The document provides an introduction to the Blender Benchmark platform, which allows users to run benchmarks on their systems and upload the anonymous results. It consists of a downloadable benchmark client and an open data portal where results are shared publicly. The benchmark collects system information during a test and users can anonymize personal details before uploading. The portal aims to host benchmark and other Blender-related data openly under defined open data principles, including making all public data available without restrictions.
Martal Group - B2B Lead Gen Agency - Onboarding Overview
Blender benchmark
1. Beginners guide: Introduction to Blender Benchmark.
Recently, Blender Benchmark, a new platform for collecting and presenting the results of hardware
and software performance tests, was introduced. This benchmark aims to provide an optimal
comparison between system hardware and installations and helps developers track performance
during Blender development.
The benchmark consists of two parts: a downloadable package that runs Blender and renders on
multiple product files and the Open Data Portal on blender.org, where the results are uploaded
(optional).
The Blender benchmark platform was developed with maximum focus on transparency and privacy.
In addition, only free and open source software (GNU GPL) was used and the test results are
anonymously distributed as public domain data – free for anyone to download and process.
This is to invite the entire Blender community to contribute the results of their performance tests
and create an open dataset for the entire CG industry.
How it works.
Users download the benchmark client and run one of the two benchmarks (“fast” or “complete”).
The benchmark collects information about the system, such as operating system, RAM, graphics
cards, CPU model, and information about the performance of the system during the execution of the
benchmark. The user can then share the result online on the Blender Open Data platform or store
the data locally.
To gain control over the data shared online, the benchmark result is first linked to the user’s Blender
ID and uploaded to mydata.blender.org, where the user can edit and anonymize the parts containing
personal data (Blender ID username and hostname). Currently, this information is removed by
default. No personal information is collected.
Blender Open Data Portal.
2. To visualize, share and explore the data, opendata.blender.org was developed. The data hosted on
the website are available under public domain, are updated in almost real time after each
benchmarking and are easy to process and well documented.
While hosting Blender benchmark results will be the main purpose of the Open Data Portal, it is
planned to host more data sets in the future. For example, information about downloads of blenders,
telemetry information, etc. Each data set published on the platform complies with the defined Open
Data principles and its collection is clearly communicated.
Principles of Blender Open Data.
As the first guideline for the definition of Blender Open Data, the developers were inspired by the
Eight Principles of Open Data. According to this, Blender Open Data should be the following:
• Completely. All public data is made available. Public data is data that is not subject to any
valid data protection, security or privilege restrictions.
• Primary. Data is collected at source, with the highest possible granularity, not in aggregate
or modified forms.
• Timely. The data will be made available as soon as possible to preserve the value of the
data.
• Accessible. The data is available to a wide range of users for different purposes.
• Machinable. The data is structured in a meaningful way to enable automated processing.
• Non-discriminatory. The data is accessible to everyone without the need for registration.
• Not uncopyright protected. The data is available in a format over which no company has
sole control.
• License-free. The data are not subject to any copyright, patent, trademark or commercial
law regulations. Appropriate privacy, security and privilege restrictions may be permitted.
• Online and free of charge. Information is not sensibly public if it is not available free of
charge on the Internet, at least not more than the marginal costs of reproduction.
• Permanent. The data should be available at a stable Internet location for an indefinite
period of time and in a stable data format as long as possible.
When dealing with benchmark data, publishers follow a data protection-conscious approach.
Thank you very much for your visit.