-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
types: use (*math/big.Int).BitLen() == 0 to check if value is 0 #8580
Conversation
Codecov Report
@@ Coverage Diff @@
## master #8580 +/- ##
==========================================
+ Coverage 61.43% 61.46% +0.02%
==========================================
Files 656 656
Lines 37533 37526 -7
==========================================
+ Hits 23059 23064 +5
+ Misses 12065 12053 -12
Partials 2409 2409
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've requested just the addition of a test case; looks good otherwise (and please feel free to dismiss my review once the test case is added)
Instead of using len((*math/big.Int).Bytes()) == 0, which expensively creates a byte slice and marshals a value, on the majority hot path, instead use the cheaper method .BitLen() to check if 0. Benchmarking results, just from types: name old time/op new time/op delta CoinsAdditionIntersect/sizes:_A_1,_B_1-8 132ns ± 2% 126ns ±13% -4.55% (p=0.050 n=10+10) CoinsAdditionIntersect/sizes:_A_5,_B_5-8 1.41µs ± 3% 1.41µs ± 2% ~ (p=1.000 n=10+10) CoinsAdditionIntersect/sizes:_A_5,_B_20-8 2.30µs ± 1% 2.27µs ± 3% ~ (p=0.066 n=10+10) CoinsAdditionIntersect/sizes:_A_1,_B_1000-8 30.9µs ± 3% 30.7µs ± 1% ~ (p=0.218 n=10+10) CoinsAdditionIntersect/sizes:_A_2,_B_1000-8 31.4µs ± 3% 30.8µs ± 2% -1.94% (p=0.015 n=10+10) CoinsAdditionNoIntersect/sizes:_A_1,_B_1-8 116ns ± 1% 114ns ± 4% ~ (p=0.142 n=10+10) CoinsAdditionNoIntersect/sizes:_A_5,_B_5-8 1.11µs ± 1% 1.08µs ± 3% -2.36% (p=0.003 n=8+10) CoinsAdditionNoIntersect/sizes:_A_5,_B_20-8 1.85µs ± 2% 1.82µs ± 1% -1.38% (p=0.001 n=10+9) CoinsAdditionNoIntersect/sizes:_A_1,_B_1000-8 30.7µs ± 1% 30.6µs ± 3% ~ (p=0.393 n=10+10) CoinsAdditionNoIntersect/sizes:_A_2,_B_1000-8 31.1µs ± 1% 30.7µs ± 2% -1.32% (p=0.015 n=10+10) CoinsAdditionNoIntersect/sizes:_A_1000,_B_2-8 31.0µs ± 2% 30.7µs ± 2% ~ (p=0.190 n=10+10) Bech32ifyPubKey-8 28.8µs ± 5% 28.8µs ± 3% ~ (p=0.965 n=10+8) GetPubKeyFromBech32-8 38.8µs ± 3% 39.4µs ± 2% +1.70% (p=0.013 n=9+10) ParseCoin-8 16.7µs ± 6% 15.8µs ± 4% -5.21% (p=0.001 n=10+10) MarshalTo-8 521ns ± 5% 508ns ± 3% -2.56% (p=0.029 n=10+10) UintMarshal-8 3.10µs ±17% 2.56µs ± 3% -17.45% (p=0.000 n=10+9) IntMarshal-8 2.52µs ±10% 1.94µs ± 2% -23.10% (p=0.000 n=10+10) name old alloc/op new alloc/op delta Bech32ifyPubKey-8 4.02kB ± 0% 4.02kB ± 0% ~ (all equal) GetPubKeyFromBech32-8 2.48kB ± 0% 2.48kB ± 0% ~ (all equal) ParseCoin-8 2.21kB ± 0% 2.21kB ± 0% ~ (all equal) MarshalTo-8 80.0B ± 0% 80.0B ± 0% ~ (all equal) UintMarshal-8 440B ± 0% 392B ± 0% -10.91% (p=0.000 n=10+10) IntMarshal-8 216B ± 0% 168B ± 0% -22.22% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Bech32ifyPubKey-8 25.0 ± 0% 25.0 ± 0% ~ (all equal) GetPubKeyFromBech32-8 85.0 ± 0% 85.0 ± 0% ~ (all equal) ParseCoin-8 71.0 ± 0% 71.0 ± 0% ~ (all equal) MarshalTo-8 2.00 ± 0% 2.00 ± 0% ~ (all equal) UintMarshal-8 31.0 ± 0% 25.0 ± 0% -19.35% (p=0.000 n=10+10) IntMarshal-8 24.0 ± 0% 18.0 ± 0% -25.00% (p=0.000 n=10+10) name old speed new speed delta UintMarshal-8 2.27MB/s ±15% 2.75MB/s ± 2% +20.87% (p=0.000 n=10+8) IntMarshal-8 2.78MB/s ± 9% 3.60MB/s ± 2% +29.69% (p=0.000 n=10+10) Fixes #8575
8c68c68
to
efa1bb4
Compare
Thank you for the review @alessio! I've added the tests, please take a look. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thank you for the PR
Instead of using len((*math/big.Int).Bytes()) == 0, which expensively
creates a byte slice and marshals a value, on the majority hot path,
instead use the cheaper method .BitLen() to check if 0.
Benchmarking results, just from types:
Fixes #8575
Before we can merge this PR, please make sure that all the following items have been
checked off. If any of the checklist items are not applicable, please leave them but
write a little note why.
docs/
) or specification (x/<module>/spec/
)godoc
comments.Unreleased
section inCHANGELOG.md
Files changed
in the Github PR explorerCodecov Report
in the comment section below once CI passes