I only had time for a small bunker, and have not been able to get many tasks since the start of the event... I will leave it running with another more reliable project at priority 0
Oh great, that will be around the same time that the prime grid tasks I have been running for the past 2 days are complete. Setting a reminder now to watch and prepare a bunker.
PrimeGrid it is! Shifted everything over but can't bunker due to travel. Barely had time for that. Let's hope some pendings last and valiidate at the right time.
I thought PPS Sieve had high GPU usage so I didn't bother checking it on my 1080Ti but it was very low even with CPU crunching stopped. 7:49 with 1 task. 9:06 with 3 tasks.
Even on the 1070 and 1070Ti they were around 10:40 and 9:18 at 2x. When at 3x they were running at 16:00 and 13:00. Improvements there too.
I thought PPS Sieve had high GPU usage so I didn't bother checking it on my 1080Ti but it was very low even with CPU crunching stopped. 7:49 with 1 task. 9:06 with 3 tasks.
Even on the 1070 and 1070Ti they were around 10:40 and 9:18 at 2x. When at 3x they were running at 16:00 and 13:00. Improvements there too.
On my 2080ti, just 1 PPS sieve task keeps utilization at 85%
I have been running the manual sieving application at the same time; I think it ends up adding 500k ppd, which is about a 20% increase
Looks like I get roughly the same PPD from running 2xPPS sieve at a time.
They have GPU apps but they are extremely CPU bound and every card requires many tasks to fully utilize the card if thats possible for higher core count systems. Many as in 7+ I've heard. Maybe more now. It basically a CPU project now. As the numbers get bigger it becomes less useful on the GPU.
CPU tasks appear to be mt at least up to 32t. Keeps the 1950x pegged with a quick dip every 20seconds or so. I can only get 64 tasks on 32t machine. Yikes 13gb memory and climbing. That might be its max as its going up and down between 10-13.4gb.
About 12:15 to 12:20 per task on my 1950x.
15.1GB of Virtual mem
13.6GB of memory according to BOINCTasks.
Some of the top computers are using CPU only. The top one using 2x RTX 2070s but some others after it are all or mostly CPUs.
Edit: Oh my... trying it on my NV GPUs in Linux. 2-5% GPU utilization. Clock speeds are way down like at stock non-boost and temps are in the upper 30s when they are usually like upper 60s-70s.
I'll be sticking with CPUs for a weekend run of Amicable. 2970WX running 42 threads, completes tasks in an average of 593 seconds, which is nearly 1 million ppd.
Tomorrow is the greatest day in auto racing with the Monaco GP, Indy 500 and Coke 600. Over 1260 miles to be watched tomorrow.
UK BOINC Team is slowly gaining on OCN in Amicable Numbers FB Sprint event at less than 300k between the teams. It'd be great if we could stay in 9th and place in a Sprint event.
That was my fault.
I had to take one of my machines off Amicable yesterday, but it is back up and crunching now. UK BOINC did pass us, but we might be able to catch them in the final four hours. We are currently just 68k behind.
A forum community dedicated to overclocking enthusiasts and testing the limits of computing. Come join the discussion about computing, builds, collections, displays, models, styles, scales, specifications, reviews, accessories, classifieds, and more!