I think we can all agree that tps numbers can be hard to compare between chains. Some chains can pin 10000s of native token transfers, but things get weird when you start to compare smart contracts.
Last year, I posted about this metric in the Algorand subreddits for determining throughput of different blockchains in the fairest way possible. That is using an AMM “uniswap-style” swap as the benchmark.
The authors of the medium article linked in my original post tested the smart contract throughput of some of the top smart contract platforms empirically, and they found that these chains to have the following AMM-swap tps limits:
Solana Mainnet Orca – 273 swaps BSC pancakeswap – 195 swaps Polygon quickswap – 95 swaps Avax Trader Joe – 176 swaps Celo Ubeswap – 50 swaps Ethereum uniswap v2 – 18 swaps
Immediately after the 6k tps upgrade, I made a post about Algorand’s ability to perform these AMM-style swaps which was estimated using the assumption that an AMM swap would require four txns per swap. This estimation came out to about 1625 swaps per second. At the time this was very impressive because the next fastest chain was Solana capped at being able to do 273 swaps per second.
People foreshadowed in the comments on one of my previous posts saying that it could definitely be done in less than four transactions, but I wanted to be conservative at the time.
Today, I saw on twitter that one of the developers from Vestige actually empirically tested this on the MAINNET and proved that Algorand can do 2881 AMM-style swaps per second. You can see the on-chain evidence in this block.
10.5x more than Solana Mainnet Orca – 273 swaps 15x more than BSC pancakeswap – 195 swaps 30x more than Polygon quickswap – 95 swaps 16x more than Avax Trader Joe – 176 swaps 57x more than Celo Ubeswap – 50 swaps 160x more than Ethereum uniswap v2 – 18 swaps 3.5x more than all of them combined – 807 swaps
If anybody can do this experiment for other popular Layer 1s like Tezos, NEAR, or Elrond or the Layer 2s on ETH, I would love to include them here.
Please be respectful in the comments 🙂