As part of this, move functions that seem reasonable to make out-of-line
to a separate `_impl.h` header file that is only included where the
explicit instantiation _definition_ is provided.
By using explicit instantiation we can make these templates behave more
like non-template classes in terms of supporting out-of-line definitions
that don't need to be compiled by every translation unit. The set of
eventual instantiations here is fundamentally known, and there tend to
be headers that define a canonical "leaf" type where it makes sense to
trigger the explicit instantiation.
Where we already had a `.cpp` file to put the explicit instantiation
definition, use it. But in some places we didn't have such a `.cpp` file
so this PR adds those.
This also requires that we have precise constraints on APIs that _can't_
be instantiated for specific argument types, as now we don't do this
lazily.
Combined, this appears to reduce the sum of object file sizes in the
`check` directory by almost 40% (122mb -> 74mb) in my measurement.
My actual goal was to improve compile times, but so far I don't have a
great methodology for measuring these... But the object file size
reduction seems to confirm this is a net win and likely represents a
non-trivial improvement in compile time.
Assisted-by: Antigravity with Gemini
Allow any type that has a mapping from Carbon to C++ to be exposed to
C++ via name lookup. This also exposes the logic to export Carbon
classes to C++ to apply during type mapping, which gives very slight
support for passing Carbon types to C++ functions from Carbon, but not
really enough to sensibly test yet.
Depends on #7042.
Add `InstIs`, `GetInstAs`, and `TryGetInstAs` which act on the
underlying constant instruction in a constant value, to save an explicit
call to `GetInstId`.
```carbon
context.insts().GetAs<InstT>(context.constant_values().GetInstId(const_id))
```
can now be written as simply
```carbon
context.constant_values().GetInstAs<InstT>(const_id)
```
For future work, we might provide `GetInst()` so that
`context.insts().Get(context.constant_values().GetInstId(const_id)` can
be shortened also.
- Distinguish attached vs. unattached constants.
- Add some missing value stores to the top-level output.
- Add missing fields to various Print methods.
The IdTag knows the type of the Id its tagging and the type of the Id
being used as the tag. This prevents mixing up tagged and untagged ids,
and avoids having to work with untyped integers.
Adds an Untagged marker struct that's used as the tag type in IdTag when
no tag is desired.
The complexity of ConstantIds and TypeIds became a bit visible: TypeIds
are concrete ConstantIds. And ConstantIds have two different tagging
schemes, one for concrete and one for symbolic ids. And ConstantIds are
actually re-cast InstIds with the same index. The LoweredTypeStore needs
to work with tagged TypeIds, but the tags actually come from an InstId
store in ConstantValueStore. Now this is expressed in the type system by
getting the tags for TypeIds from the ConstantValueStore.
ValueStores without an TagId type parameter are now visibly untagged.
IdTag is now only default constructible when it does not have a tag,
which means ValueStore is only default constructible when the TagId is
untagged. This forces tagged value stores to be constructed correctly
with a tag at compile time, and untagged ones to be constructed without.
FixedSizeValueStore has overloads for dealing with tagged and untagged
Ids, since it can't default-construct ValueStore for tagged ids, and no
longer requires passing in default-constructed tags when there is no tag
in the ids.
Move `GetWithDefault` into the `ValueStore` base class, and avoid doing
the tag -> index mapping twice.
Call `ValueStore::Get` instead of `ConstantValueStore::GetAttached` in
`GetUnattachedConstant`. This is equivalent, since we never need a
default value here, and should be faster and less surprising.
Use the `CheckIRId` as a unique identifier for the scope of an `InstId`
- if an `InstId` is created within the scope of one `CheckIRId` it must
not be used in the scope of a different `CheckIRId`.
This is achieved without extra storage, but with false negatives for
large inputs.
When an `InstId` is created, the original index of the `Inst` is XORed
with a tag derived from the `CheckIRId` to produce the final `InstId`.
When the `InstId` is used, the expected tag is XORed with the `InstId`
to get back to the original index - if the tags don't match, the
resulting index will be corrupted, likely too large - resulting in an
out of bounds index CHECK-failure.
