Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


How to increase CPU frequency ?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

How to increase CPU frequency ?

I have been using VPS for a long time and I have several of them. Today I tested Dedicated
servers from hetzner

When YABS is turned on I got This result

# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
#              Yet-Another-Bench-Script              #
#                     v2023-11-30                    #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #

Sat Dec 23 12:44:49 PM CET 2023

Basic System Information:
---------------------------------
Uptime     : 0 days, 0 hours, 2 minutes
Processor  : AMD Ryzen 9 3900 12-Core Processor
CPU cores  : 24 @ 2200.000 MHz
AES-NI     : ✔️ Enabled
VM-x/AMD-V : ✔️ Enabled
RAM        : 125.7 GiB
Swap       : 4.0 GiB
Disk       : 3.4 TiB
Distro     : Ubuntu 22.04.3 LTS
Kernel     : 5.15.0-89-generic
VM Type    : NONE
IPv4/IPv6  : ✔️ Online / ✔️ Online

IPv6 Network Information:
---------------------------------
ISP        : Hetzner Online GmbH
ASN        : AS24940 Hetzner Online GmbH
Host       : Hetzner
Location   : Nuremberg, Bavaria (BY)
Country    : Germany

fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/md2):
---------------------------------
Block Size | 4k            (IOPS) | 64k           (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 747.72 MB/s (186.9k) | 1.73 GB/s    (27.1k)
Write      | 749.70 MB/s (187.4k) | 1.74 GB/s    (27.3k)
Total      | 1.49 GB/s   (374.3k) | 3.48 GB/s    (54.4k)
           |                      |
Block Size | 512k          (IOPS) | 1m            (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 1.73 GB/s     (3.3k) | 1.77 GB/s     (1.7k)
Write      | 1.83 GB/s     (3.5k) | 1.89 GB/s     (1.8k)
Total      | 3.56 GB/s     (6.9k) | 3.66 GB/s     (3.5k)

iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
-----           | -----                     | ----            | ----            | ----
Clouvider       | London, UK (10G)          | 928 Mbits/sec   | 931 Mbits/sec   | 19.6 ms
Scaleway        | Paris, FR (10G)           | 933 Mbits/sec   | busy            | 21.5 ms
NovoServe       | North Holland, NL (40G)   | 938 Mbits/sec   | 935 Mbits/sec   | 10.7 ms
Uztelecom       | Tashkent, UZ (10G)        | 883 Mbits/sec   | 484 Mbits/sec   | 95.2 ms
Clouvider       | NYC, NY, US (10G)         | 871 Mbits/sec   | 610 Mbits/sec   | 83.7 ms
Clouvider       | Dallas, TX, US (10G)      | 844 Mbits/sec   | 271 Mbits/sec   | 121 ms
Clouvider       | Los Angeles, CA, US (10G) | 812 Mbits/sec   | 384 Mbits/sec   | 150 ms

iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
-----           | -----                     | ----            | ----            | ----
Clouvider       | London, UK (10G)          | 918 Mbits/sec   | 922 Mbits/sec   | 19.6 ms
Scaleway        | Paris, FR (10G)           | 922 Mbits/sec   | 921 Mbits/sec   | 15.5 ms
NovoServe       | North Holland, NL (40G)   | 925 Mbits/sec   | 923 Mbits/sec   | 10.7 ms
Uztelecom       | Tashkent, UZ (10G)        | 869 Mbits/sec   | 485 Mbits/sec   | 95.3 ms
Clouvider       | NYC, NY, US (10G)         | 866 Mbits/sec   | 342 Mbits/sec   | 83.8 ms
Clouvider       | Dallas, TX, US (10G)      | 834 Mbits/sec   | 339 Mbits/sec   | 120 ms
Clouvider       | Los Angeles, CA, US (10G) | 806 Mbits/sec   | 368 Mbits/sec   | 150 ms

Geekbench 6 Benchmark Test:
---------------------------------
Test            | Value
                |
Single Core     | 1780
Multi Core      | 10027
Full Test       | https://browser.geekbench.com/v6/cpu/4106163

YABS completed in 12 min 28 sec

This Is CPU info

CPU max MHz: 3100.0000
CPU min MHz: 2200.0000

.........

But when I turn on YABS it shows 2200.0000 MHz

The overall performance was " Multi Core | 10027 "

Technical support told me that the processor's maximum capacity is increased when needed "Thank you to hetzner for the support and wonderful servers "

Why did YABS not use maximum frequency?

Is the performance according to YABS for this server correct, or does it only represent a half of the actual performance of the server?

Is there a way to increase the frequency and is this recommended?

Comments

  • Looks at YABS source code in github on line 230. Its just return processor information from /proc/cpuinfo which is only show the processor base freq information, not the current processor freq.

    Thanked by 3loay aaxaa boot
  • @aaxaa said: Why did YABS not use maximum frequency?

    I think yabs shows different frequency with different runs and operating systems.

    Thanked by 1aaxaa
  • Check your current CPU Governor mode. Set it to Performance.

    After set, confirm after reboot if the change persists.

    Thanked by 2JeDaYoshi aaxaa
  • @nqservices said:
    Check your current CPU Governor mode. Set it to Performance.

    After set, confirm after reboot if the change persists.

