Any Talend benchmarking tool to compare computer configuration performances?

[Giova1]

Hello to all.

I'm new here, I use Talend in the context of my work (we have a paid licences and all) but I strongly suspect our work computers have very sub-optimal configurations, and I would like to compare Talend performances over different hardware configuration to see how we could optimize it for Talend.

My main concern is ram, almost all our computers have single channel ram, and I would like to know if dual channel would increase performances of Talend, and if yes, by how much.

So I can know if it worth "fighting" to get dual channel ram added in our hardware requirements, or if it doesn't worth it.

Just a "the studio feels smoother and some jobs faster" will have little to no effect, even if it's true.

So I would like to know if there is some "Talend-like" benchmarking tools (aka tools stressing a computer pretty much like Talend does) or some "Benchmarking job" you can run and use to see how much the configuration X outperforms the configuration Y.

If that doesn't exist, I'll try to make one, but if there is already one (probably better than what I could do by myself on my free time), I'd gladly use it.

(We do mostly load transform extract stuff)

Thank you in advance for your help 🙂

[Giova1]

Hello.

I wrote some jobs to see by myself in the end, looks like it doesn't change the performances much (A 2 min job is like 3 or 4 seconds shorter with dual channel).

Not impressive, now I'll try to see if it helps with compilation and/or studio reactivity, for that I will just use it and see if I can feel a difference.

Sadly there's no tool to measure that.

[Anonymous]

Sorry for the delay in responding here. This is an interesting subject that back when I was consulting I would have been able to give you all sorts of up to date stats on. However, more recently I have been focussing on Cloud and have not worried so much about the performance of my machine.....so long as it can complete the processing. I use a Mac and LOVE the memory handling of a Mac.

One thing I would advise testing is parallel processing. If you don't use parallel processing and you are interested in performance, try it. I suspect that you may see more of an improvement with parallel processing. The one thing to keep in mind with parallel processing is the number of CPUs vs processes rule. Think n-1 for the processes. If you have 6 processors, set up 5 parallel processes and start testing with that. 1 processor for each of the processors and a processor spare to control everything. This isn't always the most optimum config, but is a good starting point.

I would be interested to hear what sort of performance difference you see.

[Giova1]

Hello.

Thank you very much for your answer 🙂

I'll try this, we don't use much parallel executions but it could definitely help me to show that in some cases dual channel can really help, even if not always.

​We can't ask for Mac OS machines, but I do hope the person in charge of hardware orders in the company can at least provide us dual channel machines next time ...

He recently ordered a laptop for a coworker using talend intensively, the laptop has 2 slots of ram (unlike the 13" of my team) and the guy managed to order a 1x16GB config with a' empty slot while 2x8GB would cost no extra, as a hardware enthusiast that drives me crazy :(​

I'll dig into your suggestion to see if I can get metrics to show how dual channel can​ help in some use cases.

And one of my coworker is working in dual chan for a week, I'll ask for feedbacks about more general studio reactivity.

The studio can be pretty laggy sometimes on our laptops, I was wondering if dual channel could help but those have only 1 slot of sodimm so I can't try out on my machine.

​Thanks again 🙂