My second ZEC + XMR+ ETH thread builds info links thoughts and photos.

[arielbit]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

i have an issue with reduced hashrate if my cpu is mining too...the solution is free a core, like 3 cores will mine and 1core is free..my miner is an old socket 775 with core2quad cpu so i don't know if newer rigs have this issue.

[philipma1957]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

[AriesIV10]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

I had 6 MSI RX 470s and 5 would hash at 20mh and 1 would hash at 4mh.  I did not know what this was or how to fix it no matter what I did so, I returned the MSI RX 470s and am getting Sapphire RX 480s next week.  I am hoping that I do not see the same problem.

[xleejohnx]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

I had 6 MSI RX 470s and 5 would hash at 20mh and 1 would hash at 4mh.  I did not know what this was or how to fix it no matter what I did so, I returned the MSI RX 470s and am getting Sapphire RX 480s next week.  I am hoping that I do not see the same problem.

i had this issue when i didnt have the pcie lanes set too x1 in the bios. all of them were trying to run x16

[AriesIV10]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

I had 6 MSI RX 470s and 5 would hash at 20mh and 1 would hash at 4mh.  I did not know what this was or how to fix it no matter what I did so, I returned the MSI RX 470s and am getting Sapphire RX 480s next week.  I am hoping that I do not see the same problem.

i had this issue when i didnt have the pcie lanes set too x1 in the bios. all of them were trying to run x16

I will try that!

[GabryRox]

FYI for those that may have missed it... MSI has extended their rebate offering that was originally only valid the last week of September throughout the end of October.

470's have a $15 rebate making the net cost only $185 and the 480's have a $20 rebate.

[AriesIV10]

FYI for those that may have missed it... MSI has extended their rebate offering that was originally only valid the last week of September throughout the end of October.

470's have a $15 rebate making the net cost only $185 and the 480's have a $20 rebate.

Great information!  This rebate is for 1 unit per family. 

[philipma1957]

FYI for those that may have missed it... MSI has extended their rebate offering that was originally only valid the last week of September throughout the end of October.

470's have a $15 rebate making the net cost only $185 and the 480's have a $20 rebate.

Great information!  This rebate is for 1 unit per family. 

yes i got one from amazon  runs well

[AriesIV10]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

I had 6 MSI RX 470s and 5 would hash at 20mh and 1 would hash at 4mh.  I did not know what this was or how to fix it no matter what I did so, I returned the MSI RX 470s and am getting Sapphire RX 480s next week.  I am hoping that I do not see the same problem.

i had this issue when i didnt have the pcie lanes set too x1 in the bios. all of them were trying to run x16

I will try that!

I have changed every setting in the bios and still can't solve the "1 gpu at 4mh" while the others are cruising along at 20mh each.  Does anyone else have another solution?

[GabryRox]

Update on my 6800k-based rig:  got the 3rd MSI GPU up and running last night and got the same 19.4 H/s as the other 2, so good news on that front.

However, my ASRock x99 Extreme4 Mobo apparently requires extra power to it to support anything past 2 GPUs, so I had to use my 2nd molex cable to provide that needed power. Since my other molex cable is dedicated to GPUs #2 & #3, I am now in need of at least 1 additional molex cable.

Question #1:  can I buy any add-on molex cable, or does it have to be one exactly made for my modular Rosewill Quark series 1000w Plat PSU?  It seems that these cables/extensions are generic one-size-fits-all, but I thought i remembered reading somewhere that with modular PSU's, you must have the cables made exactly for the specific PSU.

I can't seem to find any of these online so Im probably calling it the wrong thing... but it's the on that has 6-pins on the PSU side (for peripherals) and 4-pin molex power on the other end.


In addition, since adding GPU #3, GPU #2 has been running hot. It's been averaging about 72 while the #1 & #3 are in the 59-62 range.  I'm assuming this is because #2 has #3 about 3 inches to its left (from its fan side), so I want to extend that gap by a few inches in the hopes it will help with ventilation.

