Skip to content

Instantly share code, notes, and snippets.

@arkadijs
Last active July 30, 2025 21:27
Show Gist options
  • Save arkadijs/82206d2635f917c6d47c4cbb9e535783 to your computer and use it in GitHub Desktop.
Save arkadijs/82206d2635f917c6d47c4cbb9e535783 to your computer and use it in GitHub Desktop.
Relative CPU speed with GCC

Compact x86 32-bit Linux v6 kernel compile times on tmpfs storage.

CPU Year Cores Threads Freq, GHz Boost RAM MT/s Elapsed, sec Pts.1
1 ²ⁿ AWS Graviton4 2023 2x 96 192 2.7 DDR5 5600 17 8813
2 Ryzen 9 9950X 2024 16 32 4.3 4.7 DDR5 6000 60 4512
3 ²ⁿ Xeon Gold 6230 2019 2x 20 80 2.1 2.4 R.DDR4 2666 63 6048
4 ²ⁿ Xeon E5-2630 v4 2016 2x 10 40 2.2 2.4 DDR4 2133 126 6048
5 Xeon W-11955M 2021 8 16 2.6 24.3 DDR4 2666 141 4850
6 Xeon E5-2680 v4 2016 14 28 2.4 2.9 DDR4 2133 192 6451
7 Ryzen 5 5600H 2021 6 12 3.3 DDR4 3200 271 5366
8 Xeon X56603 2010 6 12 2.8 3.0 DDR3 1333 454 7627
9 Core i7-2600 2011 4 8 3.4 3.5 DDR3 1333 568 7952
10 Core i5-4570 2013 4 4 3.2 3.4 DDR3 1600 610 8296
11 RK35884 2022 8 8 2.4 LPDDR4x 4266 5698 11726
12 N150 2025 4 4 0.8 2.9 DDR4 3200 719 8340
13 Phenom II X4 965 2009 4 4 3.4 DDR2 800 765 10404
14 Core 2 Quad Q6600 2007 4 4 2.4 DDR3 1066 1168 11213
15 Athlon X4 750K 2012 2 4 3.4 3.7 DDR3 1600 1678 12417
16 StarFive JH7110 2022 4 4 1.5 LPDDR4 2600 3145 18870
17 Athlon 64 3000+ 2004 1 1 1.8 DDR 400 67248 13046
18 Athlon XP 3200+ 2003 1 1 2.2 DDR 400 78290 18238

Footnotes

  1. "Points": frequency * num cores * time; lower is better, divide the numbers to get relative core speed normalized by frequency.

  2. Thermally throttled; 4.5GHz max boost.

  3. Core i7 1st gen refresh, Westmere-EP LGA1366 tripple memory channel.

  4. Orange Pi 5 RK3588S with 4x Cortex-A76 2.4GHz, 4x Cortex-A55 1.8GHz and a cooling fan attached. Powered by a 5V 2A USB phone charger.

  5. Open-air passive cooling time is 821 sec with A55 cores throttled to 1.2-1.4GHz; no frequency data for A76.

  6. Would be 7096 seconds at 100% CPU utilization.

  7. 8205 seconds of CPU time.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment