perf(internal/bits): Significantly speedup bitArray.PickRandom #2841

ValarDragon · 2024-04-18T07:22:11Z

This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!)

This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine.

OLD:

BenchmarkPickRandomBitArray-12           1545199               846.1 ns/op          1280 B/op          1 allocs/op

NEW:

BenchmarkPickRandomBitArray-12          22192857                75.39 ns/op            0 B/op          0 allocs/op

I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see!

PR checklist

Tests written/updated
Changelog entry added in .changelog (we use unclog to manage our changelog)
Updated relevant documentation (docs/ or spec/) and code comments
Title follows the Conventional Commits spec

melekes

Thanks @ValarDragon ❤️

cason

Great work.

Left some comments regarding the documentation of the introduced methods.

internal/bits/bit_array_test.go

internal/bits/bit_array.go

cason · 2024-04-18T08:23:17Z

internal/bits/bit_array.go

-	trueIndices := make([]int, 0, bA.Bits)
-	curBit := 0
+// getNthTrueIndex returns the index of the nth true bit in the bit array.
+// If there is no such value, it returns 0. It is required that the caller


Returns -1, doesn't it?

Whoops yes, thank you!

cason · 2024-04-18T08:23:44Z

internal/bits/bit_array.go

-	curBit := 0
+// getNthTrueIndex returns the index of the nth true bit in the bit array.
+// If there is no such value, it returns 0. It is required that the caller
+// ensures that n is less than or equal to the number of true bits in the bit array.


This is requirement is not really a requirement, right? The code support this case, but returns -1.

cason · 2024-04-18T08:37:54Z

I didn't add to the review, but I was thinking the following. This is only a suggestion.

We implement PickRandom() by 1. finding all bits set 1/true, lets say n, then 2. randomly taking one of these bits (random from 0 to n-1). So, we need 1. n to produce the random number used in 2.

Assuming that we know the maximum number of bits stored, lets say m, why we don't pick a random value from 0 to m, lets say i. Of course, probably in most cases, we have m > n, so potentially i > n. But in this case, actually in the general case, we just need to pick the index i % n. The problem, of course, is that we don't know n a priori (which can be solved in a different way, but lets keep this a bit simpler).

My point here is that we probably can implement PickRandom() with a single method. Of course, this method would need another loop, to handle the case where i > n. But in this case, after the first loop, we know n and we know that i > n. Therefore can compute i % n, and solve the problem in the second iteration.

cason · 2024-04-18T08:50:57Z

internal/bits/bit_array_test.go

@@ -301,3 +387,15 @@ func TestUnmarshalJSONDoesntCrashOnZeroBits(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, ic.BitArray, &BitArray{Bits: 0, Elems: nil})
 }
+
+func BenchmarkPickRandomBitArray(b *testing.B) {


I wonder if adding some contention to this benchmark would affect the results. I think the results are better with this PR, but maybe this is something to explore. By contention, I mean, another thread changing the bit set while we are retrieving random indexes.

Where this is used, these mutexes are truly useless. We are operating on a clone that is only owned in a single threaded function. Was going to make an issue to make a new bit array version that gets rid of the mutexes

This would be ideal. In particular because I assume that the methods using this are already synchronized.

For instance, consensus state, peer consensus state should be synchronized.

ValarDragon · 2024-04-18T09:29:51Z

I thought about the second method, but getting perfect distribution will require rejection sampling which is potentially much slower. (Doing i%n biases the random distribution which is more scary for edge case slowdowns imo.)

Rejection sampling seems like unneeded edge case slowdowns right now. Targeting a single pass seems like large overkill when in the systemic view or what's going on with the bit arrays, we're very wasteful at the moment. (Two Extra full clones in the relevant vote sampling code. We clone the bit array from votes, then do a sub call that allocates another bit array, and everything here is single threaded but does mutexes. The mutex call's Cas operation is likely slower than the first pass of this codes ~16 bitwise math ops per word)

cason · 2024-04-18T09:41:05Z

I didn't fully get this rejection sampling part, but I was wondering whether i % n is as random as rand(n), but I am not sure. Maybe it is not, but probably if we make m big enough, it becomes pretty much as random as.

cason · 2024-04-18T09:42:37Z

Also agree that removing locks is much more relevant for performance than saving a few loops.

cason · 2024-04-18T09:52:06Z

The problem, of course, is that we don't know n a priori (which can be solved in a different way, but lets keep this a bit simpler).

