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
You can run http://www.tpc.org/information/benchmarks.asp or http://www.spec.org/benchmarks.html. It fits more with Enterprise spec and what the larger companies would run. Please take a look.
And congrats on getting this up and going. I'm only making suggestions because I have been following this.
@concerto49 We'll check them out, we're considering some Phoronix stuff. But people are already complaining that the benchmarks take too long :P
Cheers @FTN_Kevin
@serverbear, when I first saw your script I felt a bit reserved, but honestly you made a great (and awesome) tool for everyone to use and perform a benchmark with a single command (awesome!!!) and for that I do thank you and congratulate you!
Keep up the great work!
Love,
-Lord Pie
Just would like to put love out there for server bear.
Just benched three servers today, awesome way to know how your server perform!
Virgin use BT Lines and that is about it, as BT was forced to open its infrastuctre to other competitors.
http://www.telegraph.co.uk/technology/broadband/8050218/Virgin-to-use-BT-lines-to-deliver-super-fast-broadband.html#
Talk Talk, Be, etc.. all use BT Lines, pretty much everything.
@serverbear: Cool site idea. I'd toss in another vote for making full benchmarks available to people instead of just the executive summary. I'm sending in one for a 256 BuyVM KVM in Buffalo right now.
It's done, you'll get your own report emailed to you.
We also added IOPS & Location info to these pages until we come up with a better solution for showing more benchmark info about each plan - http://serverbear.com/9806/digitalocean#view-benchmarks
How is the location determined? For one of our benchmarks it shows California (we only have servers in Florida).
Done via GEOIP. Not always accurate, but most scalable way of doing it. I had accidentally assigned a Hostigation bench to your plan, fixed now!
Hey guys, just a heads up.
We've taken a lot of feedback on board & we now allow you to browse through the public reports.
Example: http://serverbear.com/9773/hostigation#view-benchmarks
Quite a lot are still set to private, I'm just sorting through them over the next day or two. Moving forward you can set in the benchmark modal whether you want the report to be kept private or now.
NICE!! Keep up the good work @serverbear
Just doing 3x vr.org (India, NL, US)
(Can i do all for EDIS as well or am i not allowed as i am biased? I promise i just create them, run it, and dont use a special testing node.)
Most hosts supply their own set of benchmarks for plans.
I am sorry, but I wouldnt trust those (in general, not related with what William said, i trust him).
Is it possible to make it clear which is which ?
M
We could, I check everything before it goes on the site (plans & specs match etc).
Yes, i think it makes more sense if @ServerBear runs the tests himself. Otherwise the results should not be trusted much.
This does not take into consideration the case when the host offers a VPS on an empty node. Users must do it and even so, the host can make a fake user...
it will be unfair for the honest hosts I think.
M
@Maonique well, yes - volunteering users of a specific hosts can from time to time (i.e. maybe once a month) rerun the tests and resubmit the results to @ServerBear for comparison (or just give him access to rerun the tests).
Is not scalable, we aggregate the scores for runs from users, so over time you get to see more of a picture. One test means nothing.
Example: http://serverbear.com/9744/buyvm#view-benchmarks
@serverbear if possible make the users also submit the node on which they are as performance usually varies from node to node.
Most users won't know, we have that information in the tracert so we'll try to pull something out for the plan pages.
Nah you don't, we have no nodenames or IPs in the Traceroute - Both VRS (Linux-VServer) and KVM are bridged directly into the network and the next you see is the .1 (normal) or .3 (VSRP)...
@serverbear: The new layout is SUPER!
The 4-piece front page is fantastic, and most of the redirects have been fixed. Still one problem:
The only workaround at the moment is to
1. remember the name of the plan
2. click on the provider
3. go to benchmarks
4. search for the benchmark
Ideally, by clicking on the plan we should arrive directly at step 4!
But definitely, keep on the good work, it's already an amazing website!
Filtered by IOPS perhaps?
I have a feeling it is abused, Uncle is not always number 1 !!!
Dam chiterz !
:P
M
Yes, it is!
@serverbear:
Actually Host Virtual is missing from that lins with 93625 IOPS as it's somehow not considered to be using SSDs.
And VZSSD2 @ IperWeb is missing from this link for I have no idea what reason.
Please make a warning that it's OK Unixbench is going to take about 40 minutes on a multicore system! I really thought that something bad happened to the script.
VZSSD5 @ Prometeus done. 19500 IOPS
http://serverbear.com/benchmark/2012/09/22/IFc4iet9klkTfCAJ
I've just moved from my old provider having 7! IOPS to this package. A bit of a jump.
Software RAID? Their DD looks like it.
AFAIK they use SAN, but let's ask @Prometeus