AMD EPYC 4004 Benchmarks: Outperforming Intel Xeon E-2400 With Performance, Efficiency & Value

Written by Michael Larabel in Processors on 21 May 2024 at 09:00 AM EDT. Page 6 of 16. 22 Comments.
ScyllaDB benchmark with settings of Test: Writes. EPYC 4564P was the fastest.
ScyllaDB benchmark with settings of Test: Writes. EPYC 4564P was the fastest.
ScyllaDB benchmark with settings of Test: Writes. EPYC 4564P was the fastest.

Moving onto some database workloads, first up was the NoSQL ScyllaDB. The Xeon E-2488 did outperform the 8-core EPYC 4344P / 4364P processors but was no match for the higher core count SKUs. With this ScyllaDB benchmark not being too taxing on all of the cores, the power efficiency for the Intel Raptor Lake server processor was aligned with the AMD Zen 4 CPUs.

ScyllaDB benchmark with settings of Test: Writes. EPYC 4564P was the fastest.

But when looking at the performance-per-dollar is where the Xeon E-2488 outright fails at being competitive to AMD's new offerings.

DuckDB benchmark with settings of Benchmark: IMDB. EPYC 4564P was the fastest.
DuckDB benchmark with settings of Benchmark: IMDB. EPYC 4564P was the fastest.
DuckDB benchmark with settings of Benchmark: IMDB. EPYC 4564P was the fastest.
DuckDB benchmark with settings of Benchmark: IMDB. EPYC 4564P was the fastest.

It was a similar story with DuckDB of AMD offering SKUs able to accommodate much higher performance, the Intel Xeon E-2488 peak power consumption being much higher, and the entire AMD EPYC 4004 line-up offering better value.

Apache IoTDB benchmark with settings of Device Count: 500, Batch Size Per Write: 100, Sensor Count: 500, Client Number: 100. EPYC 4584PX was the fastest.
Apache IoTDB benchmark with settings of Device Count: 500, Batch Size Per Write: 100, Sensor Count: 500, Client Number: 400. EPYC 4584PX was the fastest.
Apache IoTDB benchmark with settings of Device Count: 500, Batch Size Per Write: 100, Sensor Count: 800, Client Number: 400. EPYC 4584PX was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, First Run / Cold Cache. EPYC 4584PX was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 4484PX was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 4484PX was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 4484PX was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 4484PX was the fastest.
MariaDB mariadb-slap benchmark with settings of Clients: 64. EPYC 4564P was the fastest.
MariaDB benchmark with settings of Test: oltp_write_only, Threads: 64. EPYC 4564P was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 64. EPYC 4564P was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 64. EPYC 4564P was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 64. EPYC 4564P was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 64. EPYC 4564P was the fastest.

While running MariaDB was some of the best results seen out of the Xeon E-2488 relative to the EPYC 4004 line-up but still struggling in value and power efficiency.

Related Articles