A Look At The AMD EPYC Performance On The Amazon EC2 Cloud

With Darktable for RAW image processing, the basic m5a.xlarge instance was faster than the Intel m5.xlarge instance, but with the more powerful instances, Intel ended up forming a large lead.

With the TTSIOD Renderer was another case of the m5a.xlarge instance in fact being much faster than the m5.xlarge, including with the 2xlarge size. But With 4xlarge and 12xlarge, the Intel Xeon performance came out ahead.

While on a performance-per-dollar basis, the AMD EPYC instances were competitive with the Intel instances.

With the PostgreSQL buffer test the size of the database test is determined based upon the core count and RAM size, which explains why the smaller instances ended up having slightly better performance, but the function of this testing anyhow is for comparing the Xeon vs. EPYC cloud instances. In this PostgreSQL read-write test, the EPYC performance was coming out on par with the Intel M5 instances.

The performance-per-dollar was also comparable between the Intel and AMD instances on the Amazon Elastic Compute Cloud.