Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux)

[notbatman]

Hi there,  msi rx470, with 16.11.4 I have blue screens and frezze , with 16.10.1 and 16.9.2 100% fans issuses. I'm mining for a year with these two versions of drivers, but now I have problems with one rig. What driver version is recomended, can anyone help me? All cards working at 0.98v, msi ab controls fans, clocks and voltage, hynix 1100/1900 and samsung 1100/2000. Rig with problem is Intel g1840, h81m plus, 4gb ddr3, 4 x msi rx470 4gb hynix.

MSI RX + Hynix ram = 5 weeks and counting RMA for me. MSI sucks balls.

[bulgar73]

what is now best for ETH, RX480 is sold out, RX570 or RX580 ??
I need 24-25pcs of cards for 4 new rigs.

Which PSU is best for 6x rig with AMD RX5xx series, MB will be Biostar TB85.

thnx for help
BTC

[AriesIV10]

what is now best for ETH, RX480 is sold out, RX570 or RX580 ??
I need 24-25pcs of cards for 4 new rigs.

Which PSU is best for 6x rig with AMD RX5xx series, MB will be Biostar TB85.

thnx for help
BTC

Buy just 1 RX5xx first.  Make sure that you are happy with these.  I have found that they are MUCH LESS EFFICIENT that the previous RX4xx Series!  Do this before you buy 24-25pcs of cards. 

As far as the PSU, EVGA is always a good choice with a little higher price.  This is my favorite on Amazon:
Rosewill Quark Series 1200 Watt Full Modular 80 Plus Platinum Certified ATX12V/EPS12V 1200 Power Supply Quark 1200
by Rosewill
Link: https://www.amazon.com/gp/product/B00VS8728G/ref=ox_sc_sfl_title_7?ie=UTF8&psc=1&smid=ATVPDKIKX0DER

[spyshagg]

what is now best for ETH, RX480 is sold out, RX570 or RX580 ??
I need 24-25pcs of cards for 4 new rigs.

Which PSU is best for 6x rig with AMD RX5xx series, MB will be Biostar TB85.

thnx for help
BTC

Buy just 1 RX5xx first.  Make sure that you are happy with these.  I have found that they are MUCH LESS EFFICIENT that the previous RX4xx Series!  Do this before you buy 24-25pcs of cards.  

As far as the PSU, EVGA is always a good choice with a little higher price.  This is my favorite on Amazon:
Rosewill Quark Series 1200 Watt Full Modular 80 Plus Platinum Certified ATX12V/EPS12V 1200 Power Supply Quark 1200
by Rosewill
Link: https://www.amazon.com/gp/product/B00VS8728G/ref=ox_sc_sfl_title_7?ie=UTF8&psc=1&smid=ATVPDKIKX0DER

Less efficient when underclocked to the same same clock/voltage?  I could swear I saw reviews online testing both @ 900mhz and the 5XX had >20w less power usage.

[AlainC]

Which PSU is best for 6x rig with AMD RX5xx series, MB will be Biostar TB85.

Why not 2 of 750W ? I found some German (not Advance) with PFC at 50€(55$) . you can connect them with some ATX coupler (AddPSU, Dual PSU)
There is some controversy about whether you should plug the card and the riser on the same PSU or oppositely all risers with the MB, and card where you can.

By the way, be careful before starting, ETH have gone so high, mining power exploded, and then difficulty. My date of benefit get worse since the bull rally (I got 1.9-2/month, now 1.6-1.7/month). Be careful. Dual Mining seems interesting.

[3dyyd]

I frequently observe the GPU usage dips to 0% for 0.5 second and recovers to 100%. My hash rate should be about 27.8Mh/s, but with these dips happening, it sometimes drops to 24Mh/s or even 8Mh/s
Anyone notice this? I'm running on Rx570 with moded BIOS.

[lonelyworks]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

[Teress]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

I have also this OpenCL hang aprox. 2 times a day on one card. I know which one it is but have absolutelly no idea how to avoid this hang. I tried to set stock voltages, low memory frequency, low core, nothing. It is RX 480 Nitro+ 4GB with Hynix memory. Currently running on 1145 core, 1885 mem, 27 Mhs, zero memory errors. But 2 hangs a day How to solve this hang error?

[bulgar73]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

I have also this OpenCL hang aprox. 2 times a day on one card. I know which one it is but have absolutelly no idea how to avoid this hang. I tried to set stock voltages, low memory frequency, low core, nothing. It is RX 480 Nitro+ 4GB with Hynix memory. Currently running on 1145 core, 1885 mem, 27 Mhs, zero memory errors. But 2 hangs a day How to solve this hang error?

sell it and buy another one, time is money

[powershot]

It depends a lot on the ambient temperature. If it is above 30ºC, consider it to be unnatural, but always check the exhaust air condition and inject cold air.

