Markfw
Moderator Emeritus, Elite Member
- May 16, 2002
- 25,610
- 14,590
- 136
The default team is now doing 2 BILLION PPD ! They were at 300 million
To me, that is irrelevant, if I can't get any work. I am getting pissed. This is the worst in the 20 years I have been folding. Why add these new units if they can't create work for them ?Here's the link to FAH server status.
That being said, my LAN right now is connected to my Internet Essentials connection, which until a few days ago, was only 15/2, but Comcast used the Covid-19 thing as an excuse and PR move to finally bump the connection speed up to 25/3 going forward (the minimum standard for "broadband", according to the US Federal Gov't, btw). I noticed that when my F@H program on my main PC was uploading a WU, it was tying up my entire 3.7Mbit/sec upload for a noticeable amount of time. Imagine the bandwidth that they must have, to receive so many WUs from so many people, so often! And downloading them too. (That does not seem to max out my 27Mbit/sec download speed.)Internet connection bandwidth for example. Their outbound and inbound traffic must be mindboggling.
And here is a folding forum thread which explains what the terms at the serverstats page mean.Here's the link to FAH server status.
I have the option (without changing any hardware) to get 150/150 megabit. Its like $20 a month more (I am on fiber-optic, all the way to my house, Frontier). I also volunteered a 7551 EPYC server to help them out, and no response.That being said, my LAN right now is connected to my Internet Essentials connection, which until a few days ago, was only 15/2, but Comcast used the Covid-19 thing as an excuse and PR move to finally bump the connection speed up to 25/3 going forward (the minimum standard for "broadband", according to the US Federal Gov't, btw). I noticed that when my F@H program on my main PC was uploading a WU, it was tying up my entire 3.7Mbit/sec upload for a noticeable amount of time. Imagine the bandwidth that they must have, to receive so many WUs from so many people, so often! And downloading them too. (That does not seem to max out my 27Mbit/sec download speed.)
Well, thats cool, but as I said, regardless, why add new units when you can't process them, its the horse before the cart (or is it the cart before the horse ??) situation. You just make everyone upset. I spend like $800 a month on electricity, and $8000-12,000 a year on hardware to help these causes. When they can't even use my help, why bother providing the service ?Each of their three top work servers has >100 TeraByte disk space (serverstats shows only the free space left), and is handling 10,800 work units per hour (3.0 work units per second). With 79 MB (?) per work unit, and 55 MB typically per result, that's 237 MB/s (1900 Mb/s) outbound traffic for new work sent and 165 MB/s (1320 Mb/s) inbound traffic for results received, just for one of their top work servers. On top of that comes the internal traffic e.g. for moving the results to further processing.
The big question then is the classic one in computing - how do we fix the bottleneck? Is there a good place to donate? I'd rather donate cash if my RX5700 can't do anything for them due to infrastructure issues.Each of their three top work servers has >100 TeraByte disk space (serverstats shows only the free space left), and is handling 10,800 work units per hour (3.0 work units per second). With 79 MB (?) per work unit, and 55 MB typically per result, that's 237 MB/s (1900 Mb/s) outbound traffic for new work sent and 165 MB/s (1320 Mb/s) inbound traffic for results received, just for one of their top work servers. On top of that comes the internal traffic e.g. for moving the results to further processing.
There are only two ways that this (sudden inrush of new volunteers, plus occasional Folders switching F@H back on due to the news) could have been averted: If they hadn't decided to work on COVID-19 in the first place (I for one am glad they do this work), or if they had but had kept it a secret (that would have been absurd).but as I said, regardless, why add new units when you can't process them
Please don't be upset about the fact that there are suddenly so many people interested in helping.You just make everyone upset.
Thanks for the info; then my assumption about networking constraints was mistaken.A source close to the issues says it's a storage bandwidth bottleneck now.
Well, how do they fix that ? I now have 1 of 12 video cards working.....Its getting worse.A source close to the issues says it's a storage bandwidth bottleneck now.
Tony, I already posted, but how can we have a race when NO units are available. Zero points and zero points ? we need to get this delayed until they work out their problems.Tentative race start time is 10am EST-USA tomorrow.
Whoever did the custom stats tracking last time ( @Ken g6 ? ) can you have a script that soon??? :/
I tend to agree, if Stanford isn't sending out WU regularly to all participants, we might as well hold a lotto for the winner of this "race". :|Tony, I already posted, but how can we have a race when NO units are available. Zero points and zero points ? we need to get this delayed until they work out their problems.
I mean seriously, I have 12 boxes, and NONE can get a unit ? What are the odds anyone with even 3 cards, can get a unit. Maybe I will try to get a CPU unit going.I tend to agree, if Stanford isn't sending out WU regularly to all participants, we might as well hold a lotto for the winner of this "race". :|
I mean seriously, I have 12 boxes, and NONE can get a unit ? What are the odds anyone with even 3 cards, can get a unit. Maybe I will try to get a CPU unit going.
Edit: tried CPU, 2 tries, NOTHING
I just tried re-setting-up my script. The stats server isn't working.Tentative race start time is 10am EST-USA tomorrow.
Whoever did the custom stats tracking last time ( @Ken g6 ? ) can you have a script that soon??? :/
This.we need to get this delayed until they work out their problems.