Use IBM Watsonx.data and IBM Db2 Warehouse on 5th gen Intel Xeon processors to speed up insights.
Over the past 25 years, there have been notable improvements in database speed as a result of IBM and Intel’s long-standing cooperation. The revelation that the chip maker Intel Gaudi 3 AI accelerators would be made available as a service on IBM Cloud has elevated this partnership to new heights. In with joint venture, this endeavour marks yet another significant milestone.
Furthermore, IBM’s internal testing indicates that the performance of IBM Watsonx.data and Db2 Warehouse might be greatly improved by integrating the most recent generation of Intel Xeon Scalable processors with Intel software. Whether on-premises or in the cloud, businesses want to manage important tasks with efficiency and scalability. Businesses may effectively manage complicated operations across a range of sectors with speed and accuracy by utilising this potent combination.
IBM Products
IBM Watsonx.data
A new open, hybrid, and controlled data lake house designed for workloads including data, analytics, and artificial intelligence is called IBM Watsonx.data. Among the main highlights are:
- Reducing analytics expenses through the use of analytical engines such as Presto and Spark and less expensive storage
- Offering a consistent picture of your data across hybrid cloud environments by taking an open and adaptable approach.
IBM Db2 Warehouse
With its sophisticated MPP column-store technology and intelligent workload management for quick data intake and high concurrency queries, IBM Db2 Warehouse is a cloud-native data warehouse designed for critical, analytical applications. With support for open formats, it makes data sharing easier. It also connects with Watsonx.data to provide a unified analytics and AI perspective. accessible on-premises for hybrid data management systems or as SaaS on AWS or IBM Cloud.
Intel Technologies & Products
4th & 5th Gen Intel Xeon Scalable Processors
The 5th Gen Intel Xeon scalable processor, which shares an architectural platform with the 4th Gen, has silicon-based security, higher performance and performance per watt, and lower TCO. Memory-bound and latency-sensitive applications benefit from speedier memory and bigger last-level cache.
Intel Advanced Vector Extensions 512 (Intel AVX-512)
Single Instruction several Data (SIMD) instructions like the Intel AVX-512 may handle more data per instruction and perform several operations. Faster results and an improved user experience may be achieved with this increased processing capability, which can easily handle complicated queries and analytics jobs. Watsonx.data with Db2’s optimization using Intel AVX-512 technology gives an advantage by generating high performance that leads to short time to insight, which is needed by organizations that continue to require faster data processing speeds for real-time decision-making.
Enhancing Efficiency with Open-Source Software: Spark, Presto, and Prestissimo
The two main query engines for Watsonx.data and Db2 are Presto and Spark.
The Presto query engine makes use of Intel’s garbage collection, vectorisation, and Java Virtual Machine (JVM) optimisations.
Built with the Velox library, Prestissimo is Presto’s next-generation query engine that uses C++ and SIMD instructions. Prestissimo’s benefits include a significant performance gain and the avoidance of Java trash collection and JVM performance issues. Along with IBM, Intel is a member of the Presto foundation and has provided AVX-512 optimizations to help query workers take use of vectorization.
Workload Comparison: Using 5th Gen Intel Xeon Scalable CPUs to increase performance
The comparison of 5th Gen Intel Xeon Scalable processors with their predecessors in Watsonx.data and Db2 performance testing yields genuinely revolutionary results. The IBM Big Data Insights (BDI) workload provides a realistic simulation that replicates intricate retail settings, illuminating the exceptional speed and agility of the most recent generation of Intel Xeon processors under demanding circumstances.
In order to precisely evaluate and compare processor execution times, a strong evaluation framework was developed by exposing the testing setup to 16 concurrent users at a significant 3TB scale factor. In addition to showing observable improvements in performance indicators, this highlights how important cutting-edge technology is to transforming database processes. The importance of such comparison assessments in directing future hardware expenditures towards the best efficiency and productivity gains increases as companies depend more and more on data-driven insights for strategic decision-making.
Results:
IBM watsonx.data
Using the Big Data Insights (BDI) workload, it compared Watsonx.data operating on a single node across four generations of Intel Xeon processors, from the second to the fifth.
With a query throughput that is up to 2.7 times higher than that of the 2nd Gen Intel Xeon scalable processor, the 5th Gen Intel Xeon 8592+ processor stands out. It is also 1.75 times faster than the 3rd generation Intel Xeon 8380 CPU and 1.09 times faster than the 4th generation Intel Xeon 8490H processor.
IBM Db2 Warehouse
Significant findings have been obtained from our study of Db2 Warehouse performance on a single node using the BDI workload over 4th gen of Intel Xeon processors, from the 2nd Gen to the most recent 5th Gen Xeon processor. The Queries per Hour (QpH) attained by each processor is displayed in the graph.
Compared to the 2nd Gen Intel Xeon 8280M CPU, the 5th Gen Intel Xeon 8592+ processor has a query throughput that is up to 2.5X higher. It is also 1.15X better than the 4th Gen Intel Xeon 8490H CPU and 1.64X better than the 3rd Gen Intel Xeon 8380 processor.
In conclusion
Superior performance is provided by 5th Gen Intel Xeon scalable CPUs, as demonstrated here. This workload makes use of the processor’s greater memory bandwidth, more cores, and superior design. Faster analytics query response times and increased throughput to accommodate more concurrent users would greatly benefit customers. Users want faster insights and cost reductions, and this delivers on both.
