Thanks for the suggestions, I was going to limit my CPU slot to 12 cores (of the 16 available) so I can still use the system for other work. I have noticed by monitoring my system with GPU-Z while folding on the GPU, if I allocate 12 cores to folding I get my max GPU Load, then as I increase cores to 15, I lose a 1-3% on my GPU Load for each core I engage in folding. At 15 cores I am down 5-8% on GPU Load from my max at 12 cores. As expected, if I fold on all 16 cores while the GPU is folding, production really drops. So for me at least, the one core per GPU seems to be a minimum, and my best production is if I have about 4 cores/threads not actively engaged while the GPU is folding.

Do you have any sense if it is better to setup two CPU slots with 6 cores each vs. a single CPU slot with 12 cores?

12 core would be better overall than 2 x 6's. Simpler and faster.

It is amazing to see that Nvidia Saturn V just churn through work units.

Huge volumes of work being done, but still systems are being overwhelmed even with the new work servers. I do note some new donors are dropping off unsure if it because of the occasional shortages of work or just bored that the cure not discovered in 3 weeks.

I'm still only getting units to work on about half of my available time. I'm fine with demand exceeding supply. Hope it stays that way.

DKeane wrote:I'm still only getting units to work on about half of my available time. I'm fine with demand exceeding supply. Hope it stays that way.

This is mainly due to work servers being overloaded even more servers coming on line.
I intend to rebuild my old folding rig this weekend and have an 'apprentice' who I will supervise fitting a new cpu cooler and swapping out the old (but reliable) win7 OS on an HDD to win 10 on an SSD, mother board too old for a M2. This a great time to do maintenance.

On the Server Stats Page, does a server with its row in gray indicate that server is "down"? I notice only some of them in gray actually indicate an error or warning, but all indicate zero public or beta jobs. Or are the gray servers where WUs are sent by clients? I notice they all have a status of "Accept".

Also, does anyone know what the "Has CS" column indicates? What does it mean if a server doesn't have a "CS"?

Sorry for the noob questions...

I actually try not to look at the servers too much at present, so much going on with them, things are getting better but still frequently am unable to get work. Sorry can't answer that question.

SAK917 well you asked about the downsides of being a beta tester, am experiencing extreme downsides at present in an unusual way, I have more than 10 completed GPU work units that the servers don't recognize or are refusing. Seems this affects everyone doing these 3 beta projects.

I agree kiore, I have processed more than a few GPU WUs that are being rejected or unable to upload since I enabled the "advanced" option as well. I hope that the work is still good for the science, but am not sure how if they aren't accepting the results. Is there a forum thread discussing this issue?

SAK917 wrote:I agree kiore, I have processed more than a few GPU WUs that are being rejected or unable to upload since I enabled the "advanced" option as well. I hope that the work is still good for the science, but am not sure how if they aren't accepting the results. Is there a forum thread discussing this issue?

I contacted the researcher and they restarted the server but still have some units built up waiting to send and as they always try and send 4 at a time it sort of makes it worse until a couple have gone. I have shut done my GPUs pretty much for the past day CPUs and working fine.

OK folding GPUs again but still have 4 completed beta units stuck on send from last night, hopefully they will send tonight when server not so saturated,

Also, does anyone know what the "Has CS" column indicates? What does it mean if a server doesn't have a "CS"?

Sorry forgot to answer this, CS= Collection Server . The beta workunit CS server is what has been giving me issues for the past day or two. I expect it now jammed up with waiting units completed by non beta team members using the flag to get the 'special treatment'.

My jammed beta work units have finally sent and my output starting to climb back towards normal.

Thanks for the info on the servers kiore. Looks like they are getting the servers dialed in now, I am not having too much trouble getting WUs or reporting back results and am finally back over 2M points a day. Do you know if they gave up on the AWS servers that I saw on the server list for a while when they were expanding capacity? Looks like they are all gone and they are using Azure servers now?

I am not sure about the servers but people are reporting being able to get and send work more easily now. My output measured by ppd has returned to the more normal 4mppd after dropping as low as 1mppd.

I'm still sitting at around half capacity from a ppd standpoint.

I think I am running at about 70-75% of max output the last day or so. The worst is at night when I am not able to keep an eye on things and I might go 4-6 hours without a WU before I am able to reset the counter. I understand why they do it, but the algorithm implemented for preventing a server from getting hammered with WU requests is a bit frustrating.

I understand the frustrations not getting work or just waiting for hours for a unit. New work is coming through beta testing at an incredible rate and being released to advanced after just a day or 2 so they are pumping out work.

Looks like the WUs are coming through nicely now.