[ruby-core:107779] [Ruby master Feature#18589] Finer-grained constant invalidation

Issue #18589 has been updated by byroot (Jean Boussier).


> Do you have measurements about it on some apps?

So here's some metrics from the app I investigated. The metric is not directly `RubyVM.stat(:global_constant_state)`, instead we emit an increment if the `global_constant_state` changed during a request cycle.

The straight line is the last 24h, and the dotted line is the same days 4 weeks prior:

![Constant Caches](https://user-images.githubusercontent.com/19192189/156726006-75aab77a-7fdf-47cf-88cb-1175f193c18a.png)

The number one cause by far was `Object#extend` like in [`open-uri`](https://github.com/ruby/open-uri/pull/8), [`aws-sdk`](https://github.com/aws/aws-sdk-ruby/pull/2670) as well as `ActiveRecord::Reation#extending`.

The second most common cause was gems using a lot of `autoload`, or having memoized class attribute like [tzinfo](https://github.com/tzinfo/tzinfo/pull/129).

Overall fixing these had a very significant impact on the service latency

![Latency](https://user-images.githubusercontent.com/19192189/156727814-adb0f8b5-9012-4d2c-ab9c-b29d80748a5c.png)

I'm however worried that this situation will degrade over time, as there's very little way to actively prevent this kind of problems from cropping up.

----------------------------------------
Feature #18589: Finer-grained constant invalidation
https://bugs.ruby-lang.org/issues/18589#change-96700

* Author: kddeisz (Kevin Newton)
* Status: Open
* Priority: Normal
----------------------------------------
This is related to https://github.com/ruby/ruby/pull/5433.

## Current behavior

Caches depend on a global counter. All constant mutations cause all caches to be invalidated.

```ruby
class A
  B = 1
end

def foo
  A::B # inline cache depends on global counter
end

foo # populate inline cache
foo # hit inline cache

C = 1 # global counter increments, all caches are invalidated

foo # misses inline cache due to `C = 1`
```

## Proposed behavior

Caches depend on name components. Only constant mutations with corresponding names will invalidate the cache.

```ruby
class A
  B = 1
end

def foo
  A::B # inline cache depends constants named "A" and "B"
end

foo # populate inline cache
foo # hit inline cache

C = 1 # caches that depend on the name "C" are invalidated

foo # hits inline cache because IC only depends on "A" and "B"
```

Examples of breaking the new cache:

```ruby
module C
  # Breaks `foo` cache because "A" constant is set and the cache in foo depends
  # on "A" and "B"
  class A; end
end

B = 1
```

We expect the new cache scheme to be invalidated less often because names aren't frequently reused. With the cache being invalidated less, we can rely on its stability more to keep our constant references fast and reduce the need to throw away generated code in YJIT.

## Performance benchmarks

The following benchmark (included in this pull request) performs about 2x faster than master.

```ruby
CONSTANT1 = 1
CONSTANT2 = 1
CONSTANT3 = 1
CONSTANT4 = 1
CONSTANT5 = 1

def constants
  [CONSTANT1, CONSTANT2, CONSTANT3, CONSTANT4, CONSTANT5]
end

500_000.times do
  constants
  INVALIDATE = true
end
```

In terms of macro benchmarks, I ran with this code on railsbench and there was not a statistically significant different in startup time or overall runtime performance.

## Memory benchmarks

In terms of memory, this includes an increase in VM size by about 500KiB when running on railsbench. This is because we're now tracking cache associations ({ ID => IC[] }) on the VM to know how to invalidate specific caches when constants change.

I booted Shopify's core monolith with this branch as well. It increased total retained memory from 1.23Gb to 1.3Gb (about a 0.7% increase). The memory increase is proportional to the number of constant caches found in the application. For each constant cache 1 level deep (e.g., `Foo`) the increase is about 33 bytes. For a constant cache 2 levels deep (e.g., `Foo::Bar`) the increase is about 67 bytes.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:[e-mail]?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>