Question #2:  similar to q1, can i just use a generic extension like this?

http://www.ebay.com/itm/8-Pin-PCIE-Power-Extension-Cable-for-GPU-Video-Card-PCI-Express-PCI-E-8-20CM-/141930636051?hash=item210bb9a313%3Ag%3AJhEAAOSwzaJX8Y6Z

[philipma1957]

Update on my 6800k-based rig:  got the 3rd MSI GPU up and running last night and got the same 19.4 H/s as the other 2, so good news on that front.

However, my ASRock x99 Extreme4 Mobo apparently requires extra power to it to support anything past 2 GPUs, so I had to use my 2nd molex cable to provide that needed power. Since my other molex cable is dedicated to GPUs #2 & #3, I am now in need of at least 1 additional molex cable.

Question #1:  can I buy any add-on molex cable, or does it have to be one exactly made for my modular Rosewill Quark series 1000w Plat PSU?  It seems that these cables/extensions are generic one-size-fits-all, but I thought i remembered reading somewhere that with modular PSU's, you must have the cables made exactly for the specific PSU.

I can't seem to find any of these online so Im probably calling it the wrong thing... but it's the on that has 6-pins on the PSU side (for peripherals) and 4-pin molex power on the other end.


In addition, since adding GPU #3, GPU #2 has been running hot. It's been averaging about 72 while the #1 & #3 are in the 59-62 range.  I'm assuming this is because #2 has #3 about 3 inches to its left (from its fan side), so I want to extend that gap by a few inches in the hopes it will help with ventilation.

Question #2:  similar to q1, can i just use a generic extension like this?

http://www.ebay.com/itm/8-Pin-PCIE-Power-Extension-Cable-for-GPU-Video-Card-PCI-Express-PCI-E-8-20CM-/141930636051?hash=item210bb9a313%3Ag%3AJhEAAOSwzaJX8Y6Z

the pcie cable is 20cm = 8 inches it says it is 18ga copper so assuming that is true is should work.

4pin molex are harder to find  and they are specific to the psu.{in general}

but I have that psu and I have that cable let me look.

I found it. send me a pm with a mailing address .  I can ship it out to you.

[AriesIV10]

Not sure if this is related, but I found I need to toggle automatic/manual voltage control inside of watt-man in order to make some cards work at full hash rate. Not sure if the cards are getting too warm/outside the power budget or what is going on.

My cards will from from 29MH/s down to 7MH/s when they get into this funky state.

Been meaning to post this question for a while

yeah  i saw that issue.  7mh is a heat downclock.



4mh  is weirder    as it comes and goes , varies  board to board .  Sometimes I would need to mix cards about on the boards.  or move  the dvi display jack or load drivers

I had 6 MSI RX 470s and 5 would hash at 20mh and 1 would hash at 4mh.  I did not know what this was or how to fix it no matter what I did so, I returned the MSI RX 470s and am getting Sapphire RX 480s next week.  I am hoping that I do not see the same problem.

i had this issue when i didnt have the pcie lanes set too x1 in the bios. all of them were trying to run x16

I will try that!

I have changed every setting in the bios and still can't solve the "1 gpu at 4mh" while the others are cruising along at 20mh each.  Does anyone else have another solution?

Notice GPU2 with 5mh on the first pic and GPU1 Power at 18W.  I have tested them.  If I have 1 GPU plugged it will be the low one no matter what slot I put it into.  The rest run fine.  Can someone help?

[philipma1957]

It is hard to get rid of the 4mh issue.

What Amd driver do you have?

16.9.2
16.9.1
16.8.3
16.8.2
16.8.1

Try the easy way first.

Take your cable. Hdmi dvi or display and move it one card then boot

Say 1-2-3-4-5-6


Try the 1. Boot see what happens boot again see what happens
Try the 2 boot see what happens boot again see what happens

Try all six. Some may work some may not

Some may give no picture at all.

If all six cards tested one at a time as the display driver don't solve the problem then remove the driver

And try a different driver

What sucks is 6 cards means 12 tests per Amd driver

And there are five working drivers which means it could be 60 tests!

