Howdy, Stranger!

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


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.

BERO HOST 10th Anniversary Deals – EPYC Turin / Ryzen 9950X / IPv6-only VPS

1234568

Comments

  • berohostberohost Member, Patron Provider

    @zhili said:
    I noticed in recent replies that tickets around #20900+ have already received their pre-order invoices/links. However, my ticket #20803 (queue request for Anniversary XL) hasn't received any response yet.

    Could you please take a look when you have a moment? Just wanted to make sure it didn't get missed.

    Thanks in advance!

    Most pre-orders will be shipped this evening. Please be patient a little longer.

    Thanked by 1zhili
  • berohostberohost Member, Patron Provider

    @abitstack said:

    @berohost said:

    @Hitori0221 said:
    Geekbench 7 result of Anniversary XL after 24hrs, 2197 | 9919.

    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 347.87 MB/s  (84.9k) | 2.57 GB/s    (39.3k)
    Write      | 348.79 MB/s  (85.1k) | 2.59 GB/s    (39.5k)
    Total      | 696.66 MB/s (170.0k) | 5.16 GB/s    (78.8k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 6.28 GB/s    (11.9k) | 8.65 GB/s     (8.2k)
    Write      | 6.61 GB/s    (12.6k) | 9.23 GB/s     (8.8k)
    Total      | 12.89 GB/s   (24.5k) | 17.89 GB/s   (17.0k)
    
    

    looks pretty good.

    We're also adding hardware to ensure that performance remains the same or improves.

    This is great because performance degraded pretty quickly on the new epyc machines. Geekbench is down ~15% although steal time is still 0, so probably throttling. 4k iops started at ~140k, now more in the 80-90k region (compare this to the first yabs from the beginning of this thread - very bad for dbs). Pings now have spikes vm to vm of up to several ms, overall the ping between vms on this cluster is 0.25-0.3ms - direct ping on the old cluster used to be consistent ~0.15ms - which is a bummer. +overall the send speed to the internet went down quite a bit for everyone - old + new - which is not so bad in my case, since I need this for real production and not for file sharing, but disk and vm-vm ping are.
    It's a little bit like you get what you pay for right now, so lets see how this plays out.

    Feel free to submit a ticket if things haven't improved in 2–3 days. We're already working hard to significantly improve the situation, but it will take at least another 1–2 days.

  • Base Frequency 3.30 GHz

    9575F seems not to work with boost frequency!

  • abitstackabitstack Member
    edited 10:06AM

    @berohost said:

    @abitstack said:

    @berohost said:

    @Hitori0221 said:
    Geekbench 7 result of Anniversary XL after 24hrs, 2197 | 9919.

    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 347.87 MB/s  (84.9k) | 2.57 GB/s    (39.3k)
    Write      | 348.79 MB/s  (85.1k) | 2.59 GB/s    (39.5k)
    Total      | 696.66 MB/s (170.0k) | 5.16 GB/s    (78.8k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 6.28 GB/s    (11.9k) | 8.65 GB/s     (8.2k)
    Write      | 6.61 GB/s    (12.6k) | 9.23 GB/s     (8.8k)
    Total      | 12.89 GB/s   (24.5k) | 17.89 GB/s   (17.0k)
    
    

    looks pretty good.

    We're also adding hardware to ensure that performance remains the same or improves.

    This is great because performance degraded pretty quickly on the new epyc machines. Geekbench is down ~15% although steal time is still 0, so probably throttling. 4k iops started at ~140k, now more in the 80-90k region (compare this to the first yabs from the beginning of this thread - very bad for dbs). Pings now have spikes vm to vm of up to several ms, overall the ping between vms on this cluster is 0.25-0.3ms - direct ping on the old cluster used to be consistent ~0.15ms - which is a bummer. +overall the send speed to the internet went down quite a bit for everyone - old + new - which is not so bad in my case, since I need this for real production and not for file sharing, but disk and vm-vm ping are.
    It's a little bit like you get what you pay for right now, so lets see how this plays out.

    Feel free to submit a ticket if things haven't improved in 2–3 days. We're already working hard to significantly improve the situation, but it will take at least another 1–2 days.

    Of course. Thank you for taking care.
    Before starting submitting tickets I will rather wait a couple of days for things to settle down. At the moment it makes no sense to rush things since you are expecting a second wave once restock arrives. Right now I think all customers on the new cluster have a degraded experience (I have on all my servers) so it makes sense to wait until everybody gets great performance again (this is not a problem on a particular vm or host) instead of ticketing and stealing your time. It's just a bit annoying, since one used to order a server and use it full send right away.

    Anyways congrats on the anniversary and on the fire sale - wish you all the best !

    BTW is the network setup different on this cluster - asking because of the 0.25ms ping minimum - on the old cluster ping vm to vm is consistent at ~0.15ms if not going through the gateway (5.175.220.3) ?

  • @abitstack said:

    @berohost said:

    @Hitori0221 said:
    Geekbench 7 result of Anniversary XL after 24hrs, 2197 | 9919.

    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 347.87 MB/s  (84.9k) | 2.57 GB/s    (39.3k)
    Write      | 348.79 MB/s  (85.1k) | 2.59 GB/s    (39.5k)
    Total      | 696.66 MB/s (170.0k) | 5.16 GB/s    (78.8k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 6.28 GB/s    (11.9k) | 8.65 GB/s     (8.2k)
    Write      | 6.61 GB/s    (12.6k) | 9.23 GB/s     (8.8k)
    Total      | 12.89 GB/s   (24.5k) | 17.89 GB/s   (17.0k)
    
    

    looks pretty good.

    We're also adding hardware to ensure that performance remains the same or improves.

    This is great because performance degraded pretty quickly on the new epyc machines. Geekbench is down ~15% although steal time is still 0, so probably throttling. 4k iops started at ~140k, now more in the 80-90k region (compare this to the first yabs from the beginning of this thread - very bad for dbs). Pings now have spikes vm to vm of up to several ms, overall the ping between vms on this cluster is 0.25-0.3ms - direct ping on the old cluster used to be consistent ~0.15ms - which is a bummer. +overall the send speed to the internet went down quite a bit for everyone - old + new - which is not so bad in my case, since I need this for real production and not for file sharing, but disk and vm-vm ping are.
    It's a little bit like you get what you pay for right now, so lets see how this plays out.

    Do not forget the most probably heavy YABS runnings because anyone want to see the results.

    As Berohost has said, in 2-3 days the situation will improve.

  • barbarosbarbaros Member

    All the MJJs will abuse the shit out of the nodes, last time someone was asking in this thread if it's okay to burst the core for 24hours. So performance degrade will happen.

  • allthemtingsallthemtings Member, Megathread Squad

    @barbaros said:
    All the MJJs will abuse the shit out of the nodes, last time someone was asking in this thread if it's okay to burst the core for 24hours. So performance degrade will happen.

  • LkwangLkwang Member

    Hope to restock soon, looking forward to pre-ordering or buying, the money is ready

  • berohostberohost Member, Patron Provider

    @abitstack said:

    @berohost said:

    @abitstack said:

    @berohost said:

    @Hitori0221 said:
    Geekbench 7 result of Anniversary XL after 24hrs, 2197 | 9919.

    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 347.87 MB/s  (84.9k) | 2.57 GB/s    (39.3k)
    Write      | 348.79 MB/s  (85.1k) | 2.59 GB/s    (39.5k)
    Total      | 696.66 MB/s (170.0k) | 5.16 GB/s    (78.8k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 6.28 GB/s    (11.9k) | 8.65 GB/s     (8.2k)
    Write      | 6.61 GB/s    (12.6k) | 9.23 GB/s     (8.8k)
    Total      | 12.89 GB/s   (24.5k) | 17.89 GB/s   (17.0k)
    
    

    looks pretty good.

    We're also adding hardware to ensure that performance remains the same or improves.

    This is great because performance degraded pretty quickly on the new epyc machines. Geekbench is down ~15% although steal time is still 0, so probably throttling. 4k iops started at ~140k, now more in the 80-90k region (compare this to the first yabs from the beginning of this thread - very bad for dbs). Pings now have spikes vm to vm of up to several ms, overall the ping between vms on this cluster is 0.25-0.3ms - direct ping on the old cluster used to be consistent ~0.15ms - which is a bummer. +overall the send speed to the internet went down quite a bit for everyone - old + new - which is not so bad in my case, since I need this for real production and not for file sharing, but disk and vm-vm ping are.
    It's a little bit like you get what you pay for right now, so lets see how this plays out.

    Feel free to submit a ticket if things haven't improved in 2–3 days. We're already working hard to significantly improve the situation, but it will take at least another 1–2 days.

    Of course. Thank you for taking care.
    Before starting submitting tickets I will rather wait a couple of days for things to settle down. At the moment it makes no sense to rush things since you are expecting a second wave once restock arrives. Right now I think all customers on the new cluster have a degraded experience (I have on all my servers) so it makes sense to wait until everybody gets great performance again (this is not a problem on a particular vm or host) instead of ticketing and stealing your time. It's just a bit annoying, since one used to order a server and use it full send right away.

    Anyways congrats on the anniversary and on the fire sale - wish you all the best !

    BTW is the network setup different on this cluster - asking because of the 0.25ms ping minimum - on the old cluster ping vm to vm is consistent at ~0.15ms if not going through the gateway (5.175.220.3) ?

    The network setup shouldn't be any different. We'd be happy to take a closer look if the issue persists in 1–2 days.

  • berohostberohost Member, Patron Provider

    @Lkwang said:
    Hope to restock soon, looking forward to pre-ordering or buying, the money is ready

    It will take another 2–3 days to restock

  • hyperblasthyperblast Member
    edited 11:07AM

    lucky you!

    mine:

    Basic System Information:
    ---------------------------------
    Uptime     : 1 days, 14 hours, 32 minutes
    Processor  : AMD EPYC 9575F 64-Core Processor
    CPU cores  : 4 @ 3300 MHz <----------------------base clock
    AES-NI     : ? 
    VM-x/AMD-V : ✔ Enabled
    RAM        : 20.0 GiB
    Swap       : 3.0 GiB
    Disk       : 198.9 GiB
    Distro     : Microsoft Windows Server 2025 (64-Bit)
    Kernel     : 10.0.26100 Build 26100
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 2129
    Multi Core      | 6460
    
  • networknetwork Member

    @hyperblast said:

    lucky you!

    mine:

    Basic System Information:
    ---------------------------------
    Uptime     : 1 days, 14 hours, 32 minutes
    Processor  : AMD EPYC 9575F 64-Core Processor
    CPU cores  : 4 @ 3300 MHz <----------------------base clock
    AES-NI     : ? 
    VM-x/AMD-V : ✔ Enabled
    RAM        : 20.0 GiB
    Swap       : 3.0 GiB
    Disk       : 198.9 GiB
    Distro     : Microsoft Windows Server 2025 (64-Bit)
    Kernel     : 10.0.26100 Build 26100
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 2129
    Multi Core      | 6460
    

    4.1 Ghz here.

    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 3 hours, 24 minutes
    Processor  : AMD EPYC 9275F 24-Core Processor
    CPU cores  : 4 @ 4099.992 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 11.7 GiB
    Swap       : 0.0 KiB
    Disk       : 118.0 GiB
    Distro     : Debian GNU/Linux 13 (trixie)
    Kernel     : 6.12.38+deb13-amd64
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
  • i see. and base clock for 9275f is?

  • berohostberohost Member, Patron Provider

    @hyperblast said:
    i see. and base clock for 9275f is?

    4.1 GHz

  • @berohost said:

    @hyperblast said:
    i see. and base clock for 9275f is?

    4.1 GHz

    ok. and both 9275f and 9575f always run with base block?

  • @hyperblast said:

    @berohost said:

    @hyperblast said:
    i see. and base clock for 9275f is?

    4.1 GHz

    ok. and both 9275f and 9575f always run with base block?

    No, I don’t think you ever get the actual boost clocks in VM’s.

  • ascicodeascicode Member

    Almost forgot my 3 vps from last year.

  • Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

  • @yforward77 said: as I’m very interested

    If you are - open a ticket and see what will happen.

  • LkwangLkwang Member

    @yforward77 said:
    Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

    It’s now
    I see it's in stock right now, 9950x

  • berohostberohost Member, Patron Provider

    @yforward77 said:
    Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

    We've restocked the Ryzen VPS.
    EPYC Turin will take another 2–3 days.

    Thanked by 1lovelyserver
  • LkwangLkwang Member

    @berohost said:

    @yforward77 said:
    Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

    We've restocked the Ryzen VPS.
    EPYC Turin will take another 2–3 days.

    Looking forward to it, looking forward to the stock replenishment as soon as possible - I need it

  • lbrootlbroot Member
    edited 12:23PM

    @berohost said:

    @lbroot said:
    Can I upgrade Anniversary M to Anniversary L?

    Yes, via a support ticket

    Support PIN 223289 I have top up the credit and opened a ticket, thx!

  • @Lkwang said:

    @yforward77 said:
    Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

    It’s now
    I see it's in stock right now, 9950x

    @berohost said:

    @yforward77 said:
    Hi, will the Anniversary S Ryzen VPS plan be restocked again? If so, do you know roughly when it will become available?
    Also, would it be possible for me to top up my account balance first, open a support ticket, and then wait for a server to be allocated when stock becomes available?
    I would really appreciate it if you could reply to me, as I’m very interested in purchasing this plan. Thank you!

    We've restocked the Ryzen VPS.
    EPYC Turin will take another 2–3 days.

    Thanks!

  • @barbaros said:
    All the MJJs will abuse the shit out of the nodes, last time someone was asking in this thread if it's okay to burst the core for 24hours. So performance degrade will happen.

    +1 block MJJ unless ticket open

    Thanked by 3meowwcc radex Nekopara
  • JinjoJinjo Member

    @hyperblast said:

    @berohost said:

    @hyperblast said:
    i see. and base clock for 9275f is?

    4.1 GHz

    ok. and both 9275f and 9575f always run with base block?

    The hypervisor usually exposes the base clock to the guest rather than the CPU’s real-time boost frequency.

  • LkwangLkwang Member

    If it runs out of stock and I don't get one, that would be really frustratingt would be really frustrating if I missed out on the restock.
    I need EPYC Anniversary M

    Thanked by 1akrdev
  • @Lkwang said: that would be really frustratingt would be really frustrating

    I do not do not repeat twice.

  • liwin1liwin1 Member

    @berohost Hello, my ticket number is #20822. I have pre-ordered an XXL size item, but no one has replied to me yet

Sign In or Register to comment.