The gas estimator is a tool that aggregates statistics on gas consumption for each Curve pool contract on mainnet. These estimates are derived from Parity traces of every transaction where the contract emits an event. The final product is a json output file that has information on gas statistics (you can choose the modality of the gaussian statistics used here), as well as the number of transactions that make up the stats and the min and max block ranges where these gas stats are aggregated:
Since this repository is specialised to Curve Finance pools, there are two modalities of gas statistics:
- Univariate: gas usage has a gaussian distribution
- Mixed Gaussian / Bimodal: gas usage has a bimodal distribution, with each component its own gaussian model.
The reason for a bimodal gaussian distribution is because there are two gas consumption regimes in Curve Cryptoswap pools: one where the liquidity in the pool is positions very close to where the market is trading at, and one regime (consumes more gas) where the liquidity is further away.
Additionally there are a few debugging tools built in that fetch the trace of a transaction but only decode traces for a single contract (if you only care about one contract in a transaction, then the decoding is much faster).
This is a work in progress, but users can already get estimates for Curve contracts. Bear in mind that getting accurate gas statistics over several thousand transactions takes quite some time. Ideally the user should set the scripts up in a remote server with an Erigon archive node (with debug mode enabled: else you won't get the traces).
If you are a Curve integrator, a DEX aggregator, a researcher/analyst, building at or on Curve, gas optimizoooor, or an MEV searcher who wants accurate gas estimates for their Curve swaps, then this is for you.
The scope of these gas estimates are isolated to mainnet, but the gas usage should be similar for similar contracts across other EVM chains.
The tool offers very simple stats at present. Should there be a need for more accurate modeling of gas usage, then please write an issue detailing your needs and its urgency. If you are a power user of this code and you think you can add a feature that you use on a daily basis to this repository, don't hesitate to submit a pull request from your forked repository.
Requirements:
- ApeWorx
- Erigon with Archive node and debug mode enabled
To install the scripts, you'll need python >3.10.4. Pyenv is recommended here. After cloning the repo:
python -m venv venv
source ./venv/bin/activate
pip install --upgrade pip
pip install -r ./requirements.txt
You will also need to set up your ETHERSCAN_API_KEY into your environment variables, so ape-etherscan can access it to get contract ABIs from Etherscan.
Finally, once you've set up your Erigon node, set up ape-config.yaml in the following manner:
ethereum:
default_network: mainnet
mainnet:
default_provider: geth
geth:
ethereum:
mainnet:
uri: http://localhost:{your-port-here}
> ape run gas_tools pools --max_transactions 1000 --pool_type stableswap --pool 0x4CA9b3063Ec5866A4B82E437059D2C43d1be596F
saved output file:
{
"0x4CA9b3063Ec5866A4B82E437059D2C43d1be596F": {
"univariate": {
"exchange": {
"mean": 130126,
"std": 14684,
"min": 85958,
"max": 150079,
"count": 536
},
"remove_liquidity_one_coin": {
"mean": 129403,
"std": 12454,
"min": 101533,
"max": 157163,
"count": 226
},
"get_dy": {
"mean": 23712,
"std": 164,
"min": 23226,
"max": 23767,
"count": 40
},
"add_liquidity": {
"mean": 150148,
"std": 20711,
"min": 101980,
"max": 204157,
"count": 218
},
"get_virtual_price": {
"mean": 10695,
"std": 4809,
"min": 8955,
"max": 23455,
"count": 25
},
"calc_withdraw_one_coin": {
"mean": 42581,
"std": 6196,
"min": 38581,
"max": 50581,
"count": 6
},
"coins": {
"mean": 3201,
"std": 0,
"min": 3201,
"max": 3201,
"count": 4
},
"remove_liquidity": {
"mean": 142906,
"std": 11471,
"min": 119821,
"max": 154021,
"count": 20
},
"balances": {
"mean": 1230,
"std": 0,
"min": 1230,
"max": 1230,
"count": 2
},
"ramp_A": {
"mean": 25184,
"std": 0,
"min": 25184,
"max": 25184,
"count": 1
},
"remove_liquidity_imbalance": {
"mean": 136085,
"std": 9872,
"min": 130385,
"max": 147485,
"count": 3
}
},
"min_block": 13352648,
"max_block": 15344943
}
}
the argument pool for gas_tools has three modes: stableswap, cryptoswap and all, which does both stableswap and cryptoswap pool gas estimates.
Set max_transactions to ensure that a maximum of n transactions are used in the gas stats: if the pool does not have n transactions, gas_tool will calculate stats on whatever it can find or whatever it thinks it needs. By default, the latest n transactions are chosen.
For stableswap:
ape run gas_tools pools --max_transactions 100 --pool_type stableswap
For cryptoswa:
ape run gas_tools pools --max_transactions 100 --pool_type cryptoswap
For both stableswap and cryptoswap:
ape run gas_tools pools --max_transactions 100 --pool_type all
If you want to get stats for a single stableswap or cryptoswap pool:
ape run gas_tools pools --max_transactions 100 --pool_type stableswap --pool 0x4CA9b3063Ec5866A4B82E437059D2C43d1be596F
By default all stableswap estimates are stored in ./stableswap_pool_gas_estimates.json and the same for cryptoswap is ./cryptoswap_pool_gas_estimates.json
If you want to debug a transaction for a specific contract, use the argument tx:
ape run gas_tools tx --contractaddr 0x4CA9b3063Ec5866A4B82E437059D2C43d1be596F --tx 0xa57089bbd7a7a4e7b2fb52a5def3ebc7c3c9d8e5806272c8a216c93279f223ce
This will return a trace where only the pool contract is decoded. It will also provide average gas costs for each method called in the contract.
(c) Curve.Fi, 2022 - All rights reserved.