(the tag value is derived as such:
* take the CheckIRId
* left shift one bit (padding zero)
* left shift another bit (padding 1 - used to signify that the resulting
`InstId` has a tag combined into it)
* reverse the bits
In this way, the tag is unlikely to overlap with the index for small
test cases - making it possible to separate out the `CheckIRId` from the
index in these cases to provide more meaningful debugging/CHECK
messages, and more informative `SemIR` textual dumping that can now
include the `CheckIRId` along with the `Inst`'s index in the name of an
`inst`)
The test churn here is improved printing as tagged `InstId`s can now,
with best effort (more likely for small test cases where the `CheckIRId`
and the `Inst` index aren't at risk of overlapping from the high and low
bits), render the `CheckIRId` as part of the inst's name. Going from
`instNN` to `irMM.instNN`.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
This makes convert more consistent, it always makes a FacetAccessType
for a facet value, rather than only doing so after lookup returns. The
intention for this is that FacetAccessType will evaluate to
SymbolicBindingType in the future, so this will expose that constant
value to impl lookup instead of the original facet value, which will
avoid impl lookup having to deal with `.Self` or `BindSymbolicName`
specifically.
This changes ConstantValueStore to use ValueStore so that we get the
allocation flow that we're leaning towards there. This adds
`ConstantId::SymbolicId` to represent what was previously an int32
"symbolic_index" (although called an index, it's consistent with ID
usages).
I'm also changing Chunk::at/push to be consistent with the wrapping
Get/Add naming; the naming difference sticks out more with
UninitializedFill being added.
Instead of building an eval block as a separate pass at the end of a
generic, build the eval block incrementally.
The larger change here is that asking for the type or constant value of
an instruction now always returns an unattached type or constant value,
in order to preserve the behavior that we previously achieved by doing
the rewrite to attached types and constant values at the end of handling
the generic.
This also incidentally fixes some subtle issues where attached types and
constant values would leak out into check and cause it to get confused
about differences between attached and unattached values. Check should
no longer see attached values except where it explicitly asks for them.
---------
Co-authored-by: Dana Jansens <danakj@orodu.net>
When the parameter is a deduced symbolic FacetValue, refering to a
BindSymbolicName, and the argument is a concrete FacetValue that would
match the FacetType requirements on the BindSymbolicName's type, we
currently do not deduce that the argument matches the parameter.
The argument is not _converted_ to the parameter type because they are
both FacetValues of the same FacetType type. However they are also not
equal constant values so the argument is not saved as a deduced match
for the parameter.
In order to accept the FacetValue, we need to consider them as
`deduce_through`, which attempts to deduce each of the fields in the
argument FacetValue against the fields in the parameter FacetValue.
This deduces that the argument's concrete type matches the symbolic
BindSymbolicName and its witnesses are the same.
Since the parameter is a FacetValue, its argument is not the type that
needs to be recorded as the deduced type for the binding. The
BindSymbolicName inside the parameter is the place that we need to find
the deduced type for the binding. So simply walking into the FacetValue
gets us to that position, where we eventually record the deduced
argument type as being the concrete type from the original argument
FacetValue.
Similarly, when determining what interfaces are satisfied by a
FacetValue for deduce, we want to use the full type available in the
FacetValue rather than just those from its FacetType. Determining
availability of interfaces here is equivalent to converting, and we want
converting a FacetValue to always work on the full available type info.
Only API access (member lookup) is restricted by a FacetValue to the
interfaces provided by its FacetType type.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
If the query facet type has more than one interface, we must find an
impl that provides that interface for the query type for each interface.
This just looks like a for loop over the interfaces and ensuring we
found one impl witness for every one.
However the impl matching must change since it can't look at the
constant value of the entire query facet type for comparison with the
impl, as that query facet type may be for multiple interfaces and we are
looking to match an impl of a single interface.
To do this we break the query facet type up into each interface and make
sure the interface ids match. Then ensure that the impl was able to
deduce any generic parameters using the specific of the single query
interface.
There are some TODOs left here:
1. If the facet type for the query or the impl constraint has
"other_requirements" then we can't verify that they match since they are
lost. We fall back to comparing the constant id of the query to the
impl's constraint (after deducing generics in the impl). This correctly
eliminates mismatches but eagerly eliminates impls that could match the
query interface as well when there's more than one interface in the
query.
2. We don't return a witness for every interface in the query facet
type. Since we can't demonstrate any use of the witness there yet, for
cases that can have more than one interface in the query facet type,
this doesn't break anything that was previously working. The return
value is currently treated as a bool for cases with multiple interfaces
in the facet type (as a test for "can this be converted") but the
converted-to facet value's witnesses are unused.
---------
Co-authored-by: josh11b <15258583+josh11b@users.noreply.github.com>
This allows iterating on all values in a store along with the Id for
each value, instead of `llvm::enumerate(store.array_ref())` which would
give you the indices.
While the indices are really the same as the Ids, this provides a
typesafe way to enumerate() over a store.
There's no use for this right now, but I thought I needed this, and it
helped me debug, and it was a pain to write correctly without dangling
references.
---------
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Treat template bindings as introducing template phase, and propagate it
in the same way we propagate the checked generic phase.
Rename "symbolic" to "checked symbolic" to make room for "template
symbolic". Also rename "phase" to "dependence".
High level, replacing `Id::Invalid` with `Id::None` and `Id::is_valid`
with `Id::has_value` for clarity, as discussed
[here](https://discord.com/channels/655572317891461132/655578254970716160/1331664574545395794).
The `IntId` refactoring is needed together with `AnyIdBase` because it's
also used with `ValueStore`.
Note, trying to be careful not to rewrite `EnumBase::InvalidIndex`, or
`is_valid` in general (e.g., `IdKind::is_valid`).
I've tried to sequence commits here:
1. Automatic replacements:
- `((?:Id|Index)(?: |::|\(|Base(?:\(|::)))Invalid((?:Index)?\W)` ->
`$1None$2`
- `<invalid>` -> `<none>`
- `InvalidNodeId` -> `NoneNodeId`
- `/\*invalid\*/` -> `/*none*/`
- `id((?:_|\(\))(?:\.|->))is_valid` -> `id$1has_value`
2. Manual edits:
- In `int.h` and `int_test.cpp`
- `IntT` has `is_value`, which I'm renaming to `is_embedded_value`.
- Manual edits to comments in this file.
- `AnyIdBase` and `IdBase`
- Declaration of `is_valid` -> `has_value`, `InvalidIndex` ->
`NoneIndex`.
- In `ids.h` and `ids.cpp`
- `is_valid` -> `has_value`
- `// An explicitly invalid ID.` -> `// An ID with no value.`; similar
for index
- Various math on `InvalidIndex` -> `NoneIndex`
- Various mentions of "valid" in comments
- In `value_store.h`, for `IdT::Invalid`, plus one comment
- In `impl.h` and `tokenized_buffer.h`, we had different initialization
of `::None` values (versus `ids.h` syntax) that I fixed manually.
- Spot checks to compile
- Particularly where `is_valid` replacements didn't catch spots due to
different naming.
3. Autoupdate tests
4. verbose.carbon (NOAUTOUPDATE)
5. Comment spot checks
Note there are probably other mentions of "Invalid" that should be swept
up, but I'd like to argue for merging and separating out remaining
cleanup since this is so sweeping (and likely to hit merge conflicts
from churn). We'll probably have lingering mentions of "invalid" for a
bit regardless, just because there are uses of "invalid" in non-Id APIs.
* Rewrite constraints are stored in a facet type, substituted, imported,
and formatted.
* We now distinguish `.Self` from other symbolic bindings in two ways:
* `.Self` itself now has an invalid compile time binding index (since it
doesn't bind to any of the generic parameters). As a result, we no
longer need to create a generic region in `handle_where.cpp`.
* There is a new phase tracking values that are only symbolic because
they transitively depend on `.Self`. This allows us to give the result
of a `where` expression template phase as long as it doesn't use any
symbolic constants other than `.Self` or other designators.
* `AddConstant` has been removed from `check/context` since it was only
used from `eval`. This meant less plumbing of the phase change.
* Evaluation of `BindSymbolicName` now also performs substitution into
its type.
* Include a bit more information in some diagnostics.
* `StringifyTypeExpr` outputs rewrites, which required adding support
for associated entities as well.
* Associated entities now have an entity name set when importing.
* Adds tests for some interesting cases with rewrites and uses of
`.Self` mixed with other symbolic constants.
Still to do:
* There is no validation that any particular type satisfies rewrite
constraints.
* Access to members of a facet type do not see the rewritten values.
* Impls don't recognize whether associated constants have rewrites
setting their values.
* No support for resolving facet types.
---------
Co-authored-by: Josh L <josh11b@users.noreply.github.com>
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
Previously Collect() was used for types that implemented
CollectMemUsage() but otherwise Add() was used. This required the caller
to think about the type of the field and know/decide which method to
use.
Now, the caller always uses Collect() unless they are adding specific
byte values, in which case Add is used. Typically then, Add will only be
used to implement the CollectMemUsage() function.
To do this we require all Collect() methods to be templates so that they
all be a single overload set. The Collect on BumpPtrAllocator is
converted to a template that checks
`std::same_as<llvm::BumpPtrAllocator, T>`.
Most of these are about enabling inlining, in a couple of cases moving
code to a header and throughout switching to `CARBON_DCHECK`. The code
size of `CARBON_CHECK` seems to make inliing quite unreliable. I'm going
to think about whether there are ways to improve this, but a reasonably
small number of these seem worth switching for now to get some compile
time savings.
