RTX4090 we're using here is owned by Zama. So we don't pay an
hourly rate to AWS per se. But in ordrer to compute throughput on
benchmarks results, the parser needs a numeric value corresponding
to the hardware used. Ops-per-dollar metric is not really used
today conversely ops-per-seconds is.
In the end we use an approximation of the cost for electrical
consumption.
Here we benchmark a fixed number of PBS with boolean and shortint
flavors on AWS EC2 instance. Once measurements are done, we compute
the number of operations per millisecond and also operations per
dollar we can perform for a given set of cryptographic parameters
and EC2 instance type. Data are then set to Slab that in turn send
them to a database to be plotted in Grafana.
