Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux)

[Gagana]

I'm confused. It think I found a bug. And I also want to suggest a big improvement, Mr.Claymore.
 
My rig is a mixed 290/390/480

16:27:54:976   12e4   AMD Cards available: 6
16:27:54:976   12e4   GPU #0: Hawaii, 4096 MB available, 40 compute units
16:27:54:992   12e4   GPU #0 recognized as Radeon 290
16:27:54:992   12e4   GPU #1: Ellesmere, 8192 MB available, 36 compute units
16:27:54:992   12e4   GPU #1 recognized as Radeon RX 480
16:27:54:992   12e4   GPU #2: Ellesmere, 8192 MB available, 36 compute units
16:27:55:008   12e4   GPU #2 recognized as Radeon RX 480
16:27:55:008   12e4   GPU #3: Ellesmere, 8192 MB available, 36 compute units
16:27:55:008   12e4   GPU #3 recognized as Radeon RX 480
16:27:55:023   12e4   GPU #4: Hawaii, 8192 MB available, 40 compute units
16:27:55:023   12e4   GPU #4 recognized as Radeon 390
16:27:55:023   12e4   GPU #5: Hawaii, 8192 MB available, 40 compute units
16:27:55:023   12e4   GPU #5 recognized as Radeon 390

I use -di detect option and apply the output to the config.txt

-di 540231

and it works:
16:23:16:876   5f0   args: -di 540231 -gmap 540231 -tt 75 -ttli 82 -mport 3333 -retrydelay 5 -ftime 10
...
16:23:18:516   5f0   driver 10, 0, 2117, 13
16:23:18:516   5f0   AMD Cards available: 6
16:23:18:516   5f0   GPU #0: Hawaii, 8192 MB available, 40 compute units
16:23:18:532   5f0   GPU #0 recognized as Radeon 390
16:23:18:532   5f0   GPU #1: Hawaii, 8192 MB available, 40 compute units
16:23:18:532   5f0   GPU #1 recognized as Radeon 390
16:23:18:548   5f0   GPU #2: Hawaii, 4096 MB available, 40 compute units
16:23:18:548   5f0   GPU #2 recognized as Radeon 290
16:23:18:548   5f0   GPU #3: Ellesmere, 8192 MB available, 36 compute units
16:23:18:563   5f0   GPU #3 recognized as Radeon RX 480
16:23:18:563   5f0   GPU #4: Ellesmere, 8192 MB available, 36 compute units
16:23:18:563   5f0   GPU #4 recognized as Radeon RX 480
16:23:18:579   5f0   GPU #5: Ellesmere, 8192 MB available, 36 compute units
16:23:18:579   5f0   GPU #5 recognized as Radeon RX 480

Then I see this in the console log:
16:24:51:915   aa0   GPU #5: Ellesmere
16:24:51:915   aa0   ZEC - Total Speed: 1824.257 H/s, Total Shares: 79, Rejected: 0, Time: 00:01
16:24:51:930   aa0   ZEC: GPU0 395.950 H/s, GPU1 388.046 H/s, GPU2 323.483 H/s, GPU3 158.419 H/s, GPU4 281.087 H/s, GPU5 277.272 H/s
16:24:51:930   aa0   Pool switches: ZEC - 0
16:24:51:946   aa0   Current ZEC pool share target: 0x0020c49b (diff: 2000H)
16:24:51:946   aa0   Currently -ttli reduces intensity for: GPU3 (-80%)
16:24:52:008   aa0   GPU0 t=75C fan=91%, GPU1 t=80C fan=99%, GPU2 t=79C fan=100%, GPU3 t=86C fan=100%, GPU4 t=69C fan=47%, GPU5 t=79C fan=66%

