So I'm hopeful this will be generally useful
and not just for GC.
The idea in TrapsNeverHappen mode is that traps are assumed to not
actually happen at runtime. That is, if there is a trap in the code, it will
not be reached, or if it is reached then it will not trap. For example, an
(unreachable) would be assumed to never be reached, which means
that the optimizer can remove it and any code that executes right before
it:
(if
(..condition..)
(block
(..code that can be removed, if it does not branch out..)
(..code that can be removed, if it does not branch out..)
(..code that can be removed, if it does not branch out..)
(unreachable)))
And something like a load from memory is assumed to not trap, etc.,
which in particular would let us remove that dropped load from earlier.
This mode should be usable in production builds with assertions
disabled, if traps are seen as failing assertions. That might not be true
of all release builds (maybe some use traps for other purposes), but
hopefully in some. That is, if traps are like assertions, then enabling
this new mode would be like disabling assertions in release builds
and living with the fact that if an assertion would have been hit then
that is "undefined behavior" and the optimizer might have removed
the trap or done something weird.
TrapsNeverHappen (TNH) is different from IgnoreImplicitTraps (IIT).
The old IIT mode would just ignore traps when computing effects.
That is a simple model, but a problem happens with a trap behind
a condition, like this:
if (x != 0) foo(1 / x);
We won't trap on integer division by zero here only because of the
guarding if. In IIT, we'd compute no side effects on 1 / x, and then
we might end up moving it around, depending on other code in
the area, and potentially out of the if - which would make it happen
unconditionally, which would break.
TNH avoids that problem because it does not simply ignore traps.
Instead, there is a new hasUnremovableSideEffects() method
that must be opted-in by passes. That checks if there are no side
effects, or if there are, if we can remove them - and we know we can
remove a trap if we are running under TrapsNeverHappen mode,
as the trap won't happen by assumption. A pass must only use that
method where it is safe, that is, where it would either remove the
side effect (in which case, no problem), or if not, that it at least does
not move it around (avoiding the above problem with IIT).
This PR does not implement all optimizations possible with
TNH, just a small initial set of things to get started. It is already
useful on wasm GC code, including being as good as IIT on removing
unnecessary casts in some cases, see the test suite updates here.
Also, a significant part of the 18% speedup measured in
#4052 (comment)
is due to my testing with this enabled, as otherwise the devirtualization
there leaves a lot of unneeded code.
|
| |
|
|
|
| |
Use ToolOptions there, which adds --nominal support.
We must also pass --nominal to the sub-commands we run.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A field in a struct is constant if we can see that in the entire program we
only ever write the same constant value to it. For example, imagine a
vtable type that we construct with the same funcrefs each time, then (if
we have no struct.sets, or if we did, and they had the same value), we
could replace a get with that constant value, since it cannot be anything
else:
(struct.new $T (i32.const 10) (rtt))
..no other conflicting values..
(struct.get $T 0) => (i32.const 10)
If the value is a function reference, then this may allow other passes
to turn what was a call_ref into a direct call and perhaps also get
inlined, effectively a form of devirtualization.
This only works in nominal typing, as we need to know the supertype
of each type. (It could work in theory in structural, but we'd need to do
hard work to find all the possible supertypes, and it would also
become far less effective.)
This deletes a trivial test for running -O on GC content. We have
many more tests for GC at this point, so that test is not needed, and
this PR also optimizes the code into something trivial and
uninteresting anyhow.
|
| |
|
|
|
| |
Add a new OptimizeForJS pass which contains rewriting rules specific to JavaScript.
LLVM usually lowers x != 0 && (x & (x - 1)) == 0 (isPowerOf2) to popcnt(x) == 1 which is ok for wasm and other targets but is quite expensive for JavaScript. In this PR we lower the popcnt pattern back to the isPowerOf2 pattern.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a local is say anyref, but all values assigned to it are something
more specific like funcref, then we can make the type of the local
more specific. In principle that might allow further optimizations, as
the local.gets of that local will have a more specific type that their
users can see, like this:
(import .. (func $get-funcref (result funcref)))
(func $foo (result i32)
(local $x anyref)
(local.set $x (call $get-funcref))
(ref.is_func (local.get $x))
)
=>
(func $foo (result i32)
(local $x funcref) ;; updated to a subtype of the original
(local.set $x (call $get-funcref))
(ref.is_func (local.get $x)) ;; this can now be optimized to "1"
)
A possible downside is that using more specific types may not end
up allowing optimizations but may end up increasing the size of
the binary (say, replacing lots of anyref with various specific
types that compress more poorly; also, for recursive types the LUB
may be a unique type appearing nowhere else in the wasm). We
should investigate the code size factors more later.
|
|
|
Add list tests for the help messages of all tools, factoring out common options
into shared tests. This is slightly brittle because the text wrapping depends on
the length of the longest option, but that brittleness should be worth the
benefit of being able to see the actual help text in the tests.
|
