Within the past weeks, I have made the experience that while sometimes the finished WUs are uploaded with a good speed of nearly 1000 KB/sec, sometimes the speed is awfully low, like 60 KB/sec. So it takes forever to get a 500MB file uploaded, resulting in the loss of the <24 hours / <48 hours credit bonus.

How can this happen, with only about 80 users currently? I cannot remember this kind of problem at former times when there were 1000 or more users crunching and uploading their WUs.

No idea why the problems with the upload server obviously cannot be fixed :-(
Again, a finished task is uploading at ~60 kb/sec :-(

So anyone can figure out how long the upload lasts for a 499 MB size file.

So anyone can figure out how long the upload lasts for a 499 MB size file.

now the next tasks which uploads with ~80 kb/sec :-(

Am I the only one who experiences this problem, or are there others, too?

Am I the only one who experiences this problem, or are there others, too?

All my uploads take 2+ hours because I have a miserable 1MBit upload pipe that only produces about 758KB in reality.

Oh how I wish I had a fiber connection. One can only dream I guess.

Maybe Gpugrid upload server is not the bottleneck.
I can bring an example of just the opposite.
I catched one of my hosts at the very end of uploading a 506,29 MB file.
It took 11 minutes and 15 seconds to upload, at rates above 700 KBps.
This host has a WiFi connection to an access point in my home network.
Internet access is through a fiber optic router.

All my uploads take 2+ hours because I have a miserable 1MBit upload pipe that only produces about 758KB in reality.

Oh how I wish I had a fiber connection. One can only dream I guess.

Great idea and awesome tech but comes at price. IMO it's a bit excessive even if you live in a remote area and are in dire need of a working alternative.

Just f.ex. current pricing for my service address in Germany would be 558€ plus 99€/month service fee. Nah

about 2 hours ago, a finished task was uploaded with ~ 980 KB/s, right now another task is uploading with ~ 80 KB/s.
There is never any value inbetween - either it's very fast, or very slow. This is rather strange, isn't it ?

I'm thinking on it. My brother has an order in for it. I think I will wait for him to test it and give me some feedback. He has an even slower connection than I do.

I'm a bit leery of using it for an always on connection since I have more data going upstream all the time compared to the intermittent topup downloads.

Probably not so strange considering only one host runs ALL the processes. Imagine the I/O gets bottlenecked regularly.

Probably not so strange considering only one host runs ALL the processes. Imagine the I/O gets bottlenecked regularly.

Strange indeed if the tasks go fast when another project is actively down/uploading.

Must be something to with the HTTP/S protocol handshakes. The get sorted out with another project in play but get bogged down when GPUGrid itself is uploading.

I still see the DDOS remediation from the project when several hosts from the same IP address are communicating with the project. I can wait several minutes and often the link is dropped when I try and view the website pages.

There is never any value inbetween - either it's very fast, or very slow. This is rather strange, isn't it ?