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.
VPS hosting. How many clients usually share 1 core?
What is the quantity of clients sharing 1 core if we're talking about VPS hosting? Max/min/average. I saw a lot of tests here on LET and sometimes 1 core gets 4500 points in geekbench 4 tests, sometimes 3000, even 900 points have been spotted and that led me to this question...
Comments
Hundreds.
Otherwise, you can't even break even, given the thin profit margin.
Therefore, VPS benchmarks are useless. Stop benchmarking it.
deank damn it. Its always fun to see people benchmarking shared cores and then raging on "This host sucks!!?!"
Well, it differ from node to node and host to host, sometimes you can have lots of users sharing and sometimes you can have only handful of them. Some hosts are strict about their node's stability, they kicks you if you abuse too much , some are like it slow sometimes but its ok as long as it works. But yes, as serv_ee said, it can be a lot.
And I firmly believe benchmarking is a form of abuse, wasting a sever's resources for absolutely no reason other than self gratification.
If you browse threads of VPS hosts that rose and fell on LET specifically, you will find that profit margin of each VPS account was in cents, not even a buck.
What do ya expect?
I'm just curious, nothing more. I do not expect anything, except perhaps the appearance of providers that would dot all the i's. And yeah, if I pay 10 euro for the 2000 points core VPS and see that conditional Nexusbyte/Clouvider/BuyVM selling VPSes with 4000-4500 points for the same price I will consider moving to them if I host not only VPN on my VPS but huge projects that need better core. Doesn't competition work like that?
Sure, if such scores are permanent.
VPS is shared hosting of the era. You are not going to get similar scores every time unless you are on an empty node.
Additionally, determining the value of a hosting account purely on the benchmark score itself is kinda sad as you are removing the human element.
Well, tbh, I do benchmark only in case if I see things are not working as they are supposed to be and that too frequently. What if something is really broken, then atleast I can inform my host about it to get it fixed.
Sure, one sex is fine. But OP isn't running once.
He is running it frequently like waves of PMS hitting him. That is an abuse.
May we consider your high posting rate per minute on LET as a form of abuse that wasting the server's resources and even the resources of your keyboard
Seeing you bothered to look up the post count which is a meaningless record on its own or another, it is sad indeed.
Seriously, how many times did you run the benchmark?
Wwwwait, a sec, are you running one right now?
Theres a benchmark for everything these days, wonder if theres one for my wife...
Japanese do have one. Their highest ranking tag for porn is NTR after all...
I do benchmarks every morning. I test myself for strength with simple pull-ups on the horizontal bar. Today I made 23 pull-ups!
Does it also test how long she can keep up with my bull...ehm bandwidth?
Next test: posting on LET while doing pullups.
Ya, you said it and we believed it.
Next time post a video, if we have not seen , its not there.
Benchmarks on vps is only good to tell what the cpu can take.
Although if you already know what cpu they use you can look up online what the benchmark of the cpu is already.
If you need something in particular for your application or server. Look at the type of cpu the host uses.
As far as the clients on per core. It depends on host to host.
I stick to the 25% rule per VPS depending on package.
Where did he say he's running it more than once or even implied such? He's asking about benchmarks LET users post. He's even asking hypothetically about buying a $10 server (ha!), not that he has one. It's like you're being a dick because you failed to read the thread and just venting about benchmarks and directing it at him instead of at actual fucking excessive benchmarking users.
Again, a user tries to discuss something on a discussion forum and gets shit on. It's creepy how you use any opportunity to say "PMS". I've got 5 sisters, and you say "PMS" more times in a month than my whole family, combined, in our entire lives. Fucking creepy how often PMS is on the very front of your brain. Someone in your life should have pointed this out to you already.
Benchmarking can be useful as a reference. When I get a new server, I'll benchmark it and have an idea of hardware and the load on it. Later, if in actual usage of the server I find something taking longer than expected, I'll run it again and see if there is a noticeable difference. Over the years, I've alerted providers to failed RAIDs on TWO occasions as a result. Being able to tell the provider in the ticket specific dates with DRASTIC performance changes (eg, consistent 500-800MB/s disk speeds for years dropping to 47MB/s suddenly) gets much more responsive support. I shouldn't have to inform a provider their hardware is bad, but I'm glad that I could before more data was lost.
Servers are like vehicles. Some people just need a reliable car that gets then from point A to point B and no special performance is needed. Some need cars that can do a certain 0-60, pull a boat, or use the HOV lane. The specs of the vehicle gets you so much, a test drive gets you so much, and usage over time gets you the rest when evaluating the overall value of a vehicle. Not all buyers see the value in a $100K car when a $25K car does the job. Hope the point isn't lost.
You do know we are talking about VPS here not an actual server? By your logic your only benching 2 cylinders out of an V8 and the complaining it doesnt run like an V8.
If another user on the same VPS node is maxing all of it out while youre running your bench you are not gonna get any sort of accurate reading from your bench in the end.
Many provider don't give use to 100% CPU.
You should quote who you're talking to for context. By who's logic? No one is going to know what you're talking about since no one else was talking about cylinders or V8's.
Getting a number on what CPU cycles was available is useful. If the server is busy, the score will be low. If server is light, score will be higher. Performance will vary.
Client: "How many clients do you fit on 1 core?
Provider: "Yes."
If it is cheap,
That core isn't yours to keep,
The answer you seek,
Will always end in lie,
So don't waste your money,
Because The End is Neigh.
Not even Yes...
Client: How many clients do you fit on 1 core?
Providers:
Deathly silence
VPS is able to manage per day approximately 1000–1100 users/ visitors per day. It is best for the medium and small size of business where you provide services for thousands of clients
If this isn't sustainable pricewise, how are there providers out there who can offer dedicated cores at similar price points?
Its pretty straightforward to figure out an estimate: just take the cost of a similar dedicated server and divided by 2 x the cost of the VPS. Some hosts are more greedy, some are more lax.
IMO, there is no answer other than that the number of customers per-core depends on the price. Some providers have "dedicated core" plans, sometimes it's an HT / SMT core {two customers per each}, sometimes it's a physical dedicated core.
^ Don't expect to find that on a low-budget.
If you have performance requirements and need a guarantee, the best bet is honestly to just ask for a custom plan unless it's upfront stipulated. ---> Some hosts will give you rough estimates if you ask, others won't, some will work with you ... albeit at a significantly higher cost.
Number of customers per core doesn't equate in any meaningful way to performance though. Low core count, medium density (per core) systems can often beat a high core count low density if you have some unique-need. {like game servers}
Personally I suggest looking towards GSP's {Game Server Providers} if you demand low density, high clock frequency, modern processor / high IPC. {or just move up to dedicated}
-- BuyVM might be a good option for you, DrServer, or NFO-Servers (they don't post on here).
I would agree with that. After running a bunch of cloud servers benchmarks I get pretty similar results from time to time. One core servers usually score around 3000 or more in Geekbench 4.