Remembering that the most efficient system would be that they expel the air in the open, however as we work with electronic components this becomes impossible because of the rains, but a draft is always welcome.

The good thing would be to put in an environment like I said earlier with constant or forced air exchange.

Spacing between GPUs is also important.

Standard temperatures for

R9 280x is 68 ° C ~ 75 ° C
Rx 470 4gb with Oc, 64 ° C ~ 70 ° C
In Rig, Space 8 cm

[TechPark]

Hi all, I have a weird issue with one of my rigs and I hope some may sound familiar and can be fixed;

I have one rig with 6x rx 580's (with bios mod) happily dual mining in windows 10. With bios mod they to 27 mhs, with parameters set in the config file they do 29.5 mhs.  It appears that overtime, it can be 5 minuts or 12 hours, something kicks in and resets the values I have given the miner in commands and they go back to 27 mhs.

These are the parameters;
-cvddc 980
-mclock 2150
-fanmin 80

Although I doubt fanmin works at all because Claymore miner never report a fan speed @ 80%, always less. Usually 70-75% and one card at 50% fanspeed

What memory timing do you use? What make and model of your cards? Do you use 17.4.3 and Claymore 9.3?

[coinzoid]

Anyone else having their effective hashrate drop this low?     It was like clockwork for months.

Yes, mining is like gambling Sometimes you are lucky and find a lot of shares and your effective hash rate goes way up, but sometimes you have unlucky streak and your effective hash rate will go very low. In the end it should average to your actual hash rate or very close to it over time.

Here is my current chart.

Same here.
I doubt it's an unlucky streak, since I know several people that are experiencing the same thing since a couple of days.
The number of shares just dropped by about 20% a couple of days ago on ethermine.org pool.
What could be the reason?

I'm also on ethermine and my last 24 hours looks like yours. The higher your hash rate the more noticeable the drop. My effective hash rate went from just under 1GH/s to 942MH/s. If you look at the stats for the top 10 miners on ethermine, you'll see that almost all of their charts show similar drops.

Same has happened to me for last two weeks. Effective rate drops frequently. I've thought it was due to dual mining eth + dcr but, it seems like pool has some issues.

Btw, If your rig are crashing when dual mining, check the gpu that is giving erros and re-flash its bios or reduce intensity. I've had one unlucky gpu that freezes whole system.

[djazher]

Hey everybody !

After countless hours of research, and obviously reading a lot about AMD cards Ether mining crown, I decided to do full Nvidia for a first 6xGPUs dedicated rig. I figured out that even if the hashrate is a bit lower than with RX 4XX cards on ETH, it also offers more possibilities on other cryptos. We don't want to forget that ETH is announced to go POS since last year. Plus, on the power consumption side, no battle here, Nvidia holds the 1st place by far.

I read as much as I could of this very thread, and checked the results on whattomine.com, but I'm still wondering what hashrates Nvidia users actually get in the real world, out of the theoritical field.

I'm already mining ETH+LBC with 2 GTX970 [41MHs/64MHs], for it happened to be the most valuable combo so far, but I doubt these results are accurate for a proper upgrade to GTX1070.

I also would like to ask what GTX1070 brand/model you're using, and eventually why. I'm looking for a good average between performance and longevity. With proper maintenance, I plan to get these working for at least 2 years time [Reboot and dust-off every week + full cleaning and basic maintenance every other month + full check and fan greasing twice a year]

I decided to buy GTX1070, but I've got a very nice opportunity for an EVGA GTX1080 SC, and was wondering as well, if people had performance feedback about this very GPU. As far as I can read, it seemed like 1080 weren't doing very good, and looked like it was because of a lack of optimisation [on top of the DDR5x issue]. Do you know if this has been fixed, or it will just be uncompatible at all ?

Cheers !

[notbatman]

Hey everybody !

After countless hours of research, and obviously reading a lot about AMD cards Ether mining crown, I decided to do full Nvidia for a first 6xGPUs dedicated rig. I figured out that even if the hashrate is a bit lower than with RX 4XX cards on ETH, it also offers more possibilities on other cryptos. We don't want to forget that ETH is announced to go POS since last year. Plus, on the power consumption side, no battle here, Nvidia holds the 1st place by far.

I read as much as I could of this very thread, and checked the results on whattomine.com, but I'm still wondering what hashrates Nvidia users actually get in the real world, out of the theoritical field.

I'm already mining ETH+LBC with 2 GTX970 [41MHs/64MHs], for it happened to be the most valuable combo so far, but I doubt these results are accurate for a proper upgrade to GTX1070.

