Genoil's ZEC miner: AMD ZCash miner for Windows - 0.6

[Walrusbonzo]

EDIT: Oh I forgot to add that cards #0 and #2 seem slower than card #1, even though they are all Hawaii cards with the exact same clocks. The difference is that those two are in PCI-Ex1 slots, but card #1 and the Pitcairn(#3) are in PCI-Ex16 slots. Could the PCI-E bus or the CPU be causing a bottleneck? (CPU is a 3.7GHz Thuban and I set genoil process to above-normal priority. No other programs running other than firefox and afterburner and a couple other low-resource background ones.)

I'm witnessing the same. I've got a 7x 290x rig. Only one card is connected via 16x PCI-e, the others are all 1x PCI-e. The card on 16x slot runs 35~40 sol, the other cards sometimes get into the 30s, but mostly 25-30 sol.

Having studied CPU, GPU and GPU memory usage whilst mining ZEC I thought that the bottleneck to mining appeared to be elsewhere, and so possibly PCI-e bus. My thoughts were that given the generally CPU intensive nature of mining ZEC that the CPU was heavily involved and transferring commands/data over the PCI-e bus to the GFX cards. This would explain why the cards on 1x slots run a bit slower.

In another rig I have 3x RX480, two are on 16x slot, 1 is on 1x slot. The 1x slot card is consistently slower, all settings are identical across the three cards.

i'm going to guess that ZEC lends itself very nicely to dual mining with Eth. Hope Claymore comes up with the goods.

[1715301705]

1715301705

[1715301705]

1715301705

[1715301705]

1715301705

[1715301705]

1715301705

[Tmdz]

0.5 seems more buggy to me then 0.4.1

I noticed on my rx 470 it will crash within an  hour, I have a reset timer for every hour.  So Im using 0.4.1 since that will run for more time, usually hours before crashes happen and the reset script keeps it going.

[bardacuda]

EDIT: Oh I forgot to add that cards #0 and #2 seem slower than card #1, even though they are all Hawaii cards with the exact same clocks. The difference is that those two are in PCI-Ex1 slots, but card #1 and the Pitcairn(#3) are in PCI-Ex16 slots. Could the PCI-E bus or the CPU be causing a bottleneck? (CPU is a 3.7GHz Thuban and I set genoil process to above-normal priority. No other programs running other than firefox and afterburner and a couple other low-resource background ones.)

I'm witnessing the same. I've got a 7x 290x rig. Only one card is connected via 16x PCI-e, the others are all 1x PCI-e. The card on 16x slot runs 35~40 sol, the other cards sometimes get into the 30s, but mostly 25-30 sol.

Having studied CPU, GPU and GPU memory usage whilst mining ZEC I thought that the bottleneck to mining appeared to be elsewhere, and so possibly PCI-e bus. My thoughts were that given the generally CPU intensive nature of mining ZEC that the CPU was heavily involved and transferring commands/data over the PCI-e bus to the GFX cards. This would explain why the cards on 1x slots run a bit slower.

In another rig I have 3x RX480, two are on 16x slot, 1 is on 1x slot. The 1x slot card is consistently slower, all settings are identical across the three cards.

i'm going to guess that ZEC lends itself very nicely to dual mining with Eth. Hope Claymore comes up with the goods.

Thanks, that pretty much confirms it for me then. Been waiting for Zen and Kaby Lake to upgrade the system. If ZCash is going to remain more profitable and this can't be fixed by more optimized software then I'll be sure to prioritize getting a board with multiple x16 slots.

[elzium]

If all is right, this pieceofshit miner is very stable, ran all day today without errors on several different rigs with different card combos, pitcairn to ellesmere, all clocked.
Me like.

And u guys with so many hardware, dedicate some weaker card to solve start.bat as it is.

[toptek]

EDIT: Oh I forgot to add that cards #0 and #2 seem slower than card #1, even though they are all Hawaii cards with the exact same clocks. The difference is that those two are in PCI-Ex1 slots, but card #1 and the Pitcairn(#3) are in PCI-Ex16 slots. Could the PCI-E bus or the CPU be causing a bottleneck? (CPU is a 3.7GHz Thuban and I set genoil process to above-normal priority. No other programs running other than firefox and afterburner and a couple other low-resource background ones.)

