[Costing] Provide support for multiple 'CostModel's #5851
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
effectfully
changed the title
effectfully
force-pushed
the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
from
c8abb84
to
a9a71bd
Compare
|
/benchmark validation |
|
Click here to check the status of your benchmark. |
effectfully
force-pushed
the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
3 times, most recently
from
18f2db1
to
9758d10
Compare
effectfully
force-pushed
the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
from
9758d10
to
b97e9f1
Compare
|
/benchmark validation |
|
Click here to check the status of your benchmark. |
|
Comparing benchmark results of 'validation' on '381172295' (base) and 'b97e9f11e' (PR) Results table
|
|
/benchmark nofib |
|
Click here to check the status of your benchmark. |
…to effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
effectfully
force-pushed
the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
from
b97e9f1
to
0a1af53
Compare
|
Click here to check the status of your benchmark. |
|
/benchmark lists |
1 similar comment
|
/benchmark lists |
|
Comparing benchmark results of 'validation' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
|
Click here to check the status of your benchmark. |
|
Comparing benchmark results of 'validation' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
|
Click here to check the status of your benchmark. |
|
Comparing benchmark results of 'nofib' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
|
Click here to check the status of your benchmark. |
zliu41
approved these changes
Apr 19, 2024
| @@ -1287,7 +1291,8 @@ instance uni ~ DefaultUni => ToBuiltinMeaning uni DefaultFun where | |||
| :: BS.ByteString -> BS.ByteString -> BS.ByteString -> BuiltinResult Bool | |||
| verifyEd25519SignatureDenotation = | |||
| case semvar of | |||
| DefaultFunSemanticsVariant1 -> verifyEd25519Signature_V1 | |||
| DefaultFunSemanticsVariant0 -> verifyEd25519Signature_V1 | |||
| DefaultFunSemanticsVariant1 -> verifyEd25519Signature_V2 | |||
There was a problem hiding this comment.
This mapping is potentially easy to get wrong. Is there any test for this?
There was a problem hiding this comment.
We're just replacing one implementation with another for maintenance reasons, so doing this matching here is only for the purpose of being paranoid about not breaking anything in the past in case of some unknown unknown. There's nothing additional to test therefore beyond what we already test, which is how we use ed25519_Variant0Prop to test all variants of this builtin.
Note that consByteString is untestable in the same way, but for a different reason: Variant0 and Variant1 versions of that builtin are completely identical (unlike with verifyEd25519Signature, whose variants behave identically but come from different libraries, i.e. are nominally different), so again there's nothing to test.
So for both the builtins we only make sure that whatever their variant is, it behaves as expected, we cannot test that Variant0 behaves differently from Variant1, because it doesn't.
|
Comparing benchmark results of 'nofib' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
|
Click here to check the status of your benchmark. |
|
Comparing benchmark results of 'lists' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
|
Click here to check the status of your benchmark. |
|
Comparing benchmark results of 'lists' on '16a986fd8' (base) and '9d8942da1' (PR) Results table
|
michaelpj
approved these changes
Apr 19, 2024
| @@ -85,6 +86,9 @@ defaultCekMachineCosts = | |||
| defaultCekCostModel :: CostModel CekMachineCosts BuiltinCostModel | |||
| defaultCekCostModel = CostModel defaultCekMachineCosts defaultBuiltinCostModel | |||
|
|
|||
| toCekCostModel :: BuiltinSemanticsVariant DefaultFun -> CostModel CekMachineCosts BuiltinCostModel | |||
| toCekCostModel _ = defaultCekCostModel | |||
There was a problem hiding this comment.
Should it therefore have some comment explaining that?
kwxm
approved these changes
Apr 22, 2024
| | DefaultFunSemanticsVariant2 | ||
| deriving stock (Enum, Bounded, Show) | ||
| data BuiltinSemanticsVariant DefaultFun | ||
| = DefaultFunSemanticsVariant0 |
There was a problem hiding this comment.
I vote that we change these to DefaultSemanticsVariantA/B/C or something similar. It's quite confusing having all these variants and versions in the code, so it might help if they're not all identified by numbers that might lead people to believe that the same number means the same thing in different contexts. That would mean changing the specification too though, so if we do it let's do it later in its own PR.
There was a problem hiding this comment.
I'm fine either way. I'm hopeful we'll be able to get rid of the semantics variants entirely and simply rely on the language and protocol versions directly (maybe in the "condensed" form).
| >=> mkDynEvaluationContext | ||
| PlutusV3 | ||
| [DefaultFunSemanticsVariant2] | ||
| (const Plutus.DefaultFunSemanticsVariant2) |
There was a problem hiding this comment.
Again we could probably do without the Plutus here and in the import.
There was a problem hiding this comment.
Let me check: V1 and V2 have if pv < conwayPV then ... else ... here, but it's safe to have const for V3 because we know that PlutusV3 is going to be introduced at the same time as conwayPV, so a PlutusV3 script can never be executed with a lower PV, so there's no possiblity that pv < conwayPV here (but conceivably we might need if pv < einsteinPV or something in the future). Is that right? Maybe we should try to explain that somewhere, since there's a LL/PV dependency that I don't think is captured in the code anywhere. Maybe we should have note containing a table like we do in the PLC specification (or a big table that contains all of the relevant information, since there are multiple tables in the spec).
[Later] Actually, maybe the dependencies are specified in PlutusLedgerApi.Common.Versions.
There was a problem hiding this comment.
Let me check: V1 and V2 have
if pv < conwayPV then ... else ...here, but it's safe to haveconstfor V3 because we know thatPlutusV3is going to be introduced at the same time asconwayPV, so a PlutusV3 script can never be executed with a lower PV, so there's no possiblity thatpv < conwayPVhere (but conceivably we might needif pv < einsteinPVor something in the future). Is that right?
I believe so, yes.
Maybe we should try to explain that somewhere, since there's a LL/PV dependency that I don't think is captured in the code anywhere. Maybe we should have note containing a table like we do in the PLC specification (or a big table that contains all of the relevant information, since there are multiple tables in the spec).
Good point, I'll add a small table somewhere, but I'm not totally sure where given that the logic is spread across multiple files. I guess PlutusLedgerApi.Common.Versions indeed.
| evaluation. | ||
|
|
||
| Different protocol versions may require different bundles of machine parameters, which allows us for |
There was a problem hiding this comment.
Maybe make this into a Note? It seems quite important. It would seem to have some bearing on this, for example.
There was a problem hiding this comment.
Maybe make this into a Note?
I kinda want this to be the Haddock for extra transparency. I do reference the Haddock of a function instead of a Note sometimes.
|
/benchmark validation |
|
Click here to check the status of your benchmark. |
effectfully
force-pushed
the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
from
e6afc0b
to
039dce1
Compare
…to effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
effectfully
deleted the
effectfully/builtins/both-MajorProtocolVersion-and-multiple-CostModels
branch
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This makes cost models configurable, so that the cost model can depend on the protocol version. All commentary is in the diff.