What is worse after the 60 tests it still may not work!
But there are two plus points it does use less power
And the scaling is the same.  Ie watts per mh are equal.

I have had the problem and no matter what I did. On one board with four rx 480 sapphire 8gb it never worked
So I pulled one sapphire put in a Msi rx 470 connected. Via Dvi and boom it worked

Three rx 480 and one rx 470

[citronick]

It is hard to get rid of the 4mh issue.

What Amd driver do you have?

16.9.2
16.9.1
16.8.3
16.8.2
16.8.1

Try the easy way first.

Take your cable. Hdmi dvi or display and move it one card then boot

Say 1-2-3-4-5-6


Try the 1. Boot see what happens boot again see what happens
Try the 2 boot see what happens boot again see what happens

Try all six. Some may work some may not

Some may give no picture at all.

If all six cards tested one at a time as the display driver don't solve the problem then remove the driver

And try a different driver

What sucks is 6 cards means 12 tests per Amd driver

And there are five working drivers which means it could be 60 tests!

What is worse after the 60 tests it still may not work!
But there are two plus points it does use less power
And the scaling is the same.  Ie watts per mh are equal.

I have had the problem and no matter what I did. On one board with four rx 480 sapphire 8gb it never worked
So I pulled one sapphire put in a Msi rx 470 connected. Via Dvi and boom it worked

Three rx 480 and one rx 470

force mobo to use gen1 for all PCI slots - will this help?

[xleejohnx]

It is hard to get rid of the 4mh issue.

What Amd driver do you have?

16.9.2
16.9.1
16.8.3
16.8.2
16.8.1

Try the easy way first.

Take your cable. Hdmi dvi or display and move it one card then boot

Say 1-2-3-4-5-6


Try the 1. Boot see what happens boot again see what happens
Try the 2 boot see what happens boot again see what happens

Try all six. Some may work some may not

Some may give no picture at all.

If all six cards tested one at a time as the display driver don't solve the problem then remove the driver

And try a different driver

What sucks is 6 cards means 12 tests per Amd driver

And there are five working drivers which means it could be 60 tests!

What is worse after the 60 tests it still may not work!
But there are two plus points it does use less power
And the scaling is the same.  Ie watts per mh are equal.

I have had the problem and no matter what I did. On one board with four rx 480 sapphire 8gb it never worked
So I pulled one sapphire put in a Msi rx 470 connected. Via Dvi and boom it worked

Three rx 480 and one rx 470

force mobo to use gen1 for all PCI slots - will this help?

he tried the force mobo

reset wattman. sounds crazy but i have to do it when modding bios

[citronick]

It is hard to get rid of the 4mh issue.

What Amd driver do you have?

16.9.2
16.9.1
16.8.3
16.8.2
16.8.1

Try the easy way first.

Take your cable. Hdmi dvi or display and move it one card then boot

Say 1-2-3-4-5-6


Try the 1. Boot see what happens boot again see what happens
Try the 2 boot see what happens boot again see what happens

Try all six. Some may work some may not

Some may give no picture at all.

If all six cards tested one at a time as the display driver don't solve the problem then remove the driver

And try a different driver

What sucks is 6 cards means 12 tests per Amd driver

And there are five working drivers which means it could be 60 tests!

What is worse after the 60 tests it still may not work!
But there are two plus points it does use less power
And the scaling is the same.  Ie watts per mh are equal.

I have had the problem and no matter what I did. On one board with four rx 480 sapphire 8gb it never worked
So I pulled one sapphire put in a Msi rx 470 connected. Via Dvi and boom it worked

Three rx 480 and one rx 470

force mobo to use gen1 for all PCI slots - will this help?

he tried the force mobo

reset wattman. sounds crazy but i have to do it when modding bios

is there a way to uninstall wattman, but use the latest 16.9.2 drivers?

[philipma1957]

Wattman is part of the driver if you delete the right file it may not work and the driver may also fail.  A lot of trial and error would be needed to see if a disable wattman and a working amd driver for rx 470 and or rx 480 is possible.