I'm witnessing the same. I've got a 7x 290x rig. Only one card is connected via 16x PCI-e, the others are all 1x PCI-e. The card on 16x slot runs 35~40 sol, the other cards sometimes get into the 30s, but mostly 25-30 sol.

Having studied CPU, GPU and GPU memory usage whilst mining ZEC I thought that the bottleneck to mining appeared to be elsewhere, and so possibly PCI-e bus. My thoughts were that given the generally CPU intensive nature of mining ZEC that the CPU was heavily involved and transferring commands/data over the PCI-e bus to the GFX cards. This would explain why the cards on 1x slots run a bit slower.

In another rig I have 3x RX480, two are on 16x slot, 1 is on 1x slot. The 1x slot card is consistently slower, all settings are identical across the three cards.

i'm going to guess that ZEC lends itself very nicely to dual mining with Eth. Hope Claymore comes up with the goods.

Thanks, that pretty much confirms it for me then. Been waiting for Zen and Kaby Lake to upgrade the system. If ZCash is going to remain more profitable and this can't be fixed by more optimized software then I'll be sure to prioritize getting a board with multiple x16 slots.

even on the on board x 16 slot have the same issue so i sty with how it is and hope zcash holds it value and the software is tweaked or fine tuned more.

 

[toptek]

I'm aware of the nicehash miner since genoil still mention Nvidia on his github i just put a legitimate question.
Thats all and OpenCL is not AMD exclusive..omg

hope this helps in the read me


--gpus,-g <x y z ...>     set gpu(s)               -g 0 1 2 3
--plat,-P <n>            set OpenCL platform id   -P 0
- --plat--P paramter is useful for systems with iGPU enabled or mixed AMD/Nvidia GPUs.


so it won't  mater to much unless there mixed .

[Cryptozillah]

EDIT: Oh I forgot to add that cards #0 and #2 seem slower than card #1, even though they are all Hawaii cards with the exact same clocks. The difference is that those two are in PCI-Ex1 slots, but card #1 and the Pitcairn(#3) are in PCI-Ex16 slots. Could the PCI-E bus or the CPU be causing a bottleneck? (CPU is a 3.7GHz Thuban and I set genoil process to above-normal priority. No other programs running other than firefox and afterburner and a couple other low-resource background ones.)

I'm witnessing the same. I've got a 7x 290x rig. Only one card is connected via 16x PCI-e, the others are all 1x PCI-e. The card on 16x slot runs 35~40 sol, the other cards sometimes get into the 30s, but mostly 25-30 sol.

Having studied CPU, GPU and GPU memory usage whilst mining ZEC I thought that the bottleneck to mining appeared to be elsewhere, and so possibly PCI-e bus. My thoughts were that given the generally CPU intensive nature of mining ZEC that the CPU was heavily involved and transferring commands/data over the PCI-e bus to the GFX cards. This would explain why the cards on 1x slots run a bit slower.

In another rig I have 3x RX480, two are on 16x slot, 1 is on 1x slot. The 1x slot card is consistently slower, all settings are identical across the three cards.

i'm going to guess that ZEC lends itself very nicely to dual mining with Eth. Hope Claymore comes up with the goods.

Thanks, that pretty much confirms it for me then. Been waiting for Zen and Kaby Lake to upgrade the system. If ZCash is going to remain more profitable and this can't be fixed by more optimized software then I'll be sure to prioritize getting a board with multiple x16 slots.

even on the on board x 16 slot have the same issue so i sty with how it is and hope zcash holds it value and the software is tweaked or fine tuned more.

 

I have one system with 4 x 480 8GB gpus onboard motherboard.
It works like a charm on ETH (4 x 30.5Mhs)
But when i mine ZEC the cards do like 20s/s each with this miner (should do like 30-40). There is a bottleneck somewhere while mining ZEC with many gpus onboard for sure.
I am going to rebuild that system in the next coming days (H81 + risers)

On the other hand i have two other rigs with 3 x 480 cards onboard and they are running normal on ZEC (other mobo).
I guess it could be that other mobo that dont like ZEC and many used pcie slots.

