hi,
We are currently revamping our fme server infrastructure. We're are going from a 2 physical servers on windows with two engines on both to 4 virtual server with one engine on redhat. That's a pretty big change.
We would like to compare performance before going to production with that configuration.
Is there any benchmark workspace available for this?
Thanks for your answers.
With kind regards.
Vincent.
You can compare the run time in the logging. Just take some workspaces and run them on the old and new environment. Look at the job-logs how long they take.
In addition to what @stalknecht says, I'd recommend making sure the workspaces you use are taking a long time, at least several minutes, to run. If you have a simple one that, for example, runs in 1.4 seconds on the old server and 1 second on the new one it's hard to get any meaningful results out of it. That 0.4 seconds time difference may just be a random fluctuation or due to other external factors.
It would be nice if we could come up with a meaningful benchmark for FME, both Desktop and Server. Maybe make it a challenge? (pinging @Mark2AtSafe)
In this article "https://blog.safe.com/2017/02/its-2017-introducing-the-next-generation-of-fme-cloud-instances/" it is mentioned that “Based on our benchmarking
tests, we think the vast majority of Starter users will benefit from this
change. »
I'd like to know how did Safe to do the benchmarking and if they can share something with us?
Thanks.
Vincent.
In this article "https://blog.safe.com/2017/02/its-2017-introducing-the-next-generation-of-fme-cloud-instances/" it is mentioned that “Based on our benchmarking
tests, we think the vast majority of Starter users will benefit from this
change. »
I'd like to know how did Safe to do the benchmarking and if they can share something with us?
Thanks.
Vincent.
In addition to what @stalknecht says, I'd recommend making sure the workspaces you use are taking a long time, at least several minutes, to run. If you have a simple one that, for example, runs in 1.4 seconds on the old server and 1 second on the new one it's hard to get any meaningful results out of it. That 0.4 seconds time difference may just be a random fluctuation or due to other external factors.
It would be nice if we could come up with a meaningful benchmark for FME, both Desktop and Server. Maybe make it a challenge? (pinging @Mark2AtSafe)
What about a Workspace reading/writing data stored locally in FFS file to remove any performance issues from network or particular file formats?