PrimeGrid Times

[Gilthanis]

I thought maybe people would want to make a list of the times they see on different hardware here.
Device Run time Cpu time Credit application
NVIDIA GeForce 8400 GS (512MB) driver: 34201 - 103,458.71 - 865.25 - 3,371.00 PPS (Sieve) v1.39 (cudaPPSsieve)
NVIDIA GeForce GT 430 (1024MB) driver: 38494 - 6,819.26 - 297.45 - 3,371.00 PPS (Sieve) v1.39 (cudaPPSsieve)
9,786.10 - 652.20 - 416.84 Genefer 17 Low v3.18 OCLcudaGFN17LOW)
1,087,278.21 - 370,038.90 - 141,596.12 Genefer 21 v3.18 (OCLcudaGFN)
365,865.57 - 71,145.36 - 31,617.41 Genefer 20 v3.18 (OCLcudaGFN20)
NVIDIA GeForce GTX 780 (3072MB) driver: 38205 - 616.89 - 260.89 - 3,371.00 PPS (Sieve) v1.39 (cudaPPSsieve)
106.88 - 102.25 - 25.42 Genefer 15 v3.18 (OCLcudaGFN15)
234.92 - 228.31 - 97.11 Genefer 16 v3.18 (OCLcudaGFN16)
698.44 - 299.78 - 416.82 Genefer 17 Low v3.18 (OCLcudaGFN17LOW)
3,802.53 - 190.58 - 4,043.00 AP27 Search v2.02 (OCL_cuda_AP27)
1,782.94 - 534.42 - 1,592.63 Genefer 18 v3.18 (OCLcudaGFN18)
61,622.39 - 692.56 - 141,177.41 Genefer 21 v3.18 (OCLcudaGFN)

NVIDIA GeForce GTX 970 (4095MB) driver: 38813 - 276,020.26 - 5,257.40 - 692,815.24 Genefer 22 v3.19 (OCLcudaGFNWR)
6,925.38 - 85.75 - 7,377.81 Genefer 19 v3.19 (OCLcudaGFN19)

Edit: These numbers are running stock speeds and 1 work unit at a time.

 

[MN Scout]

These times are with a NVIDIA 1080TI models. The EVGA 1080 TI SC Gaming and the ASUS ROG STRIX GTX1080TI GAMING GeForce 11GB OC Edition. The ASUS card runs ridiculously cool and can easily run at 1.9GHz at 72C forever where the EVGA downclocks to 1.6GHz to keep the same 70 to 73C. The ASUS also runs its fans speeds at around 50% whereas the EVGA is always at 90%.

Most of these times are at a 70% power target on the EVGA. If I remember right the card stayed at full clocks around 1.9GHz on the Genefer 15 to 19 tasks, then on the 20+ tasks it started throttling down because the temps were so hot. I need to get some new times with 100%PT on the ASUS card, since it can handle it.

The ASUS card gets another 50k points when running 3up PPS-Sieve tasks whereas the EVGA keeps the same PPD as 2up. I think the higher sustained clocks on the ASUS helps get the 50k. There is too many variables, so I just run the cards at 2up all the time on PPS-Sieve. There may be options for 2-upping the lower numbered Genefers, but I haven't tried yet.