Howdy, Stranger!

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


Dedipath OpenVZ - E5-2660 vs Ryzen 3900X
New on LowEndTalk? Please Register and read our Community Rules.

Dedipath OpenVZ - E5-2660 vs Ryzen 3900X

KebabKebab Member

So currently dedipath offers huge discounts of 65%, I saw that they offer 1GB OpenVZ Ryzen NVME VPS for cheap so I took that, but when I ran yabs it showed AMD Ryzen 9 3900X 12-Core @ 3800.000 MHz geekbench Single Core - 537, Multi Core - 938.

I expected more from it, so I went and bought their 1GB OpenVZ SSD VPS (intel) which is 2 dollars / month cheaper (without discount).

Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz scored Single Core - 568, Multi Core - 1023
Turns out the cpu from 2012 is faster than server from 2020.
So I wonder is this some openvz limitation, someone is hogging all the resources or is this dedipath limiting it.

ryzen

# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
#              Yet-Another-Bench-Script              #
#                     v2022-08-20                    #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #

Sat Nov 19 10:05:51 EST 2022

Basic System Information:
---------------------------------
Uptime     : 0 days, 16 hours, 0 minutes
Processor  : AMD Ryzen 9 3900X 12-Core Processor
CPU cores  : 2 @ 3800.000 MHz
AES-NI     : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM        : 1.0 GiB
Swap       : 256.0 MiB
Disk       : 29.9 GiB
Distro     : Ubuntu 20.04.3 LTS
Kernel     : 5.4.0

fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4k            (IOPS) | 64k           (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 46.66 MB/s   (11.6k) | 63.39 MB/s     (990)
Write      | 46.72 MB/s   (11.6k) | 63.84 MB/s     (997)
Total      | 93.39 MB/s   (23.3k) | 127.23 MB/s   (1.9k)
           |                      |
Block Size | 512k          (IOPS) | 1m            (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 102.42 MB/s    (200) | 72.64 MB/s      (70)
Write      | 107.86 MB/s    (210) | 77.47 MB/s      (75)
Total      | 210.29 MB/s    (410) | 150.11 MB/s    (145)

iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed
                |                           |                 |
Clouvider       | London, UK (10G)          | 47.0 Mbits/sec  | 26.1 Mbits/sec
Online.net      | Paris, FR (10G)           | 45.0 Mbits/sec  | 24.9 Mbits/sec
Hybula          | The Netherlands (40G)     | 43.7 Mbits/sec  | 24.2 Mbits/sec
Uztelecom       | Tashkent, UZ (10G)        | 3.91 Mbits/sec  | 6.72 Mbits/sec
Clouvider       | NYC, NY, US (10G)         | 1.26 Gbits/sec  | 440 Mbits/sec
Clouvider       | Dallas, TX, US (10G)      | 68.8 Mbits/sec  | 44.3 Mbits/sec
Clouvider       | Los Angeles, CA, US (10G) | 49.0 Mbits/sec  | 30.0 Mbits/sec

iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed
                |                           |                 |
Clouvider       | London, UK (10G)          | 24.8 Mbits/sec  | 25.6 Mbits/sec
Online.net      | Paris, FR (10G)           | 23.6 Mbits/sec  | 23.8 Mbits/sec
Hybula          | The Netherlands (40G)     | 22.4 Mbits/sec  | 22.5 Mbits/sec
Uztelecom       | Tashkent, UZ (10G)        | busy            | 10.2 Mbits/sec
Clouvider       | NYC, NY, US (10G)         | 1.31 Gbits/sec  | 527 Mbits/sec
Clouvider       | Dallas, TX, US (10G)      | 41.8 Mbits/sec  | 44.0 Mbits/sec
Clouvider       | Los Angeles, CA, US (10G) | 28.2 Mbits/sec  | 28.8 Mbits/sec

Geekbench 5 Benchmark Test:
---------------------------------
Test            | Value
                |
Single Core     | 537
Multi Core      | 938
Full Test       | https://browser.geekbench.com/v5/cpu/18738934

intel

# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
#              Yet-Another-Bench-Script              #
#                     v2022-08-20                    #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #

Sat Nov 19 10:03:20 EST 2022

Basic System Information:
---------------------------------
Uptime     : 0 days, 15 hours, 23 minutes
Processor  : Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz
CPU cores  : 2 @ 2699.914 MHz
AES-NI     : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM        : 1.0 GiB
Swap       : 256.0 MiB
Disk       : 29.9 GiB
Distro     : Ubuntu 20.04.3 LTS
Kernel     : 5.4.0

fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4k            (IOPS) | 64k           (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 34.59 MB/s    (8.6k) | 104.39 MB/s   (1.6k)
Write      | 34.68 MB/s    (8.6k) | 104.94 MB/s   (1.6k)
Total      | 69.27 MB/s   (17.3k) | 209.33 MB/s   (3.2k)
           |                      |
Block Size | 512k          (IOPS) | 1m            (IOPS)
  ------   | ---            ----  | ----           ----
Read       | 208.39 MB/s    (407) | 224.55 MB/s    (219)
Write      | 219.46 MB/s    (428) | 239.51 MB/s    (233)
Total      | 427.85 MB/s    (835) | 464.07 MB/s    (452)

iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed
                |                           |                 |
Clouvider       | London, UK (10G)          | 861 Mbits/sec   | 429 Mbits/sec
Online.net      | Paris, FR (10G)           | 847 Mbits/sec   | 207 Mbits/sec
Hybula          | The Netherlands (40G)     | 843 Mbits/sec   | 776 Mbits/sec
Uztelecom       | Tashkent, UZ (10G)        | 5.02 Mbits/sec  | 155 Mbits/sec
Clouvider       | NYC, NY, US (10G)         | 878 Mbits/sec   | 894 Mbits/sec
Clouvider       | Dallas, TX, US (10G)      | 856 Mbits/sec   | 545 Mbits/sec
Clouvider       | Los Angeles, CA, US (10G) | 830 Mbits/sec   | 571 Mbits/sec

iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed
                |                           |                 |
Clouvider       | London, UK (10G)          | 819 Mbits/sec   | 347 Mbits/sec
Online.net      | Paris, FR (10G)           | 813 Mbits/sec   | 355 Mbits/sec
Hybula          | The Netherlands (40G)     | 811 Mbits/sec   | 678 Mbits/sec
Uztelecom       | Tashkent, UZ (10G)        | 63.6 Mbits/sec  | 205 Mbits/sec
Clouvider       | NYC, NY, US (10G)         | 859 Mbits/sec   | 875 Mbits/sec
Clouvider       | Dallas, TX, US (10G)      | 834 Mbits/sec   | 394 Mbits/sec
Clouvider       | Los Angeles, CA, US (10G) | 811 Mbits/sec   | 544 Mbits/sec

Geekbench 5 Benchmark Test:
---------------------------------
Test            | Value
                |
Single Core     | 568
Multi Core      | 1023
Full Test       | https://browser.geekbench.com/v5/cpu/18738828

Comments

  • @Kebab said:
    So currently dedipath offers huge discounts of 65%, I saw that they offer 1GB OpenVZ Ryzen NVME VPS for cheap so I took that, but when I ran yabs it showed AMD Ryzen 9 3900X 12-Core @ 3800.000 MHz geekbench Single Core - 537, Multi Core - 938.

    I expected more from it, so I went and bought their 1GB OpenVZ SSD VPS (intel) which is 2 dollars / month cheaper (without discount).

    Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz scored Single Core - 568, Multi Core - 1023
    Turns out the cpu from 2012 is faster than server from 2020.
    So I wonder is this some openvz limitation, someone is hogging all the resources or is this dedipath limiting it.

    Openvz is kind of bad anyway, have you tried kvm?

  • Most likely ryzen servers are oversold and so these scores in geekbench are low.

  • cpu in vps is just a marketing gimmick. First, You can't hog the CPU anyway so what app are you runnning that requires a ryzen? Its like buying a ferrari and you drive in under 100km/h speed limit. Just use a Kia and you get there in the same amount of time for less money. Second, an empty 2660 node is always faster than a loaded 3900x node, which means cpu in a vm means nothing unless the load on the node is also guaranteed.

  • @jonatha said:
    Openvz is kind of bad anyway, have you tried kvm?

    Not from them, no.
    I have few KVMs from v6node with Ryzen 5 3600 and single core score is 1232

  • add_iTadd_iT Member
    edited November 19

    Looking at the score maybe You only got 2 core thread (logical core) + heavy overload node

  • cybertechcybertech Member
    edited November 20

    @mosquitoguy said:
    cpu in vps is just a marketing gimmick. First, You can't hog the CPU anyway so what app are you runnning that requires a ryzen? Its like buying a ferrari and you drive in under 100km/h speed limit. Just use a Kia and you get there in the same amount of time for less money. Second, an empty 2660 node is always faster than a loaded 3900x node, which means cpu in a vm means nothing unless the load on the node is also guaranteed.

    just wait till 2660 gets loaded and then compare.

    would you rather squeeze 10 people in a ferrari or a Kia rushing from point to point in start/stop traffic?

  • jbilohjbiloh Administrator

    @Kebab you should submit a LowEndReview and make some $$$

  • @cybertech said:
    just wait till 2660 gets loaded and then compare.
    would you rather squeeze 10 people in a ferrari or a Kia rushing from point to point in start/stop traffic?

    The 2660 won't get loaded by virtue of the marketing and people who think like you. People will go for the ryzen for heavy loads and the people who get the 2660 are the light users who believe they don't need a fast cpu. So the ryzen will end up being slower by mob mentality, just as what the OP posted. :)

  • It highly depends on your neighbors.

Sign In or Register to comment.