DeinServerHost rootservers
DeinServerHost seems to run an offer on higher end rootservers Expert, Large and Mega.
https://deinserverhost.de/root-server-ssd-mieten
I was looking for an excuse to try them, since I've never used them before, so I got a Mega, I think it's quite nice for 25 eur
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2023-04-23 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Sun Jul 2 11:04:41 AM EEST 2023
Basic System Information:
---------------------------------
Uptime : 0 days, 9 hours, 36 minutes
Processor : AMD EPYC 7443P 24-Core Processor
CPU cores : 8 @ 2849.998 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 31.3 GiB
Swap : 953.0 MiB
Disk : 298.6 GiB
Distro : Debian GNU/Linux 12 (bookworm)
Kernel : 6.1.0-9-amd64
VM Type : KVM
IPv4/IPv6 : ✔ Online / ✔ Online
IPv6 Network Information:
---------------------------------
ISP : combahton IT Services
ASN : AS30823 combahton GmbH
Host : combahton IT Services UG
Location : Mauern, Bavaria (BY)
Country : Germany
fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 14.47 MB/s (3.6k) | 191.42 MB/s (2.9k)
Write | 14.48 MB/s (3.6k) | 192.42 MB/s (3.0k)
Total | 28.96 MB/s (7.2k) | 383.84 MB/s (5.9k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 490.13 MB/s (957) | 489.76 MB/s (478)
Write | 516.17 MB/s (1.0k) | 522.37 MB/s (510)
Total | 1.00 GB/s (1.9k) | 1.01 GB/s (988)
iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed | Ping
----- | ----- | ---- | ---- | ----
Clouvider | London, UK (10G) | 904 Mbits/sec | 441 Mbits/sec | 11.0 ms
Scaleway | Paris, FR (10G) | 898 Mbits/sec | 729 Mbits/sec | 15.1 ms
NovoServe | North Holland, NL (40G) | 1.02 Gbits/sec | busy | 14.0 ms
Uztelecom | Tashkent, UZ (10G) | busy | 245 Mbits/sec | 72.7 ms
Clouvider | NYC, NY, US (10G) | busy | 60.4 Mbits/sec | 80.5 ms
Clouvider | Dallas, TX, US (10G) | 229 Mbits/sec | 69.8 Mbits/sec | 115 ms
Clouvider | Los Angeles, CA, US (10G) | 233 Mbits/sec | 72.3 Mbits/sec | 139 ms
iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed | Ping
----- | ----- | ---- | ---- | ----
Clouvider | London, UK (10G) | 993 Mbits/sec | 936 Mbits/sec | 11.0 ms
Scaleway | Paris, FR (10G) | 886 Mbits/sec | busy | 15.7 ms
NovoServe | North Holland, NL (40G) | 987 Mbits/sec | 957 Mbits/sec | 12.9 ms
Uztelecom | Tashkent, UZ (10G) | 668 Mbits/sec | 333 Mbits/sec | 72.6 ms
Clouvider | NYC, NY, US (10G) | 335 Mbits/sec | 56.3 Mbits/sec | 80.7 ms
Clouvider | Dallas, TX, US (10G) | 255 Mbits/sec | 419 Mbits/sec | 116 ms
Clouvider | Los Angeles, CA, US (10G) | 468 Mbits/sec | 238 Mbits/sec | 139 ms
Geekbench 5 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 1084
Multi Core | 5999
Full Test | https://browser.geekbench.com/v5/cpu/21406541
YABS completed in 9 min 31 sec
Thanked by (1)Ganonk
Comments
Looks a bit meh for the price to be honest
CPU should be a lot faster and IO performance is also quite low
AMD EPYC / NVMe / 10GBPs KVM in Frankfurt - https://v6node.com
Looking for an unbeatable AMD EPYC Baremetal Server in Frankfurt? Drop me a PM
Can you please suggest something else with 32GB ram on this price and better performance?
I am interested.
Well, I'm probably biased but our Node 32GB G2 might be a good fit at 23,99€ (+80ct IPv4) :)
AMD EPYC / NVMe / 10GBPs KVM in Frankfurt - https://v6node.com
Looking for an unbeatable AMD EPYC Baremetal Server in Frankfurt? Drop me a PM
Thanks. Can you
curl -sL yabs.sh | bash -s -- -5for direct comparison?What is your policy on CPU usage?
The CPU performance looks fine. We've been doing extensive testing with EPYC and sometimes the multi core might not exactly scale if the host node is using threads (hyperthreading). Single core is slightly lower but a 1084 is still pretty good. You got a good deal 👍
@iandk Keep in mind that OP posted a GB5 and not a GB6 which scores differently.
I am a representative of Advin Servers
This is like total no-starter for me. I know it's 4k block size... but too damn low, especially if you want database there or something liek that.
Yeah, around 1000 GB5 for EPYC is good, but then it all depends what do you need this server for :-)
[as I don't need 32GB RAM and personally I would rather get Ryzen with less RAM, but more CPU power]
If you just need RAM then get Contabo - they don't oversell that one ]:->. On serious note - if you happy then you happy and this is what matters the most, but I have a feeling for this price you could get something without this 4k block size speeds and on par with rest of specs.
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
Yeah, on the other hand in my case this is balanced out due to memory caching.
Usually, 25% sustained but we dont strictly measure it.
As long as you're not using all cores 24/7 you should be fine
AMD EPYC / NVMe / 10GBPs KVM in Frankfurt - https://v6node.com
Looking for an unbeatable AMD EPYC Baremetal Server in Frankfurt? Drop me a PM
yep I just noticed, I assumed it was GB6
AMD EPYC / NVMe / 10GBPs KVM in Frankfurt - https://v6node.com
Looking for an unbeatable AMD EPYC Baremetal Server in Frankfurt? Drop me a PM
@yoursunny trigger post
Haven't bought a single service in VirMach Great Ryzen 2022 - 2023 Flash Sale.
https://lowendspirit.com/uploads/editor/gi/ippw0lcmqowk.png
I've read so many bad reviews for contabo performance, that to be honest I've never used them. Only once I asked them (before buying) if they support custom ISO installation, and they were asking for additional €25.00 "due to the technical and administrative effort". That was a no-no.
Do you have any idea how this is compared to their VPS L in terms of performance?
I can't take any (EU) provider maintaining customer relations over Discord for real.
As far as I know, usage is currently not GDPR-compliant? No-go for b2c communication.
If that's acceptable, data protection is probably not a priority overall.
Support appeared to know everything. Quibbling over phrases. Nothing I'd consider a serious business.
Went with Server-Factory instead and can't complain after more than a year of usage. No-bullshit control panel, no Discord/Facebook/Twitter. Consistent performance. (~1600 single core gb5) and no "fair-use unlimited traffic" shenanigans.
(If you read the ToS of Deinserverhost, unlimited appears to be at 10TB/month(!))
Don't want to make this a sales pitch, but now that are some straight to the point ToS.
Can't judge about their legal upholding, but kudos to @mxmla for the overall no-bullshit approach.
Guess storage performance is not strictly defined because no one hit the limit yet.
Why use swap at 32gig RAM? Disable swap and rerun the test.
huh?
Isnt it like it, that swap would be used as RAM alternativ, so you got disk activity for swap thats could make your disk i/o slower.
If the server is setup correctly it will only use swap if it has no other choice. As long as the ram is not full it should not touch the disk.
Sometimes it does happen but I have not seen a partochlar case that it will be that much of an issue.
URL Shortener | YetiNode | Come join us on the MetalVPS IRC channel!!! | Don't be a jerk, let YetiNode do the work.
Short answer: no, this is not windows xp
Swap in linux can be used when the ram is nearly full, but this is not the primary reason we use it. In long running servers it's used to free up pages that are not used in order to give more ram for caches. This behavior is controlled by vm.swappiness in sysctl