Just now noticed this. My GPU, 300 task que, and my CPU, 100 task que, have no _0 work. None. Never seen that before. What's up?

Just now noticed this. My GPU, 300 task que, and my CPU, 100 task que, have no _0 work. None. Never seen that before. What's up?

That is unusual! Are you doing Separation or N-Body tasks? Because Separation has 832343 tasks left to send out while N-Body has 112242 tasks left to send out, that's alot for both which would make sense for someone to finally setup the Server to send _1 etc tasks to pc's that are trustworthy and fast to get thru the backlogs. Your RAC is almost 1 million which means you are zipping right thru the tasks, or have a boatload of pc's, but either way you can quickly go thru alot of resends.

have no _0 work.

Coincidence. Had that once too since my return to Milkyway a month ago simply because my computer only got work when huge batches timed out on some abandonned hosts. Right now I have only _0 (except one WU) to compensate. ;-)

---

That is unusual! Are you doing Separation or N-Body tasks? Because Separation has 832343 tasks left to send out while N-Body has 112242 tasks left to send out, that's alot for both which would make sense for someone to finally setup the Server to send _1 etc tasks to pc's that are trustworthy and fast to get thru the backlogs. Your RAC is almost 1 million which means you are zipping right thru the tasks, or have a boatload of pc's, but either way you can quickly go thru alot of resends.

Just separation. When the big kerfuffle was going on with n body, something like 17,000,000 waiting for validation, I did run n body then for quite a while. Yeah, that R9 280x is crankin' 1 every 28 seconds(running 5 in parallel). And 23 threads running on the 3900x. plus a 4 core laptop and a 2 core old desktop. I still think it would benefit the program to have an option for wingman only tasks.

That is unusual! Are you doing Separation or N-Body tasks? Because Separation has 832343 tasks left to send out while N-Body has 112242 tasks left to send out, that's alot for both which would make sense for someone to finally setup the Server to send _1 etc tasks to pc's that are trustworthy and fast to get thru the backlogs. Your RAC is almost 1 million which means you are zipping right thru the tasks, or have a boatload of pc's, but either way you can quickly go thru alot of resends.

Just separation. When the big kerfuffle was going on with n body, something like 17,000,000 waiting for validation, I did run n body then for quite a while. Yeah, that R9 280x is crankin' 1 every 28 seconds(running 5 in parallel). And 23 threads running on the 3900x. plus a 4 core laptop and a 2 core old desktop. I still think it would benefit the program to have an option for wingman only tasks.

Seti tried that many many years ago and it kinda sorta did work for them, they only sent the resends to hosts that had a history of returning tasks in 24 hours, and of course they got burned a few times, but for the most part it did work