The only real world way to accurately calculate the power used in a sever is to measure the power consumed on the working server you want to review. Most of the name plate information is for a maximum configuration and some times is used for peek power usage.
The available manufactures data or calculators are a good guide but not accurate. If you are trying to calculate the heat output or the spare available power in your racks or data centre you can not relay on the estimates other than the rough guide. (I’m certain someone will tell me of an exception to the rule)
If you don’t have a actual server with the same configuration you will need to find one or use the estimate/calculator available, when install then measure for future info.
You will be surprised at the difference.
But be warned if you have a lot of servers you need to allow for the much higher starting current when starting a lot of servers together, just slowly start the servers in an staged approach with .5 on a second between each one. This will also stop a large inrush current damage to the server power supplies that can occur.
Also if you are analysing capacity it is best practice not to load the supply breakers above 90%, this will allow for harmonics issues, almost eliminates failures, increases reliability and lengthens the life of the breaker.