Intel Raptor Lake 0x129 CPU Microcode Performance Impact On Linux

Written by Michael Larabel in Processors on 11 August 2024 at 01:16 PM EDT. Page 3 of 4. 26 Comments.
MariaDB mariadb-slap benchmark with settings of Clients: 32. Pre was the fastest.
MariaDB mariadb-slap benchmark with settings of Clients: 64. Pre was the fastest.

The MariaDB performance was also appearing to take a slight performance hit from the new BIOS / microcode.

WireGuard + Linux Networking Stack Stress Test benchmark with settings of . Pre was the fastest.

The biggest change in performance observed was for the WireGuard networking/VPN benchmark with around a 10% hit to the performance from the new microcode.

Perl Benchmarks benchmark with settings of Test: Pod2html. Pre was the fastest.
Perl Benchmarks benchmark with settings of Test: Interpreter. Pre was the fastest.

When running some scripting/interpretive single-threaded benchmarks, the run on the new CPU microcode/BIOS was beginning to indicate some regressions.

PyBench benchmark with settings of Total For Average Test Times. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: go. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: chaos. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: float. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: nbody. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: pathlib. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: raytrace. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: xml_etree. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: gc_collect. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: json_loads. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: crypto_pyaes. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: async_tree_io. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: regex_compile. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: python_startup. Post was the fastest.
PyPerformance benchmark with settings of Benchmark: asyncio_tcp_ssl. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: django_template. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: asyncio_websockets. Pre was the fastest.
PyPerformance benchmark with settings of Benchmark: pickle_pure_python. Pre was the fastest.

And then against all the Python scripting benchmarks run, it was a clear sweep that the new BIOS/microcode was trending lower on this Intel Core i9 14900K desktop.

Related Articles
Intel Xeon 6980P SNC3 vs. HEX Clustering Mode Performance
Intel Xeon 6980P vs. AMD EPYC Power Efficiency / Performance-Per-Watt Benchmarks
Intel Xeon 6980P "Granite Rapids" Linux Benchmarks
Intel Launches Xeon 6900P Series "Granite Rapids" Processors
AMD AGESA PI 1.2.0.2 Performance With The Ryzen 9 9950X On Linux
AMD Ryzen 5 9600X & Ryzen 7 9700X Linux Performance With 105 Watt cTDP