[markus124]

can u guys add a command that "restart" the miner if it sees a card running at 0 hashes for 5 min. That would be very helpfull.
U think u can do it? (right now we have random restarts but sometimes it happens if the miner is working great)

[MantaMine]

Have pointed all my rigs back to ETH, the stability of the current ZEC miners out there are just too much of a PITA to deal with.  Tried genoil and it is great when all the cards are hashing, but that seems to be the exception and not the rule.  Appreciate his efforts getting that miner out and it is helping with keeping the distribution of the coin on a more even keel, but I just can't dedicate the time to babysitting rigs.  I think things will change in the next three days though  

[bardacuda]

I have one system with 4 x 480 8GB gpus onboard motherboard.
It works like a charm on ETH (4 x 30.5Mhs)
But when i mine ZEC the cards do like 20s/s each with this miner (should do like 30-40). There is a bottleneck somewhere while mining ZEC with many gpus onboard for sure.
I am going to rebuild that system in the next coming days (H81 + risers)

On the other hand i have two other rigs with 3 x 480 cards onboard and they are running normal on ZEC (other mobo).
I guess it could be that other mobo that dont like ZEC and many used pcie slots.

So in the system where you get 20S/s are they on x1 risers? On the two systems that run normal what size pci-e slot are they plugged into?

[toptek]

can u guys add a command that "restart" the miner if it sees a card running at 0 hashes for 5 min. That would be very helpfull.
U think u can do it? (right now we have random restarts but sometimes it happens if the miner is working great)

try this

start it first

Startup.bat

Code:

@echo off
echo starting miner
:loop
if exist output.txt del output.txt
time /t
start /MIN nh.bat
:loop2
timeout /T 60 /NOBREAK >nul

findstr /C:"submitted and accepted." output.txt
if %errorlevel%==0 (
echo Shares Submitted!
goto loop2
)else (
taskkill /F /IM genoil.exe
timeout /T 1 /NOBREAK >nul
echo re-starting miner
goto loop
)
|| (
goto loop2
)

make sure all bat fike are in the same folder as the software miner  
name the below one  nh.bat or whatever  whatever you want  but if you do make sure to change the line in the startup.bat to this bats new name oi it won't work .


nh.bat

 you also have to edit this with your pool info and user .

Code:

set host=%COMPUTERNAME%
echo %host% 

genoil.exe -c you pool info  -u ZEC address or pool login .%host% -p z >output.txt
echo quit >output.txt
exit

 


Then use start up bat it will start nh.bat and close it to tack bar and log to a output.txt file in the  folder you put everything in .

 to see the output txt file  i use this https://www.baremetalsoft.com/baretail/ I  like it better then seeing the black screen all the time .

I got this off the Zcash forums the other day when this forum was being attacked . i did not make them they came from some else.

[YIz]

0.5 with silentarmy v2 works faster than 0.42 but crashes every once in a while. switched back to 0.42 until there's a fix.

[Walrusbonzo]

can u guys add a command that "restart" the miner if it sees a card running at 0 hashes for 5 min. That would be very helpfull.
U think u can do it? (right now we have random restarts but sometimes it happens if the miner is working great)

try this

start it first

Startup.bat

Code:

@echo off
echo starting miner
:loop
if exist output.txt del output.txt
time /t
start /MIN nh.bat
:loop2
timeout /T 60 /NOBREAK >nul

findstr /C:"submitted and accepted." output.txt
if %errorlevel%==0 (
echo Shares Submitted!
goto loop2
)else (
taskkill /F /IM genoil.exe
timeout /T 1 /NOBREAK >nul
echo re-starting miner
goto loop
)
|| (
goto loop2
)

make sure all bat fike are in the same folder as the software miner  
name the below one  nh.bat or whatever  whatever you want  but if you do make sure to change the line in the startup.bat to this bats new name oi it won't work .


nh.bat

 you also have to edit this with your pool info and user .

Code:

set host=%COMPUTERNAME%
echo %host% 

genoil.exe -c you pool info  -u ZEC address or pool login .%host% -p z >output.txt
echo quit >output.txt
exit

 


Then use start up bat it will start nh.bat and close it to tack bar and log to a output.txt file in the  folder you put everything in .

 to see the output txt file  i use this https://www.baremetalsoft.com/baretail/ I  like it better then seeing the black screen all the time .

I got this off the Zcash forums the other day when this forum was being attacked . i did not make them they came from some else.

Correct me if I'm wrong, but this will only work where you are running a instance of genoil miner for each GPU? If you run one instance for 6 cards and 5 cards stop mining, the 6th card could still be getting accepted shares and therefore the miner doesn't restart?

