The only thing I don't understand is why GPU0 is GPU1. This makes the fan speed and temp readout in the miner not line up with the correct cards.
Fan temp on GPU0 = GPU5 in Phoenix Hashrate of GPU0 = GPU1 in Phoenix
Fan temp on GPU1 = GPU1 in Phoenix Hashrate of GPU1 = GPU2 in Phoenix
Fan temp on GPU2 = GPU2 in Phoenix Hashrate of GPU2 = GPU3 in Phoenix
Fan temp on GPU3 = GPU3 in Phoenix Hashrate of GPU3 = GPU4 in Phoenix
Fan temp on GPU4 = GPU4 in Phoenix Hashrate of GPU4 = GPU5 in Phoenix
Fan temp on GPU5 = GPU6 in Phoenix Hashrate of GPU5 = GPU6 in Phoenix
Fan temp on GPU6 = GPU7 in Phoenix Hashrate of GPU6 = GPU7 in Phoenix
Fan temp on GPU7 = GPU8 in Phoenix Hashrate of GPU7 = GPU8 in Phoenix
Fan temp on GPU8 = GPU9 in Phoenix Hashrate of GPU8 = GPU9 in Phoenix
We've seen similar things on AMD cards because of broken support for bus ID in the drivers but not on Nvidia cards. The ordering is based on the bus ID (ascending order) and the fan/temp reporting is matched to the bus ID as well, but in same cases like this one the NVML library obviously doesn't report the correct bus ID for the GPU5 and it ends up as the first. Could you press the 's' key on your keyboard and copy the list of GPUs with their bus IDs (it may be easier to do it from the log file than from the screen)?
17576:08:08:13.086: main Available GPUs for mining:
17576:08:08:13.086: main GPU1: GeForce GTX 1070 (pcie 1), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU2: GeForce GTX 1070 (pcie 2), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU3: GeForce GTX 1070 (pcie 3), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU4: GeForce GTX 1070 (pcie 4), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU5: GeForce GTX 1070 Ti (pcie 5), CUDA cap. 6.1, 8 GB VRAM, 19 CUs
17576:08:08:13.086: main GPU6: GeForce GTX 1070 (pcie 6), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU7: GeForce GTX 1070 (pcie 8 ), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU8: GeForce GTX 1070 (pcie 9), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main GPU9: GeForce GTX 1070 (pcie 10), CUDA cap. 6.1, 8 GB VRAM, 15 CUs
17576:08:08:13.086: main Eth: Accepted shares 2171 (32 stales), rejected shares 2 (0 stales)
17576:08:08:13.086: main Eth: Incorrect shares 12 (0.38%), est. stales percentage 1.47%
17576:08:08:13.086: main Eth: Maximum difficulty of found share: 76.2 TH (!!!)
17576:08:08:13.086: main Eth: Average speed (5 min): 302.663 MH/s
17576:08:08:13.086: main Eth: Effective speed: 291.26 MH/s; at pool: 290.99 MH/s
17576:08:08:13.086: main
17576:08:08:15.723: main Eth speed: 303.298 MH/s, shares: 2171/2/12, time: 20:43
17576:08:08:15.723: main GPUs: 1: 34.044 MH/s (221/5) 2: 33.216 MH/s (234) 3: 33.701 MH/s (274) 4: 32.071 MH/s (226) 5: 34.033 MH/s (254) 6: 34.187 MH/s (237/7) 7: 34.237 MH/s (260) 8: 33.813 MH/s (230) 9: 33.995 MH/s (237)
17576:08:09:01.576: main GPU1: 51C 35%, GPU2: 63C 35%, GPU3: 48C 35%, GPU4: 61C 49%, GPU5: 59C 35%, GPU6: 51C 35%, GPU7: 59C 35%, GPU8: 49C 35%, GPU9: 56C 35%
@PhoenixMiner - Looks like the issue is with the odd card of the rig, the Zotac 1070 Ti AMP Edition. I added that 9th card when Windows did the fall creators update allowing Windows to find more than 8 cards. The other 8 are Zotac 1070 AMP Extremes with a fan speed of 35%. The 1070 Ti, GPU4, GPU5 in Phoenix, is in the GPU4 location of my OC tool ( Nvidia Inspector 1.9.7.8 ) with a fan speed of 49%.
(Nvidia Inspector allows Over Clock on more than 8 cards which can be set by batch file or GUI and gives all info that GPUZ provides. On top of that it's more stable than MSI's After Burner). (The ONLY downside to Nvidia Inspector is you can't program a fan curve).
WHO EVER QUESTIONS PHOENIX ISN'T AS GOOD IF NOT MORE STABLE THAN CLAYMORE? 12 INCORRECT SHARES IN 20HRS @ 302 MH/s ON 9 CARDS.
Edit: I thought only getting 12 incorrect shares was amazing at those speeds....