I say: OMG, I turn off GPU3 until I understand what's happening. And press "3" button. What I see some time later?
16:25:25:183   aa0   ZEC - Total Speed: 1656.333 H/s, Total Shares: 104, Rejected: 0, Time: 00:01
16:25:25:199   aa0   ZEC: GPU0 390.400 H/s, GPU1 387.266 H/s, GPU2 322.875 H/s, GPU3 off, GPU4 280.499 H/s, GPU5 275.293 H/s
16:25:25:199   aa0   Pool switches: ZEC - 0
16:25:25:199   aa0   Current ZEC pool share target: 0x0020c49b (diff: 2000H)
16:25:25:214   aa0   Currently -ttli reduces intensity for: GPU3 (-80%)
16:25:25:277   aa0   GPU0 t=72C fan=91%, GPU1 t=79C fan=99%, GPU2 t=79C fan=100%, GPU3 t=87C fan=100%, GPU4 t=63C fan=22%, GPU5 t=76C fan=96%

Mmm? It looks like GPU3 is still working, but GPU4 is turned off instead. Isn't it a bit confusing, Mr.Claymore? OK, I say, may be it is related to -gmap option? Comment it by # And... Yes! After deactivating -gmap I see that the key pressed matches the GPU. If I press "3", then GPU3 stops mining, as I see on the temperature values. So, here is a bug, I suppose.

There are 3 lists of GPUs.
The first list is the GPU list to use in the config. It doesn't correspond with any other software. There is no way to reorder GPUs here. So, there is no way to make it corresponding to Atiflash, Whattool, MSI Afterburner, GPU-Z, HWinfo and other tools.
The second list is about the performance in the console log. By default it is the same as the first one. But you may reorder the GPUs here by using "-di" option.
The third list is about the temperature and it does correspond with Atiflash, Whattool, MSI Afterburner, GPU-Z, HWinfo and other tools. But if you want, you may reorder it with "-gmap" option.

They are completely independent!

I don't think that using OpenCL enumeration as the main one (with no alternative) is a good idea because it falls into conflict with most all of the other software. It confuses users. But OK, it's your choice, Mr.Claymore and the community has accepted it.  

But let's explain it carefully in the docs. Let's say clearly, that there are two independent lists of GPU. That -di affects the performance list only, -gmap affects the temperature list only. And "0..9" keys in runtime matches ... which list?! And if I use "-di 210", then "0..9" keys in runtime matches ... which list?

And if I use

-tt 80,82,85 for GPU0, GPU1, GPU2 respectively, then these 0,1,2  means the ORIGINAL enumeration (from OpenCL), but not what I see in the console if I use also "-di 210"!

The things are going less complicated with "-altnum" option introduced in the latest Dual miner. With "2" value we will see exactly GPU numbers, that is great! But it's possible to solve most all the issues by implementing something like

-alo - adapter list order, that is applied to the OpenCL output first of all. All other indexes are applied to the list, already reordered by "alo" option.

Really! All I need is to order the adapters either the same way as the other software does or as these adapters physically placed. Both options are achievable with this new option.

Dear Mr.Claymore, please, stop all our troubles with those ... ATI programmers of OpenCL forever!

[Claymore]

I'm confused.

Why did you set both "-di" and "-gmap" to the same value? This way you changed order in list of GPUs, then did the same for the list of temperatures. So both list will match again. But original (default) match is not good for your system. That's the answer.
Use "-di" only, remove "-gmap".

[Lexiko]

Hi! I have two GPUs on my board: nvidia 1080ti and amd r9 370. OS - linux debian x64.
Trying to mine zec-s. NVIDIA card is ok, but I can't force to work my AMD card.
Using zecminer64:
Zecminer64 -zpool coinotron.com:3346 -zwal * -zpsw * -allpools 1
First 10-20 min's its ok, but when it's starts to mine zec.suprnova.cc, the work finish.
The log is:
DevFee: ZEC: Stratum - connecting to 'zec.suprnova.cc' <46.105.114.185> port 2242
SSL / TLS encryption is enabled
DevFee: Pool sets new share target: 0x003c3c3c (diff: 1088H)
DevFee: start mining (90sec)
GPU0 t = 61C fan = 4%
WATCHDOG: GPU 0 hangs in OpenCL call, exit
WATCHDOG: GPU 0 hangs in OpenCL call, exit
WATCHDOG: GPU 0 hangs in OpenCL call, exit
WATCHDOG: GPU 0 hangs in OpenCL call, exit
Can not even kill the task via killall -9 ... it stuck
Trying to start clinfo, it's stucked too! Only reboot can solve this problem. (But the Nvidia card at that moment is continue to mine without no problem).
There are no overheating, the temp is ok all the time before stuck
GPU0 t = 65C fan = 23%
ZEC: 04/29 / 17-21: 02: 42 - New job from coinotron.com:3346
ZEC - Total Speed: 0.000 H / s, Total Shares: 1, Rejected: 0, Time: 00:13
ZEC: GPU0 0.000 H / s