 I have detected the 4mh bug  on rx 470's rx 480's and on my biostar z170's mobo's and my asrock h81 mobo

 my four card asrock extreme3  z170 does not have the issue using 3 cards. but I will be using  a fourth card the one below so we will soon see if it shows up.



Here is good deal at newegg. spend use masterpass get 25 off a 200 dollar purchase.  the card below has a rebate.

I had to get a usb cable to go over 200.

Lastly the card will allow for a possible fix of the 4mh issue.  use 5 msi and this power color on the troubled board.

[AriesIV10]

The 4mh Issue:

Phil, I tried all tests that you recommended.

-only 1 gpu installed (I removed 5 gpu's until I can resolve this issue)
-set the Bios to Gen1 for all slots
-tried moving the 1 gpu to different slots on mobo to include installing directly on the slots on the mob
-uninstalled and reinstalled all versions of drivers (16.9.2, 16.9.1, 16.7.3)  I now have 16.9.2 installed


Still having the 4mh issue on the 1 card that I have installed.  Anyone have a solution?

[philipma1957]

The 4mh Issue:

Phil, I tried all tests that you recommended.

-only 1 gpu installed (I removed 5 gpu's until I can resolve this issue)
-set the Bios to Gen1 for all slots
-tried moving the 1 gpu to different slots on mobo to include installing directly on the slots on the mob
-uninstalled and reinstalled all versions of drivers (16.9.2, 16.9.1, 16.7.3)  I now have 16.9.2 installed


Still having the 4mh issue on the 1 card that I have installed.  Anyone have a solution?

so  this one card  with 4mh put it in a box  and do a second card different  does it work?

I was always able to get the trouble to leave.

But are you saying  you have six cards and  only one specific card does the 4mh issue?

if you put the 4mh card aside the other 5 all work?


Some have had that  issue and sent the one card in for rma.


in all of my cases  my 4mh issues  are not card specific  ie  one card would show 4mh  pull it put in a completely different card  and  still have a 4mh issue.

so my cases have all been board-driver-card conflict issues.

your description seems to say  it is always the same card.

some write ups have said this and rma was the only fix.

I am really trying to understand if you are card specific  vs floating  around  with your problem.

BTW i hate this issue  as it has been the hardest one to fix for me.

right now I have 19 cards on 5 machines all doing 20-25mh

and I dread that the 4mh issue can reappear-----------------------it has at times.

[AriesIV10]

The 4mh Issue:

Phil, I tried all tests that you recommended.

-only 1 gpu installed (I removed 5 gpu's until I can resolve this issue)
-set the Bios to Gen1 for all slots
-tried moving the 1 gpu to different slots on mobo to include installing directly on the slots on the mob
-uninstalled and reinstalled all versions of drivers (16.9.2, 16.9.1, 16.7.3)  I now have 16.9.2 installed


Still having the 4mh issue on the 1 card that I have installed.  Anyone have a solution?

so  this one card  with 4mh put it in a box  and do a second card different  does it work?

I was always able to get the trouble to leave.

But are you saying  you have six cards and  only one specific card does the 4mh issue?

if you put the 4mh card aside the other 5 all work?


Some have had that  issue and sent the one card in for rma.


in all of my cases  my 4mh issues  are not card specific  ie  one card would show 4mh  pull it put in a completely different card  and  still have a 4mh issue.

so my cases have all been board-driver-card conflict issues.

your description seems to say  it is always the same card.

some write ups have said this and rma was the only fix.

I am really trying to understand if you are card specific  vs floating  around  with your problem.

BTW i hate this issue  as it has been the hardest one to fix for me.

right now I have 19 cards on 5 machines all doing 20-25mh

and I dread that the 4mh issue can reappear-----------------------it has at times.

The 4mh Issue

in all of my cases  my 4mh issues  are not card specific  ie  one card would show 4mh  pull it put in a completely different card, in a different slot on mob and  still have a 4mh issue.

Example:  This is a different gpu in slot #3 - Same Problem: