After trying for 4 straight days to fix  this it turned out to be a stupid hardware mistake.  For 2 GPU's I had them connected with 6-pin VGA connectors when they needed to be 8-pin.  I thought both cards supported either one, but that fixed it, so I guess not.  All good now though.

A new question for everyone, I think I got ahead of myself.  I just added a 3rd GPU, another RX580, and now the rig continuously crashes in certain cases.  Here are the 2 scenarios:

• With Claymore running it tries to set the DAG epoch for each GPU and crashes.  Happens there every time
• I bought a copy of ethOS, and even though it gets past the DAG epoch, it does still eventually crash

I also now have a 1200W power supply instead of 750W, and all 3 GPU's are on risers.  Although, it doesnt seem to matter where I put the GPU's (MOBO or risers).

Any ideas from anyone on what I could try?  I've tried tweaking a lot of the BIOS with no luck, like disabling Cool and Quiet, some of the Power Controls, etc.  Maybe there is something within ethOS, or the Claymore settings I could try beyond -gsers,-wd,-di, etc.

Thanks

Another small update here...after a couple of days the card thats still in the mobo, and not the riser, keeps failing.  I think because of overheating because it starts up, mines for 10 seconds or so, I notice thru claymore that the temperature sky rockets, then either crashes the machine or freezes.  So, during the process of fixing that problem I found out that setting gpu_max_alloc_percent to 90 and not 100 (as suggested many places) keeps the gpu from overheating and has been stable ever since.

Will post here again if anything new comes up. 

woohoo!  bought a riser (thx same-day amazon) and put the 2nd card in and viola, everything works as expected.  25 mh/s each

so....not sure exactly what that means and why i needed a riser.  as far as my issue goes this is resolved, but i wouldnt mind hearing from someone who knows why this would matter, so that other users might benefit from it.

thanks again everybody.

oh interesting, i'll try that too, maybe my mobo has an equivalent setting.

and i'll research the driver issues more.  initially i thought i saw some people using a similar config and it was fine.

Yep tried both independently and got expected hashrate ranges.  Thats why its so strange, why degraded when used at the same time.

I think at this point I'm going to try a riser so at the very least I've got a 3rd slot I can test against.  If that doesnt work then I might try another GPU. 

Will post here with results.

Thanks for the help so far Afox, fanatic26

Latest progress:  I did some research into the motherboard and it sounds like one of the PCIE slots says its 16x but really runs at 4x.  Does anyone think that could cause a reduction in hashrate?  I've seen videos online that suggest there is no performance hit between 16x, 8x, 4x, and 1x (which people use with risers), what do people think?

Good call....looks like 65C and 49% fan for the slow GPU, and 63C and 40% fan for the normal GPU, so thats probably not it. 

Additionally, after looking at this again, is there a chance that the reduction in hashrate would have anything to do with the cards being placed so close together?  Probably a complete long shot dumb question, but I do see that the fan for one GPU is covered by the back of the second GPU, and may be causing it to create more heat.  Maybe a riser would help, but is that even a problem anyone has seen related to hashrate performance?

2 GPU:

Radeon RX 480
Radeon RX 580

Motherboard on Linux: ASRock 970M Pro3 AM3+

Running Claymore Miner or ethminer, I get a normal hashrate on one card at about 24.5 mh/s but on the other card I only get 12.5 mh/s.  I've tried running them separately, swapping to different PCIE slots, but nothing seems to change the performance of the "other" card.

Any ideas on how to troubleshoot this?  I'm not that familiar with how to tweak hardware for a Linux machine, but with direction I can try whatever people suggest.

Thanks for the help.