https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27233#M1561 <P>I have a dataset about 5 million rows with 14 features and a binary target. I decided to train a pyspark random forest classifier on Databricks. The CPU cluster I created contains 2 c4.8xlarge workers (60GB, 36core) and 1 r4.xlarge (31GB, 4core) driver. The GPU cluster I created contains 3 g4dn.4xlarge (64GB, 16cores) nodes, 2 as workers and 1 as driver. The hourly costs are very similar. I assumed that GPU cluster would outperform since random forest is an algorithm good for parallel computing, while the result kinda shocked me that the GPU cluster trained the model near 5 times slower than the CPU cluster. Is there anything I misunderstood about GPU acceleration or is it just not used for <A href="https://pyspark.ml" alt="https://pyspark.ml" target="_blank">pyspark.ml</A> modules?</P> Fri, 14 Oct 2022 17:07:02 GMT zzy 2022-10-14T17:07:02Z https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27233#M1561 <P>I have a dataset about 5 million rows with 14 features and a binary target. I decided to train a pyspark random forest classifier on Databricks. The CPU cluster I created contains 2 c4.8xlarge workers (60GB, 36core) and 1 r4.xlarge (31GB, 4core) driver. The GPU cluster I created contains 3 g4dn.4xlarge (64GB, 16cores) nodes, 2 as workers and 1 as driver. The hourly costs are very similar. I assumed that GPU cluster would outperform since random forest is an algorithm good for parallel computing, while the result kinda shocked me that the GPU cluster trained the model near 5 times slower than the CPU cluster. Is there anything I misunderstood about GPU acceleration or is it just not used for <A href="https://pyspark.ml" alt="https://pyspark.ml" target="_blank">pyspark.ml</A> modules?</P> Fri, 14 Oct 2022 17:07:02 GMT https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27233#M1561 zzy 2022-10-14T17:07:02Z https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27235#M1563 <P>In many cases, you need to adjust your code to utilize GPU. </P> Thu, 20 Oct 2022 12:40:36 GMT https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27235#M1563 Hubert-Dudek 2022-10-20T12:40:36Z https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27234#M1562 <P>Hi @Simon Zhang​&nbsp;, could you please go through this: <A href="https://www.databricks.com/session/gpu-support-in-spark-and-gpu-cpu-mixed-resource-scheduling-at-production-scale" alt="https://www.databricks.com/session/gpu-support-in-spark-and-gpu-cpu-mixed-resource-scheduling-at-production-scale" target="_blank">https://www.databricks.com/session/gpu-support-in-spark-and-gpu-cpu-mixed-resource-scheduling-at-production-scale</A> and let us know if it addresses your concern?</P> Tue, 18 Oct 2022 12:56:40 GMT https://community.databricks.com/t5/machine-learning/why-is-gpu-accelerated-node-much-slower-than-cpu-node-for/m-p/27234#M1562 Debayan 2022-10-18T12:56:40Z