How can I solve it?

[hawer183]

Do anyone have same problem like my Radeon 390x - it is beginning mine with speed of 400H/s and after 1 hour it is just 250H/s...

Catalist 15.12

My Radeon 390x get --» 501 H/s in Overclock
My Radeon 480 Nitro+ get --» 375 H/s in Overclock

Windows 10 64 bits and Catalist 16.12.2

Can you share the OC specs for the 480 that give you 375H/s?
Is it just one card or a full x6 Rig?

Thanks

Bullshit!!!

[Elder III]

What Do you think about replacing a r9 280 with a Gtx 1060? Same Hash, lower power..

It might be better to use the 1080 Ti.

Lol yeah, bit pricey card.. 0.07 usd/kW is my price so profit is good

You have a nice low electric cost, so I would probably stick with the R9 280 myself. Is it possible for you to add a second GPU and mine with both rather then replacing the first one? With $0.07/kwh I would work towards a nice 6 GPU rig if you can.

[FFI2013]

Could someone help me out with what settings you are running I have a Asus r9 390 only getting 380hs I also have a sapphire reference 290 only getting 315hs from what I'm seeing my rates are low my clocks are 1075 1500 and 1075 1450 how could I get better rates and less power consumption just looking for some pointers I no every card is different

[sikkan]

What Do you think about replacing a r9 280 with a Gtx 1060? Same Hash, lower power..

It might be better to use the 1080 Ti.

Lol yeah, bit pricey card.. 0.07 usd/kW is my price so profit is good

You have a nice low electric cost, so I would probably stick with the R9 280 myself. Is it possible for you to add a second GPU and mine with both rather then replacing the first one? With $0.07/kwh I would work towards a nice 6 GPU rig if you can.

As a matter of fact, my r9 280 is the least energy efficent card in my system as it turns out . It will be replaced with a third 7990 as soon as my risers comes! Turns out each 7990 only pulls 260W per card. That would mean 780W for 3 7990, 6 cores and 1680 sols/s!

[igorsantos]

Is there any forecast for a new version?

[golden0707]

Hi friends!

For those who has 290s. Im getting 360 h/s. But the card runs at 80 degree celsius. 76% fan speed. Is it normal for that card? I have several 280x. And they run at less then 70°.

[marcos5872]

Do anyone have same problem like my Radeon 390x - it is beginning mine with speed of 400H/s and after 1 hour it is just 250H/s...

Catalist 15.12

I had the same problem, using Dummy Plug and Team Viewer the problem was solved.

[kiaas]

out of curiosity, does it work on the FOSS Radeon drivers? At all?
I know some of the PPAs on Ubuntu at least enable the OpenCL support from the FOSS drivers, but before ZEC they weren't usable because they only wanted to allow allocating 1GB of VRAM for OpenCL, and Ethereum was already up to ~1.3GB DAG files. But how about with ZEC, does it work at all and what sort of performance penalties are there from it?
edit: asking because I have A Lot of SI hardware, and it working even at somewhat reduced performance with the Radeon Drivers will make my migration path a bit easier.

[golden0707]

Hi friends!

For those who has 290s. Im getting 360 h/s. But the card runs at 80 degree celsius. 76% fan speed. Is it normal for that card? I have several 280x. And they run at less then 70°.