[ryantc]

as others have mentioned, use the guide here
https://bitcointalk.org/index.php?topic=1667366.0

the bat file basically kill the mining process every T seconds, mine is on its second round

[toptek]

can u guys add a command that "restart" the miner if it sees a card running at 0 hashes for 5 min. That would be very helpfull.
U think u can do it? (right now we have random restarts but sometimes it happens if the miner is working great)

try this

start it first

Startup.bat

Code:

@echo off
echo starting miner
:loop
if exist output.txt del output.txt
time /t
start /MIN nh.bat
:loop2
timeout /T 60 /NOBREAK >nul

findstr /C:"submitted and accepted." output.txt
if %errorlevel%==0 (
echo Shares Submitted!
goto loop2
)else (
taskkill /F /IM genoil.exe
timeout /T 1 /NOBREAK >nul
echo re-starting miner
goto loop
)
|| (
goto loop2
)

make sure all bat fike are in the same folder as the software miner  
name the below one  nh.bat or whatever  whatever you want  but if you do make sure to change the line in the startup.bat to this bats new name oi it won't work .


nh.bat

 you also have to edit this with your pool info and user .

Code:

set host=%COMPUTERNAME%
echo %host% 

genoil.exe -c you pool info  -u ZEC address or pool login .%host% -p z >output.txt
echo quit >output.txt
exit

 


Then use start up bat it will start nh.bat and close it to tack bar and log to a output.txt file in the  folder you put everything in .

 to see the output txt file  i use this https://www.baremetalsoft.com/baretail/ I  like it better then seeing the black screen all the time .

I got this off the Zcash forums the other day when this forum was being attacked . i did not make them they came from some else.

Correct me if I'm wrong, but this will only work where you are running a instance of genoil miner for each GPU? If you run one instance for 6 cards and 5 cards stop mining, the 6th card could still be getting accepted shares and therefore the miner doesn't restart?

yup yuor right why i don't use it to much and only with one card , if  I use it in a rig with more then it all ways seems to stop one off ,  i think that's all he wants or you wanted if not my bad i don't use it with three cards one all ways shut down i should have said that to, i use .41 it seems to work best with my 470 and ,05 with every thing else in one card rigs till the minier is either tweaked more or some thing replaces it .

[Cryptozillah]

I just found this and now have that running on all my rigs.
https://forum.z.cash/t/genoils-zec-miner/3791/1823

Restarting miner if any gpu goes to 0

[molitar]

I actually use Windows Task Scheduler to do the same thing.  Restarts every 10 minutes.  Not only that if I select option to allow running when logged off the window is hidden.  So works great.  I have it set to trigger on start or login.  That way I know it's running all the time.

[frozn0000]

https://s15.postimg.org/vkctewh8r/gpu_1.jpg



https://s15.postimg.org/vlmr8bj2j/gpu_2.jpg

Im using my 7950 shappire rig of 6 gpus getting a media of 16 hashes per gpu. is this good for this type of gpu??
Is there any ways of use 99% of my gpu usage, im stuck at around 85%?

Thanks for the help if u could. 

If the GPU is sat between 80~85% as your graph shows it suggests that they're memory bandwidth/latency limited. Most 7950/7970 will happily do atleast 1500MHz, start there. Failing that you probably want to use a custom BIOS with tighter timings, usually copying a memory strap from 1250/1375 to 1500 does the trick. When I used to have 7950s I was using tighter timings(The Stilt's) and memory clock was around 1700MHz. This provided excellent performance for scrypt and Eth type crypto.

Kind of slow mine does 20-25 sol/s with genoils miner 0.4.2 on stock settings

Thanks for this.  I have a folder setup with this for each of my gpus. 280x x 5.  After driver crashes for that paticular gpu genoil won't mine.  Not your fault in know but is there a way to auto fix the driver issue so genoil will mine?  After driver crashes it only mines 0sols

[toptek]

I just found this and now have that running on all my rigs.
https://forum.z.cash/t/genoils-zec-miner/3791/1823

Restarting miner if any gpu goes to 0

I saw that to and even tried it but had to reformat that PC after I  used it but not because of that file, i couldn't find it again . ty .

find it and here is original version

https://github.com/Genoil/ZECMiner/issues/40 in this post .

[polylogic]

sometimes the miner shows 0 sol/s, sometimes it shows connection error, sometimes miner stopped and window disappear. What can be done to solve these ?

Please tell me a way so my miners restart if there is any error at least

change the pool, flypool is currently very unstable with 4-5000 active miners