Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 1113.

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Quote from: peoples on September 22, 2016, 01:53:30 PM

still no fancontrol with rx cards under linux ?
Please all linux users with rx cards waiting for this .. sure more important than colors in linux Wink

Be patient please. Now I'm working on updating Monero miner as promised (check its thread for details), it will take a couple of days. Then I will work on this task, as promised.

adamvp

hero member

Activity: 1246

Merit: 708

Quote from: scryptr on September 22, 2016, 01:56:37 PM

Quote from: adaseb on September 22, 2016, 06:54:08 AM

Quote from: scryptr on September 22, 2016, 01:35:16 AM

A GOOD RULE-OF-THUMB--

Place no more than 2 GPU's powered risers on a single SATA cable. Placing the powered riser for GPU 0 on a single SATA cable, along with a HDD or other low powered device, may also be good policy. --scryptr

A BETTER RULE-OF-THUMB--

Don't power any 480s from any SATA cable.

SORRY, I DON'T UNDERSTAND--

The USB powered risers require power input from a SATA cable. Are you suggesting the use of non-powered risers? The results can be a fire hazard. I was expecting that the proper PCIe connections be made with PCIe cables. Using 2x Molex-to-1 PCIe adapters on anything bigger than a 750ti is also a fire hazard. Using a Mollex-to-PCIe adapter on a 280X will result in melted cables and a resulting short circuit. --scryptr

I mean he means Molex cable instead of SATA cable should be used for "bigger" card, as I mentioned above

scryptr

legendary

Activity: 1797

Merit: 1028

Quote from: adaseb on September 22, 2016, 06:54:08 AM

Quote from: scryptr on September 22, 2016, 01:35:16 AM

A GOOD RULE-OF-THUMB--

Place no more than 2 GPU's powered risers on a single SATA cable. Placing the powered riser for GPU 0 on a single SATA cable, along with a HDD or other low powered device, may also be good policy. --scryptr

A BETTER RULE-OF-THUMB--

Don't power any 480s from any SATA cable.

SORRY, I DON'T UNDERSTAND--

The USB powered risers require power input from a SATA cable. Are you suggesting the use of non-powered risers? The results can be a fire hazard. I was expecting that the proper PCIe connections be made with PCIe cables. Using 2x Molex-to-1 PCIe adapters on anything bigger than a 750ti is also a fire hazard. Using a Mollex-to-PCIe adapter on a 280X will result in melted cables and a resulting short circuit. --scryptr

peoples

member

Activity: 82

Merit: 11

Quote from: Claymore on September 22, 2016, 10:04:53 AM

v7.1:

- now "-etha 2" is set automatically for Linux and 4xx cards. You can change it by specifying "-etha" directly if necessary.
- if miner manages fans, it returns management back to drivers at closing.
- added text coloring in Linux version. You can use "-colors 0" option to disable it if necessary.
- several minor improvements and bug fixes.
- EthMan: added ability to start .bat file if miner has problems.
- EthMan: added support for CryptoNote miner.

still no fancontrol with rx cards under linux ?
Please all linux users with rx cards waiting for this .. sure more important than colors in linux Wink

adamvp

hero member

Activity: 1246

Merit: 708

Quote from: scryptr on September 22, 2016, 01:35:16 AM

A GOOD RULE-OF-THUMB--

Place no more than 2 GPU's powered risers on a single SATA cable. Placing the powered riser for GPU 0 on a single SATA cable, along with a HDD or other low powered device, may also be good policy. --scryptr

good tip. I am trying to use Molex first, especially with old high power consumption cards,
sata only with <150W cards

Small error in Linux version .tar:

Code:

Command 'tar -xz --directory="." -f "Claymore's Dual Ethereum+Decred_Siacoin_Lbry AMD+NVIDIA GPU Miner v7.1 - LINUX - Catalyst 15.12 - CUDA 8.0.tar.gz"'
failed with return code 0 and error message
tar: Claymore's Dual Ethereum+Decred_Siacoin_Lbry AMD+NVIDIA GPU Miner v7.1 - LINUX/History.txt: time stamp 2016-09-23 00:28:48 is 11492.336828217 s in the future
tar: Claymore's Dual Ethereum+Decred_Siacoin_Lbry AMD+NVIDIA GPU Miner v7.1 - LINUX/Readme!!!.txt: time stamp 2016-09-23 00:30:02 is 11566.336600769 s in the future.

And last.. (but it is not necessary miner fault) now I can't close miner by hitting ctrl+c in mobaXterm

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Quote from: fr4nkthetank on September 22, 2016, 10:09:01 AM

Any real difference between etha 1 and etha 2 for rx470 ? using etha 1 and not seeing much incorrect shares. 0 out of 1900 so far.

