EIP-2028: Calldata gas cost reduction


This is to discuss the EIP I am currently creating. Will fill this out later

1 Like

The gas per non-zero byte is reduced from 68 to TBD. Gas cost of zero bytes is unchanged.

This EIP proposes a change to the cost of CALLDATA but doesn’t specify the new cost. It proposes a model with certain parameters to be considered when coming up with the new value.

I think there should be some kind of initial ballpark figure given to kick start discussions. Is it a 10% reduction? Is it a 50% reduction? A 90% reduction?

Or at least giving a ballpark figure for a time/effort needed to do the simulation using the models. Is this work feasible within the Istanbul timeline?

1 Like

I’d like to know how the gas costs for TXZERODATA and TXNONZERODATA were choosen in the first place. Why is there a difference in cost for zeros and non zeros?

If I would want to send a boolean (0x00 or 0x01) to the chain it would be cheaper to encode 0x01 as 0x0000. This seems weird to me.

For costs itself I don’t see why there should be a difference in data aswell. To me it seems that the data gets included (in the end) in one of the block hashes (transactionRoot?) but then it makes no sense to me why nonzero data has different cost. Storing this data would also make no difference to me? Unless I’m missing some trivial trick which you can do with this data. I’d like to learn why the cost is different in the end =)