No one knows ?

[FlensGold]

Hello,
I have a problem with my rig(s), maybe someone has an idea. I am using Awesome Miner controlling Claymore Miner. This problem is not awesome miner related since I do get the same error while using claymore without AM.
I do have some Rigs running Windows 10 and I do access them via Teamviewer. Sometimes the rigs (randomly from my point of view) stop hashing and when I connect to it it looks like this:


Since this happens quite often I had the chance to see what happens before: For some reason Windows 10 shows an error claiming that Windows blocks access to (I think it was) Radeon settings which causes Claymore Miner to stop (or maybe stop GPU working so the watchdog restarts Claymore).
I did try different driver versions as of right now I am using the recommended 16.2.2.

In those cases I have to reboot the rig, but unfortunately sometimes not all GPUs start hashing after start of CM. CM states that the GPU hangs in OpenCL call and watchdog restarts CM which can also lead to the screen above.
I do have this problem on all my rigs, so I would appreciate any idea that might help.
Thank you!

[galaxy5user]

How to use an external watchdog? I mean, for ex., sgminer has an API which allows to query the status of the miner and reboot the PC if necessary. How to do the same with Claymore's (check it's status)?

[haxiboy]

Claymore have the same jsonrpc protocol for remote monitoring/controlling, default port is 3333.

You can check it with EthMan, or you can write your own as jsonrpc is well documented. (as i did).
After my miner monitor will be stable i'll release it with a minersit hardware based that can monitor/restart your pc like simple rig resetter but with more features.

[MinermanNC]

Hi friends!

For those who has 290s. Im getting 360 h/s. But the card runs at 80 degree celsius. 76% fan speed. Is it normal for that card? I have several 280x. And they run at less then 70°.

No one knows ?

I'm getting 310 h/s on 3 x 290x and 383 h/s on a 390

Temps I have capped not to exceed 83. I am fighting the heat of the south now in my garage. But have had higher hashes during the winter. I don't want to burn em up lol

So I would say your h/s vs. temps is very good and about right  

and yes on the 280x's I run 11 of those and the 4 x 280x rigs average 1.1 kh/s and run cool around 72. Great cards for ZEC 

[Ahead]

Hi,

I'm having an issue with my Gigabyte R9 280x (windforce OC).
The card hashrate is at 300+-, (core clock at 1100) but after a while, for no reason the hashrate drops at 50% (and I have noticed that the core clock drops to 501).
I have to reboot the system or disable/enable the card in device manager to hash at 300 again (for a while, then it drops again at 150 again)

Anyone else having the same issue or any idea what is causing this?  

Thank you in advance!

[MinermanNC]

Hi,

I'm having an issue with my Gigabyte R9 280x (windforce OC).
The card hashrate is at 300+-, (core clock at 1100) but after a while, for no reason the hashrate drops at 50% (and I have noticed that the core clock drops to 501).
I have do reboot the system or disable/enable the card in device manager to hash at 300 again (for a while, after it drops at 150 again)

Anyone else having the same issue or any idea what is causing this?  

Thank you in advance!

are you running just the one card?

What fan temp control software are you running?

Sounds like it getting hot and throttling down the clock

[MinermanNC]

Anyone have any results yet for the RX 580 - 570 cards?

I noticed they are keeping prices up on the 480 line 

Please post

THX

[Ahead]

Hi,

I'm having an issue with my Gigabyte R9 280x (windforce OC).
The card hashrate is at 300+-, (core clock at 1100) but after a while, for no reason the hashrate drops at 50% (and I have noticed that the core clock drops to 501).
I have do reboot the system or disable/enable the card in device manager to hash at 300 again (for a while, after it drops at 150 again)

Anyone else having the same issue or any idea what is causing this?  

Thank you in advance!

are you running just the one card?

What fan temp control software are you running?

Sounds like it getting hot and throttling down the clock

Hi,

Thank you for the reply!
yes just one card in the pc!
I use afterburner and fan speed is set at 100% (temp at 85c-92c)