I also would like to ask what GTX1070 brand/model you're using, and eventually why. I'm looking for a good average between performance and longevity. With proper maintenance, I plan to get these working for at least 2 years time [Reboot and dust-off every week + full cleaning and basic maintenance every other month + full check and fan greasing twice a year]

I decided to buy GTX1070, but I've got a very nice opportunity for an EVGA GTX1080 SC, and was wondering as well, if people had performance feedback about this very GPU. As far as I can read, it seemed like 1080 weren't doing very good, and looked like it was because of a lack of optimisation [on top of the DDR5x issue]. Do you know if this has been fixed, or it will just be uncompatible at all ?

Cheers !

Try some high end R9's like the nano on a low epoch ETH alt like EXP.

[Teress]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

I have also this OpenCL hang aprox. 2 times a day on one card. I know which one it is but have absolutelly no idea how to avoid this hang. I tried to set stock voltages, low memory frequency, low core, nothing. It is RX 480 Nitro+ 4GB with Hynix memory. Currently running on 1145 core, 1885 mem, 27 Mhs, zero memory errors. But 2 hangs a day How to solve this hang error?

sell it and buy another one, time is money

Bulgar thats not an advice. I just wanna know what is causing this OpenCL hangs. If it is something recoverable, than ok, if it's not, than sell it is the way. But what about when the new card will do the same? Again, I need to know the reason for theese hangs.

[jackbox]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

I have also this OpenCL hang aprox. 2 times a day on one card. I know which one it is but have absolutelly no idea how to avoid this hang. I tried to set stock voltages, low memory frequency, low core, nothing. It is RX 480 Nitro+ 4GB with Hynix memory. Currently running on 1145 core, 1885 mem, 27 Mhs, zero memory errors. But 2 hangs a day How to solve this hang error?

sell it and buy another one, time is money

Bulgar thats not an advice. I just wanna know what is causing this OpenCL hangs. If it is something recoverable, than ok, if it's not, than sell it is the way. But what about when the new card will do the same? Again, I need to know the reason for theese hangs.

What is your virtual memory setting? Make it higher. It is usually a system memory issue.

[AlainC]

@Claymore

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

On Linux I have this kind of message (triggering the watchdog, and reboot) when there is amdgpu driver kernel error, typically some GPU segfault fault. For me I blame it on RX480 overclocked/undervolted/overheating... another symptom is that ethdcrminer64 thread is hard sto stop, and even reboot is hard (I found a flag in grub to reboot reliably).
I imagine you are on Windows, and I don't know if you can catch the same hardware message ? can you check in windows event manager?
Note that for me this message can happen after 6-12 hours of stable work, but only near the hardware limits of each card (which are very diffrent for my RX480 - one stable at 29.6, other not so at 27.7).

On claymore 9.2 with a defunct RX470 I never had GPU hangs, but "incorrect shares", caused by hardware instabilities from overclocking and overheating.
NB: I have 20GB or swap and 4GB of DRAM.

[an6227]

Please claymore make the GPU list sorted as shown in GPU Z which is physically layouted

[AlainC]

Please claymore make the GPU list sorted as shown in GPU Z which is physically layouted

I don't know GPUZ (on linux), but it should be coherent with atiflash and amd numbering.
for Nvidia+ATI mixes, I'm afraid it will be complex.
Whatever numbering is done, something should allow a script to know which atiflash "slot" (and similar tool for Nvidia) is associated with a GPUnumber in ethdcrminer. Why not an initial table printer at boot, or a command-line option to dump that numbering. Today I am happy.

NB: I'm probably not the only guy to have automated some maintenance... eg I flash a more stable version of the VBIOS when I detect incorrect share or GPU hanging on one given GPU.

[Teress]

@Claymore

I have a problem with one of my RX480 cards. In the log I see before every crash the same message:

Code:

WATCHDOG: GPU 1 hangs in OpenCL call

Now I want to find out which of my GPUs this is! When I start claymore with the parameter "DI -1", can I be sure that it is directly pointing to GPU 1 or does the ordering changes?

What I do personally, I go on MSI Afterburner and I put the fan to the maximum. Now you just have to check the louder one.

I have also this OpenCL hang aprox. 2 times a day on one card. I know which one it is but have absolutelly no idea how to avoid this hang. I tried to set stock voltages, low memory frequency, low core, nothing. It is RX 480 Nitro+ 4GB with Hynix memory. Currently running on 1145 core, 1885 mem, 27 Mhs, zero memory errors. But 2 hangs a day How to solve this hang error?

sell it and buy another one, time is money

Bulgar thats not an advice. I just wanna know what is causing this OpenCL hangs. If it is something recoverable, than ok, if it's not, than sell it is the way. But what about when the new card will do the same? Again, I need to know the reason for theese hangs.

What is your virtual memory setting? Make it higher. It is usually a system memory issue.

16GB physical RAM, the same alias 16384 virtual