{{ post.title }}
글 편집
{{ post.author.name }}
Posted on
| Version | {{ post.target_version }} | Product |
{{ product.name }}
|
|---|---|---|---|
| Tutorial/Manual | {{ post.manual_title }} | Attached File | {{ post.file.upload_filename }} |
Single (float) precision vs double precision
In CAE software, you often see the terminologies such as float and double or single and double precision.
- Float = Single Precision
- Double = Double Precision
What you need to remember are:
1. float:
- number of the significant number: about 7
- the maximum value: about 3.4e+38.
- single precision.
- 32 bit value
2. double:
- number of the significant number: about 16
- the maximum value: about 1.8e+308
- double precision
- 64 bit value
So, if you need more than 7 significant number or bigger value than 3.4e+38, then you must use double (double precision).
In other cases, float (single precision) can be better choice and you can save the memory usage or file size (which means you can get faster simulation speed or file generation performance.)
For the more details, please refer to the following link.
https://docs.microsoft.com/en-us/cpp/c-language/type-float?view=vs-2019
For your information,
In RecurDyn
, you can find those terminologies in RFI optimizer or Output File Generator dialog.
In Particleworks , you can choose Single or Double in Solving Option.
Simply speaking, float (=Single precision) is less accurate than double (=double precision) but float uses smaller memory or generates smaller size of file than double. The less memory or smaller file size means the simulation speed or file generation performance can get faster.
*Single menas Single precision, and Double means Double precision in the above image.
