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’ve had the old ks2 that was supposed to be 300mbit but I got 10g/10g and pushed over 250tb in a month so I’m not sure how they monitor it but it’s definitely not capped at the threshold it’s meant to be (or atleast wasn’t)
Yeah OVH doesn’t care about bandwidth usage. I’ve been using 400TB of bw on their cheap 7 eur vps
It’s probably too much of a hassle to monitor each dedicated server given how big OVH is and how many servers they have. APAC is the only region they actively put limits on.
May I asks how you used 400TB on a 7 euro vps? Free P0rn?
If someone interested, i have an upgraded KS-LE-C grabbed few days ago from beloved @fredo1664 to transfer: 1630v4 64GB RAM 2x480GB SSD 500mbps UP / 1 GBPS down - GRA1 (renew price: 12.99€)
Just got a brand new gigahost dedi + already 2 KS-LE-B. Would be more than enough for the year
Transfer price: 29€
Tell us more about your gigahost dedi
Tor Relay!
Yes do tell
Yes, tell us!
I'll always have the image in my head that @remy looks like young Travolta IRL now
Yes i sort of look like him but not as handsome and not as young (so not at all sadly)
allthemtings already poking his head into my inbox
Why do you assume it’s a good deal right away?!
I have luck only for ovh lottery
If you know me then you know in a fiend for a giga dedi I try every year on the Black Friday ones but always sadly miss
It's not those gems. It's an auction server
Not cheap, just wanted that location.
@FAT32 has very good gigahost dedi
But he does not want give one to me
This year it will happen
I gave up on the idea. I don't think it's possible without a bot
or i'm just too old and too slow (or both)
I grab nothing on megathread.
Anyone got a spare
SYS-STOR-LE-2 ?
LLMs ?
yis
Thanks for the quick transfer
Also just had a KS-STOR delivered in GRA - No hardware upgrades but 10g down and 500mpbs up like others
If anyone has a SYS-STOR-LE-2 in EU and want to sell it, let me know.
What llm you running and how is the throughput?
usually no dense LLM's or small dense LLM's.
Mostly from unsloth, e.g Q5_K_M or higher.
For example
https://huggingface.co/unsloth/aquif-3.5-Max-42B-A3B-GGUF
Also good ones.
https://huggingface.co/unsloth/Qwen3-30B-A3B-Thinking-2507-GGUF
https://huggingface.co/unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF
If you really want the best you can fit on a 64gig kimsufi.
https://huggingface.co/unsloth/GLM-4.5-Air-GGUF
4bit should run tight with small context size.
How fast those LLM(s) are? With CPU?
The best I get here with CPU-only are 2 to 3 response tokens/s for complex tasks (coding etc.).
(Different larger LLMs though).
Just let them write some easy text (travel plan for Puerto Rico or the like) is much faster though.
dense models above 8B are slow.
However non dense models, such as the 30B, 3B active params, are fast enough to chat.
Isn't it worse than other major models in every way? Isn't the only reason to use it privacy?