We can have this as a private field:

SetIndex adds or subtracts 1
copy() can be adapted easily
IsEmpty() and IsFull() become trivial
Update(), I don't even know the point, same as copy()
UnmarshalJSON() uses SetIndex, so nothing new here
FromProto() can have this field initialized with a private method (the one added here)

I should probably create an issue from this.

ValarDragon · 2024-04-18T11:04:05Z

Your totally right, that would work! I expect that the difference in performance for the PickRandom call won't be that high (I think were already fast enough here / remaining time is coming from the Mutex), but agreed that seems like a better solution!

This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- #### PR checklist - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec (cherry picked from commit b139e13)

This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- #### PR checklist - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec (cherry picked from commit b139e13) # Conflicts: # libs/bits/bit_array_test.go

This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- #### PR checklist - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec (cherry picked from commit b139e13) # Conflicts: # .changelog/v0.37.5/improvements/2841-speedup-bits-pick-random.md # libs/bits/bit_array_test.go

…ort #2841) (#2887) This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- #### PR checklist - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request #2841 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

Follow-up to #2841

Follow-up to #2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec

Follow-up to #2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec (cherry picked from commit b831f95)

Follow-up to #2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request #2899 done by [Mergify](https://mergify.com). Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

…ort cometbft#2841) (cometbft#2887) This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2841 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

#28) * perf(internal/bits): Significantly speedup bitArray.PickRandom (backport cometbft#2841) (cometbft#2887) This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2841 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> * fix failing test * changelog --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

#28) * perf(internal/bits): Significantly speedup bitArray.PickRandom (backport cometbft#2841) (cometbft#2887) This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2841 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> * fix failing test * changelog --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> (cherry picked from commit f6abfce)

#28) (#32) * perf(internal/bits): Significantly speedup bitArray.PickRandom (backport cometbft#2841) (cometbft#2887) This PR significantly speeds up bitArray.PickRandom which is used in VoteGossip and BlockPart gossip. We saw for a query serving full node, over an hour, this was a very large amount of RAM allocations. (75GB of RAM!) ![image](https://github.com/cometbft/cometbft/assets/6440154/755918a5-0cef-4e67-a47e-ce8a56aa1cd5) This PR drops it down to 0 allocations, and makes the routine 10x faster on my machine. OLD: ``` BenchmarkPickRandomBitArray-12 1545199 846.1 ns/op 1280 B/op 1 allocs/op ``` NEW: ``` BenchmarkPickRandomBitArray-12 22192857 75.39 ns/op 0 B/op 0 allocs/op ``` I think the new tests I wrote make this more tested than the old code that was here tbh, but pls let me know if theres more tests we'd like to see! --- - [x] Tests written/updated - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2841 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> * fix failing test * changelog --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> (cherry picked from commit f6abfce) Co-authored-by: Adam Tucker <adam@osmosis.team>

…cometbft#2902) Follow-up to cometbft#2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2 C95D 899 done by [Mergify](https://mergify.com). Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

…cometbft#2902) (#35) Follow-up to cometbft#2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2899 done by [Mergify](https://mergify.com). Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>

…cometbft#2902) (#35) Follow-up to cometbft#2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2899 done by [Mergify](https://mergify.com). Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> (cherry picked from commit 08efc4c)

…cometbft#2902) (#35) (#37) Follow-up to cometbft#2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2899 done by [Mergify](https://mergify.com). Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> (cherry picked from commit 08efc4c) Co-authored-by: Adam Tucker <adam@osmosis.team>

…cometbft#2902) (#35) (#36) Follow-up to cometbft#2841 --- #### PR checklist - [x] Tests written/updated - [ ] ~~Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog)~~ - [ ] ~~Updated relevant documentation (`docs/` or `spec/`) and code comments~~ - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2899 done by [Mergify](https://mergify.com). Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com> (cherry picked from commit 08efc4c) Co-authored-by: Adam Tucker <adam@osmosis.team>

Speedup bitArray.PickRandom

193d173

ValarDragon requested review from a team as code owners April 18, 2024 07:22

ValarDragon changed the title ~~perf(internal/bits) Significantly speedup bitArray.PickRandom~~ perf(internal/bits): Significantly speedup bitArray.PickRandom Apr 18, 2024

ValarDragon added 2 commits April 18, 2024 16:27

Changelog + two more test cases

6274481

Delete now unused method

4c9e021

melekes approved these changes Apr 18, 2024

View reviewed changes

melekes added backport-to-v0.37.x Tell Mergify to backport the PR to v0.37.x backport-to-v0.38.x Tell Mergify to backport the PR to v0.38.x backport-to-v1.x Tell Mergify to backport the PR to v1.x labels Apr 18, 2024

cason approved these changes Apr 18, 2024

View reviewed changes

adizere added this to the 2024-Q2 milestone Apr 18, 2024

cason reviewed Apr 18, 2024

View reviewed changes

Apply @cason's comment improvements!

f9656f8

cason mentioned this pull request Apr 18, 2024

libs/bits: store the number of non-zero bits in a private field #2849

Closed

melekes added this pull request to the merge queue Apr 24, 2024

Merged via the queue into cometbft:main with commit b139e13 Apr 24, 2024
36 checks passed

mergify bot mentioned this pull request Apr 24, 2024

perf(internal/bits): Significantly speedup bitArray.PickRandom (backport #2841) #2885

Merged

4 tasks

This was referenced Apr 24, 2024

perf(internal/bits): Significantly speedup bitArray.PickRandom (backport #2841) #2886

Merged

perf(internal/bits): Significantly speedup bitArray.PickRandom (backport #2841) #2887

Merged

melekes added a commit that referenced this pull request Apr 26, 2024

fix(bits): handle last element in PickRandom

ac7b875

Follow-up to #2841

melekes mentioned this pull request Apr 26, 2024

fix(bits): handle last element in PickRandom #2899

Merged

4 tasks

This was referenced Apr 26, 2024

fix(bits): handle last element in PickRandom (backport #2899) #2900

Merged

fix(bits): handle last element in PickRandom (backport #2899) #2901

Merged

fix(bits): handle last element in PickRandom (backport #2899) #2902

Merged

czarcas7ic mentioned this pull request Apr 29, 2024

perf(internal/bits): Significantly speedup bitArray.PickRandom (backp… osmosis-labs/cometbft#28

Merged

7 tasks

mergify bot mentioned this pull request Apr 30, 2024

perf(internal/bits): Significantly speedup bitArray.PickRandom (backp… (backport #28) osmosis-labs/cometbft#32

Merged

7 tasks

czarcas7ic mentioned this pull request May 2, 2024

fix(bits): handle last element in PickRandom (backport #2899) (#2902) osmosis-labs/cometbft#35

Merged

7 tasks

mergify bot mentioned this pull request May 2, 2024

fix(bits): handle last element in PickRandom (backport #2899) (#2902) (backport #35) osmosis-labs/cometbft#36

Merged

7 tasks

mergify bot mentioned this pull request May 2, 2024

fix(bits): handle last element in PickRandom (backport #2899) (#2902) (backport #35) osmosis-labs/cometbft#37

Merged

7 tasks

dynst mentioned this pull request Jun 14, 2024

chore: 🌠 upgrade to cometbft v0.37.6. penumbra-zone/penumbra#4369

Closed

faddat mentioned this pull request Oct 15, 2024

chore: use latest cometbft-db in v0.38.x #4296

Closed

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

perf(internal/bits): Significantly speedup bitArray.PickRandom #2841

perf(internal/bits): Significantly speedup bitArray.PickRandom #2841

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

perf(internal/bits): Significantly speedup bitArray.PickRandom #2841

perf(internal/bits): Significantly speedup bitArray.PickRandom #2841

Uh oh!

Conversation

Uh oh!

PR checklist

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!