Also moves VLOG out of the hot path which helps a bit as well.
All combined, this net a bit over 10%, although it varies a bit exactly
how much. We're now pretty consistently over 800k lines/second for check
in the compilation benchmark for files >=4k lines, which makes me happy.
That's remarkably close to our original target.
Not really planning to keep optimizing here, just was glancing at the
profile and many of these stood out to me and were easy to fix.
This better follows the principle that types are simply constants of
type `type`, and allows more uniform treatment of types as just another
kind of constant from generics handling.
Use a hash table to map from `TypeId` to information about the complete
type. This makes basic operations on types a bit simpler, and operations
that actually need to access the complete class information a bit more
complex.
For each generic, build a list of instructions describing the
computations we need to do when resolving an instance of the generic:
this is a list of the instance-specific constants and types that the
generic uses. Another way of viewing this list is as a block of Carbon
SemIR code that is evaluated in order to form an instance of the generic
-- this is referenced in the code as the "eval block" for the generic.
For each instruction in the generic whose type or value is a symbolic
constant, replace that type or constant value with a symbolic reference
that says "to find the actual type or value, look at index N in the list
of values for the generic instance".
For an instruction with a symbolic constant value, we can just add that
instruction to our list. For an instruction with a symbolic constant
type, however, we may not have a corresponding instruction computing the
type within the generic and may need to build a new instruction, but
will reuse one where possible. In the case where we build a new
instruction, we use the existing substitution code to build the type
within the eval block.
For now, this transformation is only done in the declaration region of
the generic, not in the definition region. Also, we map back from the
symbolic references to the underlying constant value in a few places
where we will eventually need to do a lookup into a generic instance, in
order to avoid regressing the tests.
This works to leverage the capabilities of the hashtable as much as
possible, for example using the key context in the value stores.
However, there may still be opportunities to refactor more deeply and
use the functionality even better. Hopefully this is at least
a reasonable start and gets us a clean baseline.
On an Arm M1, this is a 15% improvement on my large lexing stress test,
but ends up a wash on my x86-64 server. This is a smaller benefit than
I expected, and it's because we're using a set-of-IDs and looking up
values with a key context for things like identifiers. This pattern has
a surprising tradeoff. The new hashtable uses significantly less memory,
a 10% peak RSS reduction just from the hashtable change. But indirecting
through the vector of values makes growing the hashtable dramatically
less cache-friendly: it causes growth to randomly access every key when
rehashing. On x86, everything gained by the faster hashtable is lost in
even slower growth. And even on Arm, this eats into the benefits.
But I have a plan to tweak how identifiers specifically work to avoid
most of the growth, and so I suspect this is the right tradeoff on the
whole. It gives us significant working set size reduction and we can
likely avoid the regressed operation (growth with rehash) in most cases
by clever reserving and if necessary by adding a hash caching layer to
the table infrastructure.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
When forming a `ConstantId` for a symbolic constant, add storage to
track the generic in which the constant was formed and the index within
that generic. These fields are not yet populated.
Require mapping from a `ConstantId` to an `InstId` to go through the
`ConstantValueStore`.
This is a preparatory step for an upcoming generics change where
symbolic `ConstantId`s are no longer just a thin wrapper around an
`InstId` but instead are indexes into a table with additional
information about the symbolic constant beyond its `InstId`.
Switch from recursing into non-canonical instruction fields to
separately canonicalizing those fields. This means we now form canonical
`InstBlockId`s, `TypeBlockId`s, `IntId`s, `FloatId`s, and `BindNameId`s
at least in the cases when they're referenced by a constant instruction.
This reduces the overall runtime for @chandlerc's 10MLoC example by
27.5% on my machine.
Make constant emission non-recursive, and stop building a bogus
FunctionContext to emit constants.
To support this, move `InstConstantKind` from the typed instruction
definition into the `.def` file, and add more macros to allow us to
generate case labels based on whether an instruction is a constant.
---------
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
Co-authored-by: Carbon Infra Bot <carbon-external-infra@google.com>
I'm proposing a different split, along the line of "what does this
relate to". I view impl.h as having started down this route. Moving the
inst store stuff to inst.h feels odd to me given how much else is there
right now, but maybe it's still the best approach. Some files only
contain a store, no structured class, but I felt the consistency in file
naming (without _store suffixes) might help.
Chandler Carruth
Richard Smith
Dana Jansens
Geoff Romer
David Blaikie
Jon Ross-Perkins
Dana Jansens
