Author

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 807. (Read 6590718 times)

newbie
Activity: 28
Merit: 0
Hi all,

I started mining with Claymore a few days ago and I am seeing an upsetting thing.
My GPU is an EVGA GTX 1080 FTW, watercooled, MSI Afterburner reports a RAM frequency of 5000 MHz, which is fine... but when mining with Claymore, the RAM frequency automatically drops to 4500 MHz. If I stop mining and start a game, the frequency goes back to 5000 - the 4500 MHz behavior only appears when mining using Claymore.

I suspect the driver somehow knows I am using Claymore and taxes me for that.
If I boost RAM frequency by +1000, it goes up from 4500 to 5500 MHz and the hash speed goes from 20.5 MH to over 25 MH - 25% boost - but this solution is clumsy, because whenever I stop mining I have to dial back the frequency boost, otherwise as soon as the GPU encounters any regular load, it tries to boost from 5000 to 6000 MHz and the memory can't handle it.

What should I do to keep mining with Claymore, but at correct RAM frequencies?

Hi, I don't have GTX1080 card, but I am currently mine on GTX970 which have similar issue. Try to use the NVSMI tools to raise the card power state from P2 to P0.
NVIDIA card stay in P2 power state when CUDA computing, memory speed will be limited in P2. Raise to P0 to unleash the full clock speed. Hope it will solve your problem.

ref: http://cryptomining-blog.com/tag/nvidia-force-power-state/

This doesn't work on 10 series cards right now, there is a nvidia driver bug.
newbie
Activity: 2
Merit: 0

Any ideas on whats causing this and how to fix it?


Are you OCing them? Maybe too much OC on memory?

No, I havent overclocked them just yet. Still having the same problem
newbie
Activity: 3
Merit: 0
Simple stupid question (because my brain went to bed hours ago and I can't find the answer)

Does it work on a 2 gig rx460 anymore ? I can't write the dag file (31 and 60) which made me think 2 gig isn't enough ram on the card

This is with 16 gig of system ram and a 16 gig pagefile.
newbie
Activity: 2
Merit: 0
Hi all,

I started mining with Claymore a few days ago and I am seeing an upsetting thing.
My GPU is an EVGA GTX 1080 FTW, watercooled, MSI Afterburner reports a RAM frequency of 5000 MHz, which is fine... but when mining with Claymore, the RAM frequency automatically drops to 4500 MHz. If I stop mining and start a game, the frequency goes back to 5000 - the 4500 MHz behavior only appears when mining using Claymore.

I suspect the driver somehow knows I am using Claymore and taxes me for that.
If I boost RAM frequency by +1000, it goes up from 4500 to 5500 MHz and the hash speed goes from 20.5 MH to over 25 MH - 25% boost - but this solution is clumsy, because whenever I stop mining I have to dial back the frequency boost, otherwise as soon as the GPU encounters any regular load, it tries to boost from 5000 to 6000 MHz and the memory can't handle it.

What should I do to keep mining with Claymore, but at correct RAM frequencies?

Hi, I don't have GTX1080 card, but I am currently mine on GTX970 which have similar issue. Try to use the NVSMI tools to raise the card power state from P2 to P0.
NVIDIA card stay in P2 power state when CUDA computing, memory speed will be limited in P2. Raise to P0 to unleash the full clock speed. Hope it will solve your problem.

ref: http://cryptomining-blog.com/tag/nvidia-force-power-state/
sr. member
Activity: 438
Merit: 250
Claymore,
do you have any ideas how to overcome Polaris hashrate drop for epoch >129 ?
Genoil said that DAG now took 100% of first memory bank (2gb) and go to second memory bank, so access two banks instead of one leads significant hashrate drop.
Do you make any research in this direction, or it is hopeless? What you think, any comments?

I did not do any research yet, but I will do it very soon.

disclaimer: i just gave this as a possibly oversimplified answer to the question. like you explain to kids that thunder comes from clouds bumping into eachother  Cool
newbie
Activity: 22
Merit: 0
Claymore,
do you have any ideas how to overcome Polaris hashrate drop for epoch >129 ?

Genoil said that DAG now took 100% of first memory bank (2gb) and go to second memory bank, so access two banks instead of one leads significant hashrate drop.

Do you make any research in this direction, or it is hopeless? What you think, any comments?

I have this probleme with 8g rx 480 hash drop every new dag epoche on ethereum and ethereum classic

Minergate miner say with her laste version we do not loss effective Hashrate see here;

https://minergate.com/blog/v-6-5-no-hashrate-drop-after-dag-epoch-switch/
donator
Activity: 1610
Merit: 1325
Miners developer
Had no failing rigs - until today. Sadly I now can confirm that EthMan still did not detect that the rig was up and running again ("View miner console" worked).
I started to use your logging debug version now and will inform you as soon as I have the error "recorded".

Did you try to reproduce the lock with sending the restart command to a read-only mamangement interface like I stated before?

Yeah I got some logs and found the reason of the issue, I will post EthMan v3.3 within 1-2 days.
newbie
Activity: 50
Merit: 0
I think I have found a bug. I identified it at v9.3 and I tried 9.5 to check if it has been fixed, but bug stays there. The problem is; I have two RX570 graphics cards. One is MSI and one is Sapphire. In MSI Afterburner, GPU-Z and CPUID Hwmonitor all values are same, I can recognize which RX570 is MSI and which is the sapphire according to GPU-Z basic info-sensors screens. and this values are appropriate with Miner at line which writes GPU temperatures with purple color. But the line which writes hashrates with cyan color is not true. GPU numbers are invers there. I have tested a few times to be sure, and I am sure now. It writes Sapphire's hashrate in cyan as GPU0 but writing sapphire's temperature as GPU1 in purple. and for MSI same nice versa.
Hard to read so big text to understand that the problem is already discribed in faq, try to put problem in first part of text.

look readme.txt you need to use -di detect or -di 10 command to change videocards order.  to switch temperatures order use -gmap 10 command
newbie
Activity: 8
Merit: 0
I think I have found a bug. I identified it at v9.3 and I tried 9.5 to check if it has been fixed, but bug stays there. The problem is; I have two RX570 graphics cards. One is MSI and one is Sapphire. In MSI Afterburner, GPU-Z and CPUID Hwmonitor all values are same, I can recognize which RX570 is MSI and which is the sapphire according to GPU-Z basic info-sensors screens. and this values are appropriate with Miner at line which writes GPU temperatures with purple color. But the line which writes hashrates with cyan color is not true. GPU numbers are invers there. I have tested a few times to be sure, and I am sure now. It writes Sapphire's hashrate in cyan as GPU0 but writing sapphire's temperature as GPU1 in purple. and for MSI same nice versa.
full member
Activity: 186
Merit: 100
@Claymore
 I Have some gtx 1070 rigs. If I use one of the gtx 1070 primary vga for windows System becomes too slow. But if I open intel gpu and make it primary vga for windows everything ok.
Ethman monitor can't detect the restarted rig and it can't connect if I don't close and reopen the ethman
GH
member
Activity: 117
Merit: 10
Had no failing rigs - until today. Sadly I now can confirm that EthMan still did not detect that the rig was up and running again ("View miner console" worked).
I started to use your logging debug version now and will inform you as soon as I have the error "recorded".

Did you try to reproduce the lock with sending the restart command to a read-only mamangement interface like I stated before?
hero member
Activity: 520
Merit: 500
Got one asus r9 290
GPU temprerature increases up to 90 C and fan speed to 100%
Is that okay or i need something to do?

My bat:

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 80
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 80
setx GPU_SINGLE_ALLOC_PERCENT 80

EthDcrMiner64.exe -epool stratum+tcp://eth-eu.dwarfpool.com:8008 -ewal wallnumber -epsw x  -tt 84 -ttli 86 -r 30  -tstop 90 -ethi 6



You can try this.
At the same Mh you get less temp.
member
Activity: 129
Merit: 10
I have a gigabyte GA-Z170 Gaming K3 m/b running windows 10 64bit  with a powercolor  r 9 290 running Claymore 9.4 duel mining eth sia  well on it.
 i have ordered a EVGA NVIDIA GeForce GTX 1060 6 GB GDDR5 Memory PCI Express 3 card
other than the latest driver 382.53  will i need to install anything else to run it.
And would it be better to run the 2 cards on different instances.

Thanks in advance for any help
hero member
Activity: 676
Merit: 500
Just curious, has anyone tried both AMD and Nvidia in a single rig?
yep here you can see https://bitcointalksearch.org/topic/m.17966253

thanks seems to have trouble trying to boot up with 4 rx570 and 1 gtx1070 on my TB250 mobo
I don`t know about this board , but put cards one buy one and boot and install drivers , and if you in legacy mode try to setup windows in only uefi mode in the bios!  
newbie
Activity: 3
Merit: 0
Just curious, has anyone tried both AMD and Nvidia in a single rig?
yep here you can see https://bitcointalksearch.org/topic/m.17966253

thanks seems to have trouble trying to boot up with 4 rx570 and 1 gtx1070 on my TB250 mobo
newbie
Activity: 64
Merit: 0
I've been mining Eth only for about a month now and wanted to setup dual mining for ZEC, but i'm unsure of if I have to use the claymore-zec or can I dual mine it with just adding the appropriate settings into my claymore-eth folder?

I have been running a zcash miner along side of claymore now a bit and it does ok. Just watch the over clock speed for errors. That is on 1060 gtx cards with 6 gig ram. Works the card really hard and heats up so run fan high. Also you will see significant drop in hash power each. One has a sol > 1100 but drops back half when clay runs.
newbie
Activity: 72
Merit: 0
  A good word to the wise, do not buy a MSI  Z170 M5 if you wish to mine. I have tried win7 win10 and ubu1604 and none of them will install and run.
Not sure what the problem with it is but seems to be the old style of bios and it will not see win7 and win 10 may or may not boot. Then cards not show
up. Go Asus or something else your style.

thay


Did you try upgrading the bios ?
newbie
Activity: 64
Merit: 0
  A good word to the wise, do not buy a MSI  Z170 M5 if you wish to mine. I have tried win7 win10 and ubu1604 and none of them will install and run.
Not sure what the problem with it is but seems to be the old style of bios and it will not see win7 and win 10 may or may not boot. Then cards not show
up. Go Asus or something else your style.

thay
hero member
Activity: 676
Merit: 500
Just curious, has anyone tried both AMD and Nvidia in a single rig?
yep here you can see https://bitcointalksearch.org/topic/m.17966253
newbie
Activity: 72
Merit: 0
Anyone experience dev pool fee connection failing for past few hours ?
Jump to: