Howdy, Stranger!

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


HostBrr | 1 Year celebration deals | EPIC EPYC DEALS + Storage | Flash deals - Page 9
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.

HostBrr | 1 Year celebration deals | EPIC EPYC DEALS + Storage | Flash deals

1679111221

Comments

  • @Porlam said:

    @hyperblast said: whois behind this script?

    It's me. And, I already fixed it problem.

    hmm... here all provider are failed @Porlam

  • @DigiGoon said:
    @labze i have 5 vccore 7950xd 8gb ram 150gb nvme plan from hostbrr, should i switch to the epyc plan of 6vcores from anniversary sale? i mean which plan will have more performance?

    What price? Can I still benefit of this 7950xd offer? @labze

  • kvz12kvz12 Member
    edited April 25

    @labze said:
    Just to add to the Geekbench discussion, do note that Geekbench 5 is a much better indicator for multi-core performance than Geekbench 6. As soon as you go above 4 cores, Geekbench 6 scales terribly.

    Geekbench 5 is more tailored towards server workloads whereas Geekbench 6 was tailored towards desktop PCs and workloads associated with this.

    Fun fact, this 128 core / 256 core beast only barely beats a 16 core / 32 thread 7950XD in Geekbench 6.

    Fun fact, Geekbench in general is a bad benchmark and can be faked, seemingly easily. Here's a "Xiaomi 2107113SI (Mi 11T Pro)" "AMD Ryzen 5 7600" beating everything else in single core.

    https://browser.geekbench.com/v6/cpu/singlecore

  • zrj766zrj766 Member

    A friend did a non-rigorous test, and when importing a large number of databases, EPYC processed them faster than Ryzen.

    Disk is not the bottleneck, they are both fast. The processor used EPYC 9754, 9554, compared to 7950X. basically the same results whether it was HostBrr, Hybula, Layerstack, server-factory, or AWS, Hetzner, Azure.

    Thanked by 1nick_
  • lukast__lukast__ Member
    edited April 25

    @kvz12 said: Fun fact, Geekbench in general is a bad benchmark and can be faked, seemingly easily. Here's a "Xiaomi 2107113SI (Mi 11T Pro)" "AMD Ryzen 5 7600" beating everything else in single core.

    But that might only be a problem with the geekbench website - basically I don't see how they could verify whether an uploaded result is true if you invest enough time to disassemble geekbench/log the network communication of geekbench/manipulate the values of geekbench in-memory/...
    And I think that this benchmark you mention seems pretty much completely faked - I don't think you could install a 65-watt cpu inside of an ARM smartphone, add 3 GB memory, remove/disable 2 cores of the cpu, halve the base tact, and get 3 times single core geekbench/2 times multi core.
    I think that the website simply doesn't validate the results enough to trust some benchmarks submitted by someone, which doesn't mean that there is a general problem with geekbench.

  • dev_vpsdev_vps Member

    @zrj766 said:
    A friend did a non-rigorous test, and when importing a large number of databases, EPYC processed them faster than Ryzen.

    Disk is not the bottleneck, they are both fast. The processor used EPYC 9754, 9554, compared to 7950X. basically the same results whether it was HostBrr, Hybula, Layerstack, server-factory, or AWS, Hetzner, Azure.

    Try running HDR image compression on multiple threads and measure that performance, if you are looking to benchmark the computing power of the VPS.

    Loading database is not directly proportional to computing power.

  • kvz12kvz12 Member

    @lukast__ said:
    I think that the website simply doesn't validate the results enough to trust some benchmarks submitted by someone, which doesn't mean that there is a general problem with geekbench.

    And that's the issue. There's obviously fake benchmark results on their own website, with no way of reporting them and no indication to someone who doesn't know any better that they are fake. They just don't seem to care. The actual tests performed might be fine, but I lose trust in what they mean if they let that happen.

  • @dev_vps said:

    @zrj766 said:
    A friend did a non-rigorous test, and when importing a large number of databases, EPYC processed them faster than Ryzen.

    Disk is not the bottleneck, they are both fast. The processor used EPYC 9754, 9554, compared to 7950X. basically the same results whether it was HostBrr, Hybula, Layerstack, server-factory, or AWS, Hetzner, Azure.

    Try running HDR image compression on multiple threads and measure that performance, if you are looking to benchmark the computing power of the VPS.

    Loading database is not directly proportional to computing power.

    Some of the people faking Benchmark scores like to disguise their devices as Xiaomi phones. Geekbench lacks a verification feature like CPU-Z and always uploads the results, causing the scores in the database much less informative

  • @kvz12 said: And that's the issue. There's obviously fake benchmark results on their own website, with no way of reporting them and no indication to someone who doesn't know any better that they are fake. They just don't seem to care. The actual tests performed might be fine, but I lose trust in what they mean if they let that happen.

    It's of course bad you can't report fakes and that they don't even seem to notice such obvious fakes.
    It's also bad that often the overview over a specific CPU is often simply inaccurate as for example (I think) many VPS benchmarks influence the scores, but still, geekbench is, in my opinion, perfectly suitable for comparing servers (though GB6 multi core score not so much).

  • kvz12kvz12 Member

    @lukast__ said:

    @kvz12 said: And that's the issue. There's obviously fake benchmark results on their own website, with no way of reporting them and no indication to someone who doesn't know any better that they are fake. They just don't seem to care. The actual tests performed might be fine, but I lose trust in what they mean if they let that happen.

    It's of course bad you can't report fakes and that they don't even seem to notice such obvious fakes.
    It's also bad that often the overview over a specific CPU is often simply inaccurate as for example (I think) many VPS benchmarks influence the scores, but still, geekbench is, in my opinion, perfectly suitable for comparing servers (though GB6 multi core score not so much).

    I agree.

  • Are there specific instructions on cpu frequency? Such as the maximum limit for a certain period of time.

  • btselembtselem Member

    Is it a general network thing with rather big variance in pings if one leaves mtr running for a while or is it something that is likely to be isolated to certain nodes and somebody blasting the port?

    I have a valentines day offer and the performance is great and nothing to complain about when it comes to that but there seems to be some ping jumps/jitter and I wonder if I should expect same on the epyc offer or if it might be different/this is something that is expected to be resolved in the future?

  • labzelabze Member, Patron Provider

    @btselem said:
    Is it a general network thing with rather big variance in pings if one leaves mtr running for a while or is it something that is likely to be isolated to certain nodes and somebody blasting the port?

    I have a valentines day offer and the performance is great and nothing to complain about when it comes to that but there seems to be some ping jumps/jitter and I wonder if I should expect same on the epyc offer or if it might be different/this is something that is expected to be resolved in the future?

    I am not familiar with this issue. Does the ping fluctuate at the VPS or the routing inbetween?

  • btselembtselem Member

    @labze said:

    @btselem said:
    Is it a general network thing with rather big variance in pings if one leaves mtr running for a while or is it something that is likely to be isolated to certain nodes and somebody blasting the port?

    I have a valentines day offer and the performance is great and nothing to complain about when it comes to that but there seems to be some ping jumps/jitter and I wonder if I should expect same on the epyc offer or if it might be different/this is something that is expected to be resolved in the future?

    I am not familiar with this issue. Does the ping fluctuate at the VPS or the routing inbetween?

    If I mtr any of my vps at other frankfurt hosts (aurologic, clouvider, ovh etc) then I get intermittent +10ms spikes and also in general some jitter

  • DigiGoonDigiGoon Member
    edited April 25

    @labze Can you?

    Also, can you create any custom plan for me? For Android ROM compilation?
    Ryzen 7950XD - 10 core or above
    32GB RAM
    1 TB NvME

    What might be the price of it?

  • @labze possible to announce flash deals little earlier in time? Eliminate timezone bias basically ....

  • i think it's good to be in the same time zone as @labze in order to have an advantage. but with greencloud i'm usually in the tube!

  • ehabehab Member

  • ehabehab Member

    better idea

  • Happy birthday to a fantastic provider! @labze
    It is an absolute pleasure doing business with you.
    Great offer but am saving up for something else from Hostbrr!

  • I'm very sleepy, hope you guys get nice deals! 😭

    Thanked by 2ehab amaeva080
  • Let's goooo

    Thanked by 1amaeva080
  • amaeva080amaeva080 Member
    edited April 25

    Sleepy now ....
    @labze show me a dreamy deal, while I'm still awake. :)

    Thanked by 1sh97
  • sh97sh97 Member

    @amaeva080 said:
    Sleepy now ....
    @labze show me a dreamy deal, while I'm still awake. :)

    Pretty please @labze

    Thanked by 1amaeva080
  • labzelabze Member, Patron Provider
    edited April 25

    Flash deal 3

    1 vCore AMD Ryzen 7950XD
    4 GB DDR5 ECC RAM 
    100 GB NVMe Gen4 Storage
    5000 GB HDD Raid-6 storage
    5000 GB Bandwidth @ 1 Gbps
    IPv4
    IPv6/64 
    Hosted in Frankfurt, Germany
    €5/month
    $5.5/month

    Provision time: Later

  • marquelinmarquelin Member
    edited April 25

    @labze said:

    Flash deal 3

    1 vCore AMD Ryzen 7950XD
    4 GB DDR5 ECC RAM 
    100 GB NVMe Gen4 Storage
    5000 GB HDD Raid-6 storage
    5000 GB Bandwidth @ 1 Gbps
    IPv4
    IPv6/64 
    Hosted in Frankfurt, Germany
    €5/month
    $5.5/month

    Provision time: Later

  • labzelabze Member, Patron Provider

    @labze said:

    Flash deal 3

    1 vCore AMD Ryzen 7950XD
    4 GB DDR5 ECC RAM 
    100 GB NVMe Gen4 Storage
    5000 GB HDD Raid-6 storage
    5000 GB Bandwidth @ 1 Gbps
    IPv4
    IPv6/64 
    Hosted in Frankfurt, Germany
    €5/month
    $5.5/month

    Provision time: Later

    Ok this is amazing. I didnt even get to post this and it is out of stock. Keep your F5 ready, restock at 25 minutes past this hour (5 minutes from now)

  • lc475lc475 Member

    Out of stock :o

  • labzelabze Member, Patron Provider

    @marquelin said:

    @labze said:

    Flash deal 3

    1 vCore AMD Ryzen 7950XD
    4 GB DDR5 ECC RAM 
    100 GB NVMe Gen4 Storage
    5000 GB HDD Raid-6 storage
    5000 GB Bandwidth @ 1 Gbps
    IPv4
    IPv6/64 
    Hosted in Frankfurt, Germany
    €5/month
    $5.5/month

    Provision time: Later

    Its $5/year on the site lol

    That'll be flash deal of the year I guess.

    Thanked by 2MateiSR amaeva080
  • @labze said:

    @marquelin said:

    @labze said:

    Flash deal 3

    1 vCore AMD Ryzen 7950XD
    4 GB DDR5 ECC RAM 
    100 GB NVMe Gen4 Storage
    5000 GB HDD Raid-6 storage
    5000 GB Bandwidth @ 1 Gbps
    IPv4
    IPv6/64 
    Hosted in Frankfurt, Germany
    €5/month
    $5.5/month

    Provision time: Later

    Its $5/year on the site lol

    That'll be flash deal of the year I guess.

    Reoccuring? I finally got one :smile:

Sign In or Register to comment.