Optimized Machine Learning Libraries For CPUS Exceed GPU Performance
Historically, many open source machine learning software packages have been optimized to run on GPUs… but new research suggests that when code is optimized for CPUs, performance can match and even exceed that of GPUs. Is your organization making hardware procurement decisions based on biased benchmark results? We’re taking a deep dive into processor issues, and exploring how new optimized libraries can cut away the benchmarking bias.
This 10 page, exclusive whitepaper covers:
- Why HPC leaders such as Kyoto University are choosing CPUs for machine learning
- Capabilities of the CPU-optimized DAAL and MKL-DNN libraries
- How developers are leveraging multicore processors to speed data analytics, training and prediction
- …and more!
Please register to download this special report, Optimized Machine Learning Libraries For CPUS Exceed GPU Performance, produced by Tabor Custom Publishing in conjunction with Datanami, and sponsored by Intel.