v0.9.12-rc2 #4108
v0.9.12-rc2 #4108
Conversation
s3krit
commented
Oct 19, 2021
•
s3krit
commented
- Cherry-pick Remove stale migrations #4107
- Cherry-pick polkadot: remove call filters on registrar pallets #4093
- Apply benchmarks
s3krit
added
A3-in_progress
github-actions
bot
added
A0-please_review
* remove stale migrations * fix warnings * Update runtime/westend/src/lib.rs Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com> Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com>
|
New weights will be applied when the jobs in this pipeline complete: https://gitlab.parity.io/parity/polkadot/-/pipelines/162841 |
| @@ -42,69 +44,97 @@ use sp_std::marker::PhantomData; | |||
| /// Weight functions for `pallet_vesting`. | |||
| pub struct WeightInfo<T>(PhantomData<T>); | |||
| impl<T: frame_system::Config> pallet_vesting::WeightInfo for WeightInfo<T> { | |||
| fn vest_locked(l: u32, s: u32) -> Weight { | |||
| @@ -42,17 +44,17 @@ use sp_std::marker::PhantomData; | |||
| /// Weight functions for `pallet_utility`. | |||
| pub struct WeightInfo<T>(PhantomData<T>); | |||
| impl<T: frame_system::Config> pallet_utility::WeightInfo for WeightInfo<T> { | |||
| fn batch(c: u32) -> Weight { | |||
| (15_136_000 as Weight) | |||
There was a problem hiding this comment.
The batch and batch_all in every pallet seem to have a consistent upward, especially trend per input but this is def the biggest
| @@ -42,68 +44,96 @@ use sp_std::marker::PhantomData; | |||
| /// Weight functions for `pallet_vesting`. | |||
| pub struct WeightInfo<T>(PhantomData<T>); | |||
| impl<T: frame_system::Config> pallet_vesting::WeightInfo for WeightInfo<T> { | |||
| fn vest_locked(l: u32, s: u32) -> Weight { | |||
| (93_789_000 as Weight) | |||
There was a problem hiding this comment.
The weights for this whole pallet fairly consistently change. It looks like all the weights are more sensitive to inputs and have lower base weights. Not sure what to make of it
| fn force_lease() -> Weight { | ||
| (92_502_000 as Weight) |
There was a problem hiding this comment.
I'm running the benchmarks for runtime_common::slots manually to see if we get the same result again
| @@ -42,69 +44,97 @@ use sp_std::marker::PhantomData; | |||
| /// Weight functions for `pallet_vesting`. | |||
| pub struct WeightInfo<T>(PhantomData<T>); | |||
| impl<T: frame_system::Config> pallet_vesting::WeightInfo for WeightInfo<T> { | |||
| fn vest_locked(l: u32, s: u32) -> Weight { | |||
There was a problem hiding this comment.
I've sen these before in other pallets sometimes. The base is reduced, a component is increased. Ideally the benchmark values could be tweaked to not behave like this, but suppressing that in favour of time, I usually assume this specific patter as "it is fine"
There was a problem hiding this comment.
fwiw seems like vesting consistently changed across runtimes
|
/benchmark runtime kusama runtime_common_slots |
|
/benchmark runtime kusama runtime_common_slots |
|
Error running benchmark: mp-v0.9.12-rc2 stdoutFrom https://github.com/paritytech/polkadot * branch master -> FETCH_HEAD |
|
I'm unsure how to proceed with the changed weights... my feeling is to create followup issues to track whether the weight changes are 'organic' or rather an issue with the benchmarking setup (though I feel it is the former). @kianenigma @emostov do you have thoughts on this? edit Continued discussion of the weights can be discussed in this PR: #4113 |
