Damir Vodenicarevic

> [perf.txt](https://github.com/massalabs/massa/files/9631362/perf.txt) That's an old one (has secp256k1 stuff). Here is the new one: [perf.txt](https://github.com/massalabs/massa/files/9632446/perf.txt)

Perf for 2 minutes under 1000 txps load (ops sent by another server), compiled with cargo build --release: https://we.tl/t-fQI5O7jQB1 * sudo apt install hotspot * unzip file * open file...

This can be interesting to revisit after Storage is used in the whole codebase

> wondering if we should remove src files from main repo? @damip I guess yes since otherwise it would be a duplicate (right ?)

> > To counter it, the only option I see for now is to drop operations whose validity range starts more than 1 period in the future. > > If...

@adrien-zinger Good proposal. Here are some of my thoughts. 1- if we're only looking at a batch of ops at every iteration, we cannot account for the whole fee spending...

Let's prioritize memory usage analysis.

Didn't manage to reproduce it on the tesnet with heaptrack running.  

> @damip should we make `max_gas` an `Amount` also while we are at it? no, gas should never be made into an Amount (see my comments in the PR)

> @damip Thanks for the clarification! So the `fee` in all cases would come from that found on the operation, except for the `send_message` ABI where we'd have to add...