can you still improve this miner?

-Both Rigs are Win10 Pro x64
-Drivers 376.66
-GPU's 2x 1070 EVGA ACX 2.0 Superclocked in 1070 rig and 1x MSI GTX 1060 3GT OC (B01KUCL25Y) in 1060 rig.
-1070 rig is a gigabyte z97x-ud3h-cf motherboard and 1060 rig is an optiplex 390 desktop with PS replaced.
-1070 rig has I5-4460 and optiplex 390 1060 rig has an I5-2500s
-1070 rig has a PC power & cooling 950W Silencer MK-II PS and Optiplex has Thermaltake SP-650
-No risers.
-The 1070's all have two 8 pin connectors coming from the power supply (no adapters) and the 1060 has one 6 pin power cable direct from the power supply
-miner --server us1-zcash.flypool.org --port 3333 --user address.phoop --solver 1 --pass x --cuda_devices 0 1 2 Note: I've tried using all solvers 0-3 per ewbf's suggestion and I've also tried running the cards individually by using cuda_devices selection with the same results.
I start the miners from an account with administrator privileges, but I don't right click and run as admin.  If I run it from an admin command prompt I do get the same problem though.  I assume if it was admin related it wouldn't work at all, not work for 5-10 minutes then stop working.
The benchmarks will run on whatever you have set as your primary display.  I have one display connected to each video card so to make benchmark run off other gpu I just have to set the other monitor as my primary display.  No SLI mode is not on with hardware or software bridge.  No I have not tested the cards one by one in main GPU slot.  I have this case http://www.antec.com/product.php?id=705069&lan=nz so adding/removing cards can be a bit of a pain.

There are some very smart people here that can code this stuff, @EWBF, you are one of them, so please dont take offense to this question.  Are we at a point of technology or knowledge that hinders progress?

You tried to rescan the driver's devices in Devices Manager?
Sometimes the drivers get corrupted... Microsoft advanced housekeeping.

I've met this. Lowering clocks did help.

Just got the same glitch, restarting the script worked fine.
Some time ago I had kind of the same problem with Claymore and my AMD's.
hint to EWBF: Setting the (-r 240) "restart by user" switch to every 4 hours solved my problem.
[/quote]

Thanks for the restart by user hint, didn't even know it existed till I saw your post.  I've set it up on all my miners as a just in case measure.  I believe I have found a solution for Thread exited with code: 46 , Thread exited with code: 77 , and Thread exited with code 30 .  After hours of tinkering and sifting through logs in EVGA precision X I noticed a correlation between power % and this error.  Every time it would happen on my 1060 and 1070 rigs they would be within 4% of 100%.  One of the times I even saw GPU voltage at 0.000% the line before the miner crashed.  I've found on both my rigs that were having this problem, setting the power % to 94% has stopped the problem and allowed me to returned to an overclocked (rather than under clocked) configuration.  My 1060 is now getting me a consistent 280-300 sol/sec with these overclocked settings on an MSI 1060 3GT OC 3GB.
http://i128.photobucket.com/albums/p193/dtawom/1060_zpsghnrimkv.jpg

I have never believed the story that ewbf has copied someone or other miners, I'm sure that if he wants he can improve his miner without help from others.

"miner --help or -h" ?

need to put a space between devices, there are examples in the miner help.

--cuda_devices 0 1 2

thx...figured it out by googling for 15 minutes   my mixed GTX Rig keeps dying so i needed to switch both GTX 1060 off. now it works. will need to plug them to an another PC as it seems