Summary

Companies should always be looking for ways to better serve their customers. Customers are overwhelmed with information and often make buying decisions based on existing relationships. Companies looking to expand their relationships with customers can benefit from combining Machine Learning technologies with Data Mining to better understand their customers’ needs and to tailor their offerings to those needs.

Earlier this year, Dell and Intel conducted testing to determine how the new PowerEdge Server family utilizing Intel^® 4th Generation Xeon^® Scalable Processors could improve a company’s Data Mining efforts with Machine Learning technologies. 

HiBench is a big data benchmark suite that helps evaluate different big data frameworks in terms of speed, throughput, and system resource utilizations.  Part of the HiBench framework focuses on Machine Learning and utilizes Bayesian Classification and K-Means Clustering to effectively measure the relative performance of systems in a Machine Learning environment. The information below highlights the performance differences between a Dell PowerEdge R750 server with 3rd Generation Intel^® Xeon^® Scalable processors compared to the new Dell PowerEdge R760 with 4th Generation Intel^® Xeon^® Scalable processors.

All testing was conducted in Dell Labs by Intel and Dell Engineers in January of 2023.

Solution Overview

One of the primary benefits of the new 4th Generation Intel^® Xeon^® Scalable processors is core count. The previous generation of processors offered a maximum of 40 cores while the new processor family scales up to 56 cores. For the testing outlined in this report, we decided to use the new Intel^® Xeon^® Platinum 8470 processor which provides 52 cores.  For the previous generation processor, we chose the Intel^® Xeon^® Platinum 8380 which provides 40 cores.

In addition, to increased core count, the 4th Generation processors also support faster memory. The Dell R750 system we tested were configured with 512GB of memory (16x32GB DDR4) running at 3200MT/s. The new Dell R760 system was also configured with 512GB of memory (16x32GB DDR5) which operates at 4800MT/s.

Our testing utilized the HiBench K-Means elements of the test. This Algorithm aims to partition n observations into k clusters as shown in the graphic below:

Methodology

Each system was configured with the same number of processors, memory, and the configuration of hard drives. Each test bed was then subjected to two “warm up” cycles prior to running three iterations of the benchmark. The results for each test were averaged to measure processing time.

Hardware Configurations tested 

Software Configuration 

Test Results 

Key takeaways:

• 78% performance gain with the 4th Generation Intel^® Xeon^® 8470 compared to 3rd Generation Intel^® Xeon^® 8380 for Spark K-Means algorithm using OpenBlas library
• 4th Generation Intel^® Xeon^®Scalable processors benefits are results of:
  1. Innovative CPU microarchitecture providing up to a 37% performance boost
  2. Increased Parallelism (30% more cores) 

Conclusion

Implementing Machine Learning technologies with Big Data can help Companies better serve their customers. As shown in the testing above, the new Dell PowerEdge R760 with 4th Generation Intel^® Xeon^® Scalable processors can significantly reduce processing times leading to faster decision making.