Topic: My second ZEC + XMR+ ETH thread builds info links thoughts and photos. - page 37. (Read 147998 times)

I am having both black screens and 4mhs bug, so sad

and how do we do that mate?

Was wanting to do the same thing with 470's. But with the downturn in Zec all the sudden I don't know if it is worth it now to start. 

Make sure crossfire and ULPS are disabled. Regedit is the best way to go.

 

yeah sapphire and xfx have this issue. 

swap the monitor cable  from card to card.

also if you are an hdmi cable try using this

https://www.amazon.com/Cable-Matters-Gold-Plated-DisplayPort/dp/B004C9P9TM/ref=sr_1_2?


or if the card has a dvi  jack

https://www.amazon.com/dp/B014I8UU2W/ref=sr_ph_1?


all miners should have a few of the above  they have helped me with black screen issues and with the dreaded 4mh bug

Is it the card with the monitor hooked up to it? If so, put the monitor on a different card

I have exactly the same bug.

MB - Z97 GAMING 5. All cards are RX 480 Nitro+ (on powered risers)

Tried everything - always one card  is 4MH/s

I am using some reference rx480s some reference r470s and some nitro rx470s, has anyone managed to fix it?

Cheers

that bug has occurred with a few of us.

what cards do you use.

i am using the same build as phills. however I am having some trouble

I am having trouble having all 4 cards to mine at full speed, when all 4 of them are plugged in one gpu always mines at 4mhs instead of 24mhs, i have tested all cards individually and all work, i have testeded all slots individually and they work,
please please help me kind souls

I used all the same risers for all 6 slots in BTC Pro - everything was good.

Hey John, thanks for this tips.  Lone's tip did indeed fix my current speed issue, but what you mention about changing the x16 slot to Gen1 in the BIOS might actually solve the other issue I am having.  I don't currently have anything plugged into the x16 slot because I cannot get past post with a USB powered riser plugged in there.  Phil kindly sent me a x16 to x16 ribbon riser, but it isn't long/twistable enough to make work in my current configuration.  So... maybe setting this slot to Gen1 in the BIOS will fix this issue.  Can you provide and tips on how to do this?  Seems that every time i go into BIOS it takes me forever to find things.

The mere act of toggling it snaps my cards out of the funk.  Some I end up turning it 'off' and some I end up turning it 'on.  Makes no difference, just the act of fiddling with it.  This does not appear to happen on W10, as far as I can tell.

Check temps make sure thermals are good

Sometime when you run something in the x16 it gets all the resources it needs
then everything left over is divided out
You can change in bios of mobo an make all slots Gen1
just to make sure it isn't bugging out
I always change it even if it isn't needed


Edit: Seems you fixed it nevermind on this then lol

Thanks Lone!  Amazingly, it seems like just opening up Wattman has resolved this issue.  When I opened the AMD utility, I went to the Wattman tab for my first GPU.  The Voltage control was set to Auto so I toggled it to Manual.  But either right before or right around the time i did this (before clicking Apply, mind you), the speed started to increase on all 4 GPUs... pretty much to their expected rate.  I went ahead and applied the change to manual on all 4 cards and they are now running at the expected pre-overclocking speed.

Out of curiosity, is there any downside to leaving the Voltage control toggled to Manual?

Thanks again for the tip!

I have to go inside AMD WattMan from time to time after the cards heat up and toggle 'voltage control' on each card. This only needs to be done on my W7 boxes, the W10 boxes are fine.

Happy weekend everyone!  I have a weird issue that is occurring on my 2nd mining rig and was wondering if anyone had experienced similar and/or had any suggestions on what to tweak to try and fix it.  Here's the scenario...

Rig Details:
- i4790k CPU
- ASRock H81 Pro BTC R2 Mobo
- 16GB RAM (2x8)
- Win7 Pro
- 16 GB Virtual Memory
- 1200W Plat PSU
- Claymore ZEC Miner v8

I had 2 MSI 470's running for 3-4 days very stably.  With no tweaks at all, were running about 160 S/s.  After undervolting and setting core/mem clocks to 1300/1900, they were producing about 170-175 S/s.  I had an issue with strapping the first GPU, so reverting back to original BIOS.

After the molex cables from Phil arrived yesterday, I hooked up the 3rd GPU, a Sapphire 470 8GB.  This also ran fine, providing about 180 S/s with no mods (but it comes factory over-clocked).  In addition, the 2 MSI 470's were still running at previous speeds.

The problem started after I hooked up the 4th GPU, another MSI 470.  While my system recognized it fine, after running Claymore for a while, I realized that the hash-rate on all 4 GPUs had dropped.  Now, the 3 MSI's are barely getting 150 and the Sapphire about 155... so, all are about 25 S/s short.  Even stranger, when I set the core/clock speeds of the MSIs to 1300/1900 in afterburner, it has no apparent effect on speeds.

I am using the exact same config file with Claymore v8 as I am on my first rig and all other settings seem to also be inline, but all 4 cards on rig #2 are running about 25 S/s slower than they should be since I added the 4th GPU.  I double checked virtual memory and it is set to 16GB just like rig #1, so not sure what else to check.

Not sure this matters, but I have the 4 GPU's plugged into slots 1, 3, 4, & 5 (all x1 slots).  Slot #2 in the x16 slot and although Phil sent me a ribbon riser, I still need to figure out how to connect it given the design of my rig.  So, after running with USB powered risers in slots 1 & 3 stably, I added the Sapphire to slot 4 and then the 3rd MSI to slot 5.

Anybody seen anything like this before on this or similar Mobo?  I know there are tons of people running this Mobos at full speeds with 6 GPUs, so I must be missing something.  Any suggestions would be highly welcomed.

it isnt the pool thats the problem..
even at nicehash with claymore its all over the place ill be at 700 one min then 1300 the next

i only found that optiminer gives the right amount more consistent than claymore but it is slower than claymore anywhere from 20-50 sol/s

 v9 is apparently worth trying - quite a few folks have reported better hashrate out of it when it works at all.


 Scrypt mining, like Dnet work, was compute bound and soaked a LOT more power than ETH or ZEC do out of the same GPU at the same clocks.
 Probably easier and quicker to set up the dnet client for "high stress" testing though.


 Wasn't that long ago that a lot of "review" sites used the dnet client and/or Prime95 as a deliberate "stress test".

I recently switched from flypool to suprvova, since I've seen reports ( probably here ), as a test. It will take a few days to see
[/quote]

I notice the same on the flypool. It is usually 5-7% lower than the reported by Claymore.

The Dwarf is similar to that reported by Claymore. So it is quite strange, should be at least 2.5% lower due to the 2.5% fee.