About Update benchmark Setting

[rainmaple]

if i want to use all params in validation_params.csv , which was producted by using update streams. How many operation_count i need to set when i set up benchmark.properties?

[szarnyasg]

Hi @rainmaple,

• When you run in benchmark mode with benchmark.properties, the framework does not use validation_params.csv.
• What is the goal with your run? If you would like to conduct a benchmark that would pass an audit, you have to run it such that the warmup lasts for at least 0.5 hours and the benchmark runs for at least 2 hours. For this, getting the correct operation_count is currently done by trial and error. There's an old issue about allowing users to specify a time range (Run benchmark based on total time, rather than operation count ldbc_snb_interactive_v1_driver#38) but this has not been addressed yet.

[szarnyasg]

Feel free to open an issue if you have any further questions.

[rainmaple]

As far as I know, all the update parameters generated by validation file are full update parameters, that is, if a test uses all the parameters contained in the validation file, it is a full update load, then how can I set it if I want to run a full update test。 发自我的iPhone

…

------------------ Original ------------------ From: Gabor Szarnyas ***@***.***> Date: Tue,Feb 15,2022 11:50 PM To: ldbc/ldbc_snb_interactive ***@***.***> Cc: erchuan ***@***.***>, Mention ***@***.***> Subject: Re: [ldbc/ldbc_snb_interactive] About Update benchmark Setting (Issue#236) Closed #236. — Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you were mentioned.Message ID: ***@***.***>

[szarnyasg]

The script validate.sh will take a validation_params.csv file and run its workload (reads/updates) sequentially.