    Thank you .. I Set it to Performance

    YABS Now

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2023-11-30                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Sat Dec 23 11:09:20 PM CET 2023
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 4 hours, 34 minutes
    Processor  : AMD Ryzen 9 3900 12-Core Processor
    CPU cores  : 24 @ 3100.000 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 125.7 GiB
    Swap       : 8.0 GiB
    Disk       : 1.7 TiB
    Distro     : Ubuntu 22.04.3 LTS
    Kernel     : 5.15.0-89-generic
    VM Type    : NONE
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    IPv6 Network Information:
    ---------------------------------
    ISP        : Hetzner Online GmbH
    ASN        : AS24940 Hetzner Online GmbH
    Host       : Hetzner
    Location   : Nuremberg, Bavaria (BY)
    Country    : Germany
    
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/md2):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 550.43 MB/s (137.6k) | 1.23 GB/s    (19.2k)
    Write      | 551.88 MB/s (137.9k) | 1.24 GB/s    (19.3k)
    Total      | 1.10 GB/s   (275.5k) | 2.47 GB/s    (38.6k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 1.22 GB/s     (2.3k) | 1.28 GB/s     (1.2k)
    Write      | 1.29 GB/s     (2.5k) | 1.37 GB/s     (1.3k)
    Total      | 2.51 GB/s     (4.9k) | 2.66 GB/s     (2.6k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
    -----           | -----                     | ----            | ----            | ----
    Clouvider       | London, UK (10G)          | 928 Mbits/sec   | 920 Mbits/sec   | 18.7 ms
    Scaleway        | Paris, FR (10G)           | busy            | busy            | 21.5 ms
    NovoServe       | North Holland, NL (40G)   | 938 Mbits/sec   | 935 Mbits/sec   | 10.7 ms
    Uztelecom       | Tashkent, UZ (10G)        | 880 Mbits/sec   | 572 Mbits/sec   | 97.9 ms
    Clouvider       | NYC, NY, US (10G)         | 870 Mbits/sec   | busy            | 83.8 ms
    Clouvider       | Dallas, TX, US (10G)      | 846 Mbits/sec   | 402 Mbits/sec   | 122 ms
    Clouvider       | Los Angeles, CA, US (10G) | 813 Mbits/sec   | 446 Mbits/sec   | 152 ms
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
    -----           | -----                     | ----            | ----            | ----
    Clouvider       | London, UK (10G)          | 918 Mbits/sec   | 914 Mbits/sec   | 18.7 ms
    Scaleway        | Paris, FR (10G)           | busy            | busy            | 15.6 ms
    NovoServe       | North Holland, NL (40G)   | 925 Mbits/sec   | 923 Mbits/sec   | 10.8 ms
    Uztelecom       | Tashkent, UZ (10G)        | 868 Mbits/sec   | 468 Mbits/sec   | 97.9 ms
    Clouvider       | NYC, NY, US (10G)         | 867 Mbits/sec   | 382 Mbits/sec   | 83.7 ms
    Clouvider       | Dallas, TX, US (10G)      | 833 Mbits/sec   | 387 Mbits/sec   | 122 ms
    Clouvider       | Los Angeles, CA, US (10G) | 807 Mbits/sec   | 316 Mbits/sec   | 152 ms
    
    Geekbench 6 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 1781
    Multi Core      | 10276
    Full Test       | https://browser.geekbench.com/v6/cpu/4112708
    
    YABS completed in 13 min 55 sec
    

    The performance is the same. This means that the server performs maximum performance when needed

  • @aaxaa Can you please share where to order this exact server from Hetzner ?

    Thanks.

  • @adilolv said:
    @aaxaa Can you please share where to order this exact server from Hetzner ?

    Thanks.

    It Is "SERVER AUCTION" You get it at an acceptable price

    This is the service link
    https://www.hetzner.com/sb

  • The one you did YABS on, is this one right?

  • rm_rm_ IPv6 Advocate, Veteran
    edited December 2023

    @aaxaa said: YABS Now

    As you can see your scores did not improve whatsoever, except now your server runs at full MHz all the time, consuming more power and possibly heating up more. In case of some kinds of loads this may cause it to throttle sooner, and actually reduce performance at prolonged loads. I suggest that you change it back to whatever it was (probably "ondemand").

    Thanked by 1aaxaa
  • @rm_ said:

    @aaxaa said: YABS Now

    As you can see your scores did not improve whatsoever, except now your server runs at full MHz all the time, consuming more power and possibly heating up more. In case of some kinds of loads this may cause it to throttle sooner, and actually reduce performance at prolonged loads. I suggest that you change it back to whatever it was (probably "ondemand").

    That's right thank you

    I returned it to ondemand

  • @adilolv said:
    The one you did YABS on, is this one right?

    Yes, it's the same. I chose FSN

  • AdvinAdvin Member, Host Rep
    edited December 2023

    YABS will not show you the turbo speeds

    If you want to check the real clocks, run a CPU stress test (maybe stress -c 24 or similar) in another session (or you can use screen/tmux/etc) and then run watch -n.1 "grep \"^[c]pu MHz\" /proc/cpuinfo" which will give you a live output of the CPU clock speeds of every core :)

    Looking at the average for a Ryzen 9 3900, it seems that your server is performing above average: https://browser.geekbench.com/processors/amd-ryzen-9-3900

    Thanked by 2tiendungdev aaxaa
Sign In or Register to comment.