Howdy, Stranger!

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


server-factory.com AMD EPYC VDS review
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.

server-factory.com AMD EPYC VDS review

I have been into VDS (Virtual Dedicated Server) for quite some time. A VDS is perfect for CPU-intensive tasks, such as video encoding and file conversion because you get fully dedicated resources to use however you like whereas only the disk and network are still shared. You don't have to worry about getting suspended for high-CPU usage either.

As you know, there are a few providers in the low-end market who offer such a service at an affordable price. This year I came across a new provider in this niche named "server-factory.com". I gave them a try, and I'm satisfied with their services so far. So I think I should share my experience with them on LET.

Anyway, I'm not affiliated with them. Just a normal customer like many of you. So this is not a paid review and no aff links. Here is my honest review which has both pros and cons.

It started from their deals below in March that were hard to resist. I have never seen any provider who offers an EPYC VDS starting with 4 GB RAM 50 GB NVMe SSD for around 2.5 euros a month anywhere before.

https://lowendtalk.com/discussion/178249/40-off-ryzen-epyc-virtual-dedicated-servers-in-nl-alipay-btc-cc-pp/p1

One thing to note is that their 1 CPU core means 2 threads. You get 1 dedicated thread (0.5 EPYC core) per 4 GB RAM. So you get 2 dedicated threads (1 full EPYC core but displayed as 2 CPUs) in the 8 GB RAM plan and so on. This was confusing at first for me.

Pros:
- Very affordable prices, especially for a VDS at around 2.5 euros per month. It's really hard to beat even on BFD.
- You get fully-dedicated resources for high-CPU intensive tasks. And it's 2nd gen EPYC which is powerful enough to do most tasks. New orders will likely be placed on an EPYC 7452 node like mine.
- RAID-10 storage for EPYC and RAID-1 for Ryzen. Both are fast NVMe SSD.
- Unmetered incoming traffic. Only outgoing traffic is counted. Both are at 1Gbps.
- Custom ISOs are supported. You can install any OS.
- Their servers are at Serverius which is an excellent data center in the Netherlands from my experience.
- Great uptime. I'm not using a monitoring tool, but I have never experienced any downtime when I use my VDS on a daily basis.
- Free one snapshot for backing up your VDS. It can be scheduled too. I just noticed this feature and it's worth mentioning.
- They're innovative and keep adding new services, such as IPv6 only VPS, NAT VPS, and storage VPS.
- @mxmla is a very nice and friendly person.

Cons:
- Only 1 location in the Netherlands, for now.
- They're a one-man provider and are still new in the market and need more time to prove to be an established one.
- Email-related ports are blocked. This limits a number of projects relying on sending out emails that can be deployed on the VDS. However, they have an email relay service via support. You have to let them know the domain names. This may not be as flexible as using an external email service, but it gets the issue fixed.
- Their website still needs a professional-looking landing page.

This is my AMD EPYC S yabs. I yabs this VDS occasionally and the results stay similar to day one. I bought this VDS annually in May and got a 3-month extension from the promo back then. I restart my VDS from time to time, especially after a kernel update. So the recent uptime is 15 days. It has been working great so far.

I hope you enjoy my review. Thanks for reading and have a great day.

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

Wed 07 Sep 2022 02:43:08 AM UTC                                                       

Basic System Information:                                                             
---------------------------------                                                     
Uptime     : 15 days, 20 hours, 42 minutes                                            
Processor  : AMD EPYC 7452 32-Core Processor                                          
CPU cores  : 2 @ 2345.592 MHz              
AES-NI     : ✔ Enabled                                                                
VM-x/AMD-V : ✔ Enabled                                                                
RAM        : 7.8 GiB                                                                  
Swap       : 2.0 GiB                                                                  
Disk       : 97.9 GiB                                                                 
Distro     : Debian GNU/Linux 11 (bullseye)                                           
Kernel     : 5.10.0-17-amd64                                                          

fio Disk Speed Tests (Mixed R/W 50/50):                                               
---------------------------------          
Block Size | 4k            (IOPS) | 64k           (IOPS)                              
  ------   | ---            ----  | ----           ----                               
Read       | 24.22 MB/s    (6.0k) | 820.89 MB/s  (12.8k)                              
Write      | 24.23 MB/s    (6.0k) | 825.21 MB/s  (12.8k)                              
Total      | 48.45 MB/s   (12.1k) | 1.64 GB/s    (25.7k)                              
           |                      |                                                   
Block Size | 512k          (IOPS) | 1m            (IOPS)
  ------   | ---            ----  | ----           ---- 
Read       | 2.32 GB/s     (4.5k) | 3.07 GB/s     (3.0k)
Write      | 2.45 GB/s     (4.7k) | 3.27 GB/s     (3.2k)
Total      | 4.77 GB/s     (9.3k) | 6.35 GB/s     (6.2k)

iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed     
                |                           |                 |                
Clouvider       | London, UK (10G)          | 1.04 Gbits/sec  | 992 Mbits/sec  
Online.net      | Paris, FR (10G)           | 1.03 Gbits/sec  | 988 Mbits/sec  
Hybula          | The Netherlands (40G)     | 1.04 Gbits/sec  | 995 Mbits/sec  
Uztelecom       | Tashkent, UZ (10G)        | 388 Mbits/sec   | 869 Mbits/sec  
Clouvider       | NYC, NY, US (10G)         | 429 Mbits/sec   | 930 Mbits/sec  
Clouvider       | Dallas, TX, US (10G)      | 301 Mbits/sec   | 871 Mbits/sec  
Clouvider       | Los Angeles, CA, US (10G) | 305 Mbits/sec   | 824 Mbits/sec  

iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider        | Location (Link)           | Send Speed      | Recv Speed     
                |                           |                 |                
Clouvider       | London, UK (10G)          | 1.02 Gbits/sec  | 980 Mbits/sec  
Online.net      | Paris, FR (10G)           | 1.02 Gbits/sec  | 971 Mbits/sec  
Hybula          | The Netherlands (40G)     | 1.04 Gbits/sec  | 982 Mbits/sec  
Uztelecom       | Tashkent, UZ (10G)        | 470 Mbits/sec   | 836 Mbits/sec  
Clouvider       | NYC, NY, US (10G)         | 434 Mbits/sec   | 923 Mbits/sec  
Clouvider       | Dallas, TX, US (10G)      | 261 Mbits/sec   | 880 Mbits/sec  
Clouvider       | Los Angeles, CA, US (10G) | 292 Mbits/sec   | 840 Mbits/sec  

Geekbench 5 Benchmark Test:
---------------------------------
Test            | Value                         
                |                               
Single Core     | 947                           
Multi Core      | 929                           
Full Test       | https://browser.geekbench.com/v5/cpu/17102794

Comments

  • Thanks for your review. The yabs result looks good. Hi @mxmla any eta for other locations?

  • I bought the 1tb ipv6 storage deal and am also very happy with it.

  • Great service so far. Can totally recommend.

  • agreed, rock solid!

  • I agree. I got a second VPS with server-factory because it's been fantastic so far and having dedicated CPU is pretty unusual at this price point.

  • If I renew, is it at the current price after using the discount code?

  • I have one epyc vds too from server-factory. still testing

  • @kaiant said:
    If I renew, is it at the current price after using the discount code?

    Yes, their promo is recurring.

  • I'd like to update this review with a new yabs result of the same EPYC VPS above.

    Earlier today, server-factory.com fixed a bug that affected the performance of certain VPSs including mine (probably those that were provisioned months ago). As a result, their 1 CPU core now means 1 physical core, not a thread anymore as you can see in the GB5 score below.

    This is a significant change regarding performance. I have to update this review to inform the correct information to make it fair to the provider.

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-08-20                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed 26 Oct 2022 12:32:59 PM UTC
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 14 hours, 20 minutes
    Processor  : AMD EPYC 7452 32-Core Processor
    CPU cores  : 2 @ 2345.592 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 7.8 GiB
    Swap       : 2.0 GiB
    Disk       : 97.9 GiB
    Distro     : Debian GNU/Linux 11 (bullseye)
    Kernel     : 5.10.0-19-amd64
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 261.50 MB/s  (65.3k) | 1.57 GB/s    (24.6k)
    Write      | 262.19 MB/s  (65.5k) | 1.58 GB/s    (24.7k)
    Total      | 523.69 MB/s (130.9k) | 3.16 GB/s    (49.4k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 2.76 GB/s     (5.4k) | 3.80 GB/s     (3.7k)
    Write      | 2.91 GB/s     (5.6k) | 4.05 GB/s     (3.9k)
    Total      | 5.68 GB/s    (11.0k) | 7.85 GB/s     (7.6k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 942 Mbits/sec   | 989 Mbits/sec
    Online.net      | Paris, FR (10G)           | 1.03 Gbits/sec  | 967 Mbits/sec
    Hybula          | The Netherlands (40G)     | 1.03 Gbits/sec  | 983 Mbits/sec
    Uztelecom       | Tashkent, UZ (10G)        | 475 Mbits/sec   | 847 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 383 Mbits/sec   | 839 Mbits/sec
    Clouvider       | Dallas, TX, US (10G)      | 277 Mbits/sec   | 890 Mbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 320 Mbits/sec   | 799 Mbits/sec
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 1.03 Gbits/sec  | 964 Mbits/sec
    Online.net      | Paris, FR (10G)           | 1.02 Gbits/sec  | 932 Mbits/sec
    Hybula          | The Netherlands (40G)     | 1.04 Gbits/sec  | 965 Mbits/sec
    Uztelecom       | Tashkent, UZ (10G)        | 484 Mbits/sec   | 850 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 476 Mbits/sec   | 921 Mbits/sec
    Clouvider       | Dallas, TX, US (10G)      | 236 Mbits/sec   | 886 Mbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 285 Mbits/sec   | 780 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 963
    Multi Core      | 1787
    Full Test       | https://browser.geekbench.com/v5/cpu/18226767
    
  • I am including my test results, to complete review after the bug fix.

    Processor : AMD EPYC 7702P 64-Core Processor
    CPU cores : 1 @ 1999.999 MHz
    RAM : 3.8 GiB

    Before:

    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 485
    Multi Core      | 473
    

    Now:

    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 1000
    Multi Core      | 983
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 995
    Multi Core      | 988
    
    Thanked by 3nick_ Falzo xms
  • ralfralf Member
    edited October 2022

    @nick_ said:
    Earlier today, server-factory.com fixed a bug that affected the performance of certain VPSs including mine (probably those that were provisioned months ago). As a result, their 1 CPU core now means 1 physical core, not a thread anymore as you can see in the GB5 score below.

    I just tagged him in the LET thread, I think this is actually a mistake as he'd previously advertised them as dedicated vCores.

    Previously if you had 1/2 core, 1 thread, it was limited to 1 thread that was limited to 50% CPU time of one core, so 2 VPS each with 1 dedicated thread would get 1/2 a core of performance.

    Likewise, if you have 1 core, 2 threads, both threads were pinned to the same CPU, so it was effectively dedicated resources.

    Now, with the current situation, users can steal CPU time from other people on the same node, which I think is a step backward - it means we can't run jobs at 100% any more without causing others any problems, and our stuff will now have unpredictable performance rather than being pretty much constant as before.

    I'm hoping he'll change it back as it seems to have been an accidental change while fixing the NUMA issue.

    Thanked by 1nick_
  • nick_nick_ Member
    edited October 2022

    I have two more recent VPSs at server-factory.com, EPYC 4GB and Ryzen 4GB which 1 CPU = 1 physical core before this fix. They were provisioned just a month or two ago unlike the one in my review which was ordered in May. We'll have to wait for Max to confirm though.

  • nick_nick_ Member
    edited October 2022

    According to Max's response, it will be back to normal soon. 1 CPU core = 1 thread as usual. It all makes sense now.

    The benchmark results are due to a CPU Limit Bug with Proxmox VE. We are already in touch with the Developers to fix this.

  • I am happy with EPYC VDS by @mxmla
    But why I always get "Busy" status for iperf3 ipv4 ?
    Is it yabs's bug or Almalinux's bug ?

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2022-08-20                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Thu Oct 27 06:07:12 WIB 2022
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 23 hours, 21 minutes
    Processor  : AMD EPYC 7702P 64-Core Processor
    CPU cores  : 2 @ 1999.999 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 7.6 GiB
    Swap       : 8.0 GiB
    Disk       : 101.4 GiB
    Distro     : AlmaLinux 8.6 (Sky Tiger)
    Kernel     : 4.18.0-372.26.1.el8_6.x86_64
    
    fio Disk Speed Tests (Mixed R/W 50/50):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 216.00 MB/s  (54.0k) | 3.13 GB/s    (48.9k)
    Write      | 216.57 MB/s  (54.1k) | 3.15 GB/s    (49.2k)
    Total      | 432.58 MB/s (108.1k) | 6.28 GB/s    (98.2k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 4.71 GB/s     (9.2k) | 7.51 GB/s     (7.3k)
    Write      | 4.96 GB/s     (9.7k) | 8.01 GB/s     (7.8k)
    Total      | 9.68 GB/s    (18.9k) | 15.53 GB/s   (15.1k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | busy            | busy
    Online.net      | Paris, FR (10G)           | busy            | busy
    Hybula          | The Netherlands (40G)     | busy            | busy
    Uztelecom       | Tashkent, UZ (10G)        | busy            | busy
    Clouvider       | NYC, NY, US (10G)         | busy            | busy
    Clouvider       | Dallas, TX, US (10G)      | busy            | busy
    Clouvider       | Los Angeles, CA, US (10G) | busy            | busy
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed
                    |                           |                 |
    Clouvider       | London, UK (10G)          | 1.03 Gbits/sec  | 968 Mbits/sec
    Online.net      | Paris, FR (10G)           | 1.02 Gbits/sec  | 965 Mbits/sec
    Hybula          | The Netherlands (40G)     | 1.04 Gbits/sec  | 974 Mbits/sec
    Uztelecom       | Tashkent, UZ (10G)        | 463 Mbits/sec   | 761 Mbits/sec
    Clouvider       | NYC, NY, US (10G)         | 452 Mbits/sec   | 914 Mbits/sec
    Clouvider       | Dallas, TX, US (10G)      | 313 Mbits/sec   | 875 Mbits/sec
    Clouvider       | Los Angeles, CA, US (10G) | 318 Mbits/sec   | 566 Mbits/sec
    
    Geekbench 5 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 986
    Multi Core      | 1955
    Full Test       | https://browser.geekbench.com/v5/cpu/18239150
    
Sign In or Register to comment.