The difference is very, very small. I get about 0.1% incorrect shares on -etha 0 and 1. On -etha 2 it must be even less, but -etha 2 is a bit slower.
But for Linux and gpu-pro drivers if you don't set -etha 2, you will see high CPU load because of a lot of system logging about GPU segfaults.

fr4nkthetank

legendary

Activity: 2294

Merit: 1182

Now the money is free, and so the people will be

Any real difference between etha 1 and etha 2 for rx470 ? using etha 1 and not seeing much incorrect shares. 0 out of 1900 so far.

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

v7.1:

- now "-etha 2" is set automatically for Linux and 4xx cards. You can change it by specifying "-etha" directly if necessary.
- if miner manages fans, it returns management back to drivers at closing.
- added text coloring in Linux version. You can use "-colors 0" option to disable it if necessary.
- several minor improvements and bug fixes.
- EthMan: added ability to start .bat file if miner has problems.
- EthMan: added support for CryptoNote miner.

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Quote from: Prelude on September 22, 2016, 09:23:26 AM

Hi Claymore,

I have an issue with DAG generation with some Hawaii cards that are extremely undervolted. The cards will mine perfectly without ever crashing, but at that low voltage they can't generate the initial DAG or the following DAG when it comes time to switch. They'll just crash while other cards in the system keep mining.

This can be solved by raising the voltage, obviously, but that means an extra 30-40w of power consumed on most of these cards which I'd like to avoid. Seems to me that there would possibly be two ways to fix this, while keeping the undervolts where they are.

1) Reduce the intensity of DAG generation? (Not sure this is possible...)

2) Add an option to the config file to lower GPU's core clock while generating the DAG? IE When mining use regular stable mining clock speed of 1100MHz, next DAG generation comes along drop to 900MHz while generating, when completed increase back to 1100MHz. Flags would be -cclock 1100 and -cclockdag 900.

Any thoughts? Hope this makes sense!

Thanks!

Option #1 is better, PM me and in a day I will send you a version for tests.

CoRpO

full member

Activity: 185

Merit: 100

Quote from: Maksim Popov on September 22, 2016, 09:52:35 AM

https://blog.ethereum.org/2016/09/22/ethereum-network-currently-undergoing-dos-attack/

How to set this settings in Claymore?
–cache-size-db 1024 –gas-floor-target 1000000 –gasprice 500000000000

Applies only to pools and solo miners, not to pool miners

Maksim Popov

newbie

Activity: 24

Merit: 0

https://blog.ethereum.org/2016/09/22/ethereum-network-currently-undergoing-dos-attack/

How to set this settings in Claymore?
–cache-size-db 1024 –gas-floor-target 1000000 –gasprice 500000000000

Prelude

legendary

Activity: 1596

Merit: 1000

Hi Claymore,

I have an issue with DAG generation with some Hawaii cards that are extremely undervolted. The cards will mine perfectly without ever crashing, but at that low voltage they can't generate the initial DAG or the following DAG when it comes time to switch. They'll just crash while other cards in the system keep mining.

This can be solved by raising the voltage, obviously, but that means an extra 30-40w of power consumed on most of these cards which I'd like to avoid. Seems to me that there would possibly be two ways to fix this, while keeping the undervolts where they are.

1) Reduce the intensity of DAG generation? (Not sure this is possible...)

2) Add an option to the config file to lower GPU's core clock while generating the DAG? IE When mining use regular stable mining clock speed of 1100MHz, next DAG generation comes along drop to 900MHz while generating, when completed increase back to 1100MHz. Flags would be -cclock 1100 and -cclockdag 900.

Any thoughts? Hope this makes sense!

Thanks!

BTCBusinessConsult

sr. member

Activity: 546

Merit: 250

Active Trading on EPIC5k and Spectre.Ai

Any idea as to why I would be getting rejects all of the sudden? When it was Epoc 74 and 75, I never got a single reject with my 2x AMD 480s.

Now that Epoc 76 is here, my hashrate has dropped by 0.1 MHs per card ( expected) and I am now getting rejects. It is still a very very low percentage. This morning it was 5,000 accepted shares and 46 rejected.

adaseb

legendary

Activity: 3808

Merit: 1723

Quote from: scryptr on September 22, 2016, 01:35:16 AM

A GOOD RULE-OF-THUMB--

Place no more than 2 GPU's powered risers on a single SATA cable. Placing the powered riser for GPU 0 on a single SATA cable, along with a HDD or other low powered device, may also be good policy. --scryptr

A BETTER RULE-OF-THUMB--

Don't power any 480s from any SATA cable.

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Quote from: GandalfG