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.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I have way worse IO on mine which is set to UHP max too... Hmm... What location?
I'm in switzerland
Did you stop the instance, remove boot volume from it, re attach it and start it again?
Colo at inleed.se in Sweden 31 eur/month excl vat
My laptop, built-in 4G modem
My ampere instance in London also has pretty bad IO performance, but then again it is free. And I'm actually liking the CPU performance, especially as probably my entire application and database will be able to fit in RAM, so the IO speed isn't that much of a big deal.
disk performance is according to block storage size. 200GB will make it faster.
`# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Yet-Another-Bench-Script
v2022-05-06
https://github.com/masonr/yet-another-bench-script
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Mon May 23 10:50:10 UTC 2022
Basic System Information:
Uptime : 0 days, 2 hours, 0 minutes
Processor : AMD Ryzen 9 3900X 12-Core Processor
CPU cores : 1 @ 3792.874 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 3.8 GiB
Swap : 0.0 KiB
Disk : 78.7 GiB
Distro : AlmaLinux 8.5 (Arctic Sphynx)
Kernel : 4.18.0-348.20.1.el8_5.x86_64
fio Disk Speed Tests (Mixed R/W 50/50):
iperf3 Network Speed Tests (IPv4):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 849 Mbits/sec | 233 Mbits/sec
Online.net | Paris, FR (10G) | 1.12 Gbits/sec | 572 Mbits/sec
Hybula | The Netherlands (40G) | 1.11 Gbits/sec | 960 Mbits/sec
Clouvider | NYC, NY, US (10G) | 483 Mbits/sec | 59.0 Mbits/sec
Velocity Online | Tallahassee, FL, US (10G) | 1.02 Gbits/sec | 188 Mbits/sec
Clouvider | Los Angeles, CA, US (10G) | 257 Mbits/sec | 47.5 Mbits/sec
Geekbench 5 Benchmark Test:
Test | Value
|
Single Core | 702
Multi Core | 723
Full Test | https://browser.geekbench.com/v5/cpu/15076790
`
is it normal for AMD Ryzen 9 3900X Processor to have 702 single core score?
maybe you have 50% of one core
buyvm
I am on 4g ram plan with 1 dedicated cpu core ,lux location
@Francisco
Set higher speed in dashboard, is free
Often cpu is undervolted to save electricity and wear. Not much you can do on kvm level
"Often"? Name at least two providers ever that undervolted CPUs.
I contacted support , they replied
"For a more consistent and predictable performance without overheating of CPUs and throttling we sometimes disable boost on them. That results in lower benchmark scores but gives overall a more stable system."
Ok I meant overall lower performance, others just disable boosting like the guy above said
Avesome peering. I rarely see 800+ Mbps from one location to EU and US.
nice
RPI 3A+ 512MB RAM
It seems geekbench performs badly in low memory conditions.
For comparison, here's a 2-core KVM with more memory created on the same host (KS-LE-1): https://lowendtalk.com/discussion/comment/3418484#Comment_3418484
Yesterday I made a very small instance for a very specific purpose that didn't need much RAM, so I set it up with 256MB and 1GB swap. Today, I decided to yabs it and was quite surprised because it was using very little CPU time on the host...
So, yabs first:
Now that performance is substantially lower than the 863 single core result from the yabs from the other VM, and the host is basically idle. Out of curiosity, while it was running (VERY slowly), I ran top on the VM. It was swapping badly, as you might have guessed:
But what's interesting is that it looks like while the VM was swapping like crazy, and having a massive performance hit, it was all just hitting the RAM cache on the host and barely making any impact at all:
For actual use, I'm not too concerned, as this VM is just there to run one very specific weekly task, and it had about 100MB headroom without hitting swap last night, but I was surprised that GB5 needs so much RAM for the CPU test.
Swiss VPS SSDVPS2 - https://swiss-vps.com/ssd-vps-hosting/
Hetzner Cloud CPX11 - Nuremberg Location
root@server:~# curl -sL yabs.sh | bash
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Yet-Another-Bench-Script
v2022-06-11
https://github.com/masonr/yet-another-bench-script
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Wed 20 Jul 2022 04:30:04 PM UTC
Basic System Information:
Uptime : 0 days, 1 hours, 27 minutes
Processor : AMD EPYC Processor
CPU cores : 2 @ 2495.310 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ❌ Disabled
RAM : 1.9 GiB
Swap : 0.0 KiB
Disk : 37.5 GiB
Distro : Debian GNU/Linux 11 (bullseye)
Kernel : 5.10.0-16-amd64
fio Disk Speed Tests (Mixed R/W 50/50):
iperf3 Network Speed Tests (IPv4):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 1.25 Gbits/sec | 6.56 Gbits/sec
Online.net | Paris, FR (10G) | 8.55 Gbits/sec | 7.62 Gbits/sec
Hybula | The Netherlands (40G) | 10.2 Gbits/sec | 9.62 Gbits/sec
Uztelecom | Tashkent, UZ (10G) | 1.09 Gbits/sec | 2.12 Gbits/sec
Clouvider | NYC, NY, US (10G) | 1.79 Gbits/sec | 2.26 Gbits/sec
Clouvider | Dallas, TX, US (10G) | 479 Mbits/sec | 709 Mbits/sec
Clouvider | Los Angeles, CA, US (10G) | 945 Mbits/sec | 1.21 Gbits/sec
iperf3 Network Speed Tests (IPv6):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 6.01 Gbits/sec | 7.04 Gbits/sec
Online.net | Paris, FR (10G) | 8.32 Gbits/sec | 7.62 Gbits/sec
Hybula | The Netherlands (40G) | 9.44 Gbits/sec | 9.22 Gbits/sec
Uztelecom | Tashkent, UZ (10G) | 2.01 Gbits/sec | 1.70 Gbits/sec
Clouvider | NYC, NY, US (10G) | 2.26 Gbits/sec | 2.34 Gbits/sec
Clouvider | Dallas, TX, US (10G) | 1.24 Gbits/sec | 1.40 Gbits/sec
Clouvider | Los Angeles, CA, US (10G) | 950 Mbits/sec | 1.23 Gbits/sec
Geekbench 5 Benchmark Test:
Test | Value
|
Single Core | 603
Multi Core | 1108
Full Test | https://browser.geekbench.com/v5/cpu/16128007
You might have a better outcome trying to run GB4, GB5's Djikstra test is too bloated for it to be able to run. GB5 ~ GB4 * 0.25.
Here is a YABS test from one of our recent builds, unfortunately we did not test YABS on it while it was un-used, this thing is basically full already so yea...
Virmach BF-SPECIAL-2020 in LA ColoCrossing ($4.95/y) after migrate to ryzen in Frankfurt, DE
Virmach Ryzen Special 2560 in Tokyo ($21.85/y)
Virmach special offer in Buffalo Colocrossing (free forever) that i got it on easter after migrate to ryzen in NY
ShedDC has rid itself of the horrible OS that is Ubuntu:
Benchmark
Edit: My SFP+ 10 Gbit/s module isn't working (fiber) on my NAS connected to my switch over a single gigabit connection for the time being (i.e. slow outgoing internet speeds while I perform file transfers to this server).
Edit #2: Provided some speedtests from my PC (which has a working SFP+ module).
Speedtest.net
Local:
Japan:
UK:
Argh… English is hard. Fixed the horribly written paragraph.
Hetzner dedicated AX41 (34e + VAT per month)
A raspberry pi?
Hetzner Auction 29e/mo