Howdy, Stranger!

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


Shells Virtual Desktop
BMail.ag - Secure Email Service
Server.net
CPLicense.net
VPS Server
Buy VPN
Vultr
VMs for AI
HostDare
ReliableSite White-Label Dedicated Hosting for Resellers
InterServer VPS
BMail.ag - Secure Email Service
Best VPN
High-Performance Bare Metal Server Solutions
Karvl.com
Server Mania Cloud Hosting
DataWagon Hosting
AlphaVPS Hosting
Evoxt.com
Clouvider
VPS Hosting with NVMe
Residential IPs in the US & 4G Mobile Proxies in EU & US with Unlimited Bandwidth
ReliableSite White-Label Dedicated Hosting for Resellers
Rabisu - Hosting Solutions
Shells Virtual Desktop
Home โ€บ Offers
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.

Advin Servers - AMD Ryzen 9950X 10 Gbps VPS - Starting at $6/month for 3GB RAM

12467

Comments

  • @advinservers Why is the refund not in my bank account but in my credit balance?

  • advinserversadvinservers Member, Patron Provider

    @phanduynam said:
    @advinservers Why is the refund not in my bank account but in my credit balance?

    Open a ticket if you need it refunded to your original payment method, by default it is back to credit balance

  • @advinservers Yes, I texted yesterday but no one has replied yet.

  • advinserversadvinservers Member, Patron Provider

    Nuremberg Ryzen VPS has been restocked

  • lirrrlirrr Member

    @advinservers said:
    Nuremberg Ryzen VPS has been restocked

    as soon as i see the stock I get one immediately
    fking beast 9950x

    Thanked by 1admax
  • lirrrlirrr Member

    am i trippin or the 9950 DE plan just downgrade from 4 cpu -> 2 cpu

  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    Hi, we've recently done 2 things:

    • Lowered the cores on the 9950X plans. We noticed that the amount of cores that we allocated may have simply been too much. We try to keep our hypervisors <50% CPU load, which may be infeasible with the amount of cores/RAM that we were allocating.
    • Lowered the price. We charge $1.25 per GB of RAM now to compensate for the lower CPU allocation.
    Thanked by 1sliix
  • sliixsliix Member

    @advinservers said:
    Hi, we've recently done 2 things:

    • Lowered the cores on the 9950X plans. We noticed that the amount of cores that we allocated may have simply been too much. We try to keep our hypervisors <50% CPU load, which may be infeasible with the amount of cores/RAM that we were allocating.
    • Lowered the price. We charge $1.25 per GB of RAM now to compensate for the lower CPU allocation.

    Nice. Saw the base plan RAM is now 4GB instead of 3GB.

  • @advinservers changing from 6 USD to EUR shows 7.03 but it should be less than 6.

    Also is it possible that we pick between 9950X, 7950X3D or one of the EPYCs as listed on the page before ordering?

  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    @siemens said:
    @advinservers changing from 6 USD to EUR shows 7.03 but it should be less than 6.

    Also is it possible that we pick between 9950X, 7950X3D or one of the EPYCs as listed on the page before ordering?

    We don't have a selection yet, but I will say that we're only provisioning servers on 9950X for right now.

    I'll send a post here once/if we launch the EPYCs, and when orders might start getting placed on those. If you are placed on a 9950X, you will probably stay on a 9950X.

    The idea is that because we (plan to) pair more cores per GB of RAM on 4545P, 7950X3D, and 9575F, the performance would be similar to a 9950X hypervisor.

    The currency prices have now been updated.

    Thanked by 1siemens
  • mwmw Member

    where are all these 9950s and 7950s coming from bro everyone has so many recently

  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

  • equalzequalz Member

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    Thanked by 1advinservers
  • mwmw Member

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    thermals?

    Thanked by 1advinservers
  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    Thanked by 3equalz Tarball dev077
  • advinserversadvinservers Member, Patron Provider

    @mw said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    thermals?

    Thermals are fine, we use Dynatron L25's which can handle it.

  • @advinservers said:

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    You have Turin now ? How to select this cpu.

  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    @Motion3549 said:

    @advinservers said:

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    You have Turin now ? How to select this cpu.

    We don't have 9575F yet, it's coming soon around July hopefully. We have an EPYC 9745 / 9B45 in Los Angeles that is deployed and some KVM Premium customers have been placed on there, but we haven't setup a specific product line for it yet. (Please don't buy a KVM Premium VPS with the expectation that you'll get Turin).

  • equalzequalz Member

    thank you very much for your reply mate, great insight ๐Ÿ‘

  • mwmw Member

    @advinservers said:

    @mw said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    thermals?

    Thermals are fine, we use Dynatron L25's which can handle it.

    oh heavy, you run all owned or leased?

  • @advinservers said: 4545P, [...], and 9575F

    ๐ŸŽ†๐ŸŽ‡

  • advinserversadvinservers Member, Patron Provider

    @mw said:

    @advinservers said:

    @mw said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    thermals?

    Thermals are fine, we use Dynatron L25's which can handle it.

    oh heavy, you run all owned or leased?

    All hardware across all of our regions is owned

  • some yabs please.

    isn't the yabs for 9950X still outstanding?

  • DessgunDessgun Member
    edited June 2025

    @advinservers said:

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    Have you considered increasing the number of nodes per location? When you have only 2 nodes per location and a couple of VM start to load heavily, which affects the whole node, why couldn't you make at least 10 nodes and move VM with heavy load between them?

  • lirrrlirrr Member

    @Dessgun said:

    @advinservers said:

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    Have you considered increasing the number of nodes per location? When you have only 2 nodes per location and a couple of VM start to load heavily, which affects the whole node, why couldn't you make at least 10 nodes and move VM with heavy load between them?

    this only happens in Malaysia location ngl, so need for every location

  • xemapsxemaps Member

    Waiting Frankfurt DE - Germany B)

  • advinserversadvinservers Member, Patron Provider
    edited June 2025

    @Dessgun said:

    @advinservers said:

    @equalz said:

    @advinservers said:

    @mw said:
    where are all these 9950s and 7950s coming from bro everyone has so many recently

    It's just a coincidence, unless...

    In all seriousness, we are disappointed in the 9950X and will probably hold off on building more.

    whats up with the 9950X?

    They take up too much power, nearly 300W under full load. It's feasible if we do 256GB RAM configurations, but after doing some experimentation, we feel that it's simply too much RAM for only 16 cores.

    Before we reduced the core count, we were seeing some hypervisors peaking to >70% CPU usage. Because it starts to use hyperthreading cores at that level, CPU steal might present and single threaded scores will start to drop quite heavily. This (inevitably) results in customer complaints when they run a GB6 benchmark and it shows lower than expected results. Even if we did allocate dedicated vCPU cores to all VMs, this would still actually happen, which is also partly why we don't do VDS anymore.

    To be quite frank, the CPU load is usually fine, but it's usually the few customers that burst/sustain their CPU which really hurt. Even bursting to 4 vCPU is literally 25% of the physical cores.

    Meanwhile, on KVM Standard VPS or KVM Premium VPS, it is exceptionally rare that a hypervisor manages to reach past 50% CPU usage unless there's some new cryptocurrency mining thing or something else going on. We almost never have to restrict VMs, and whenever we do, they're usually very light caps. Customers on these plans are also more likely to be okay with some fluctuations in performance, whereas Ryzen customers are generally more sensitive due to running game servers and single-threaded dependent tasks.

    We are hopeful for the 9575F and 4545P. For the 9575F, having more VMs on one CPU will likely result in more idle CPU capacity, thus allowing us to increase core counts without compromising in performance. Similarly, we can pair the 4545P with 128GB RAM instead of 256GB, which is functionally the same CPU as the 9950X but with a 65W TDP limit. The additional cores should help avoid going past 50%, even if the TDP and clocks might be lower.

    We could change our decision in the future. But for now, we just aren't happy with it. We want to aim for consistent performance, but Ryzen prevents us from doing that. This is also why we decreased the price, we don't feel that it's worth the $2/GB RAM premium.

    Have you considered increasing the number of nodes per location? When you have only 2 nodes per location and a couple of VM start to load heavily, which affects the whole node, why couldn't you make at least 10 nodes and move VM with heavy load between them?

    While true, we don't really want to play the game of constantly live migrating VMs around whenever there's a sudden burst in CPU load.

    We did actually have more nodes per location that were intended to go, but a few of them have hardware problems so we had to postpone that.

  • equalzequalz Member

    looking forward to Turin

  • PepePepperoniPepePepperoni Member
    edited June 2025

    Your prices are stellar. What's the catch? Casual downtimes? second-hand hardware? new company? south asian owners?

  • advinserversadvinservers Member, Patron Provider

    @PepePepperoni said:
    Your prices are stellar. What's the catch? Casual downtimes? second-hand hardware? new company? south asian owners?

    Hi,

    To answer your questions:

    1. Downtime can happen with any provider, but we do our best to minimize the potential for outages. Downtime is definitely not common.
    2. We use a mixture of new and second-hand hardware. We do usually build our own servers in-house, and most of the hardware we do buy is new.
    3. We've been in business for almost 5 years now.
    4. I'm not particularly sure why the last question matters. I am Indian, but I've lived in the United States for my whole life.
    Thanked by 1yoursunny
This discussion has been closed.