Is the "Aggregate Runtime" the same as the Compute-Hours cost?

Ask the Community
,
1

Is the “Aggregate Runtime” - which is visible in the Spark Details of a job - the same as the compute-hours cost of a build in RMA?

Specifically, will the aggregate runtime take into account if I exceed the memory-to-core ratio?

1 Like
2

No, memory to core ratio is taken into account in Foundry’s compute hours/compute usage formulas, but not into aggregate runtime which is a sum of all task durations in a spark job.

3

So if I want to get the cost in compute hours, can I just multiply the aggregate runtime with the memory coefficient?

(Trying to understand what a job costs without access to RMA.)

4

+1 to see the compute hour/RMA metric on an individual job

5

No, I guess you would need to apply the formulas described in the compute usage. There is examples too.

Optimizing and debugging pipelines • Spark • Understand compute usage • Palantir