Benchmark computation time of txs #3898
Assignees
Labels
required
issue is required to be closed before workstream can be closed
WS: V3 3️⃣
item is directly relevant to the v3 hardfork
Comments
evan-forbes
added
WS: V3 3️⃣
rach-id
added a commit
that referenced
this issue
Oct 21, 2024
Works on this: #3898 To run the benchmarks, run the following in the root directory: ```shell go test -bench=<benchmark_name> app/benchmark_test.go ``` with `<benchmark_name>` is: - `BenchmarkCheckTx_MsgSend_1`: for a benchmark of a `MsgSend` transaction in `CheckTx` - `BenchmarkCheckTx_MsgSend_8MB`: for a benchmark of 8mb block worth of `MsgSend` transactions in `CheckTx` - `BenchmarkDeliverTx_MsgSend_1`: for a benchmark of a `MsgSend` transaction in `DeliverTx` - `BenchmarkDeliverTx_MsgSend_8MB`: for a benchmark of 8mb block worth of `MsgSend` transactions in `DeliverTx` - `BenchmarkPrepareProposal_MsgSend_1`: for a benchmark of a block containing a single`MsgSend` transaction in `PrepareProposal` - `BenchmarkPrepareProposal_MsgSend_8MB`: for a benchmark of an 8mb block containing `MsgSend` transactions in `PrepareProposal` - `BenchmarkProcessProposal_MsgSend_1`: for a benchmark of a block containing a single`MsgSend` transaction in `ProcessProposal` - `BenchmarkProcessProposal_MsgSend_8MB`: for a benchmark of an 8mb block containing`MsgSend` transactions in `ProcessProposal` - ... Note: keeping this as a draft because it doesn't necessarily need to be merged Benchmark run on: Macbook pro M3 max 48GB RAM --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: Rootul P <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Summary
As we look to increase block size, we'd like to make sure that there's no path for transactions that are computationally heavy from significantly affecting the block production rate. To understand this we should research the effect of the following:
Concretely, I assume for the first point we could use IBC UpdateClient transactions (as they require signature verification). For the second we could try
MsgSend. The metric would be to measure the time taken of:PrepareProposalandProcessProposalCheckTxDeliverTxThese could be done as benchmarks so we can monitor them (or modify them) in the future
(We may need a reference point to compare with but we can decide this after getting the initial numbers)
The text was updated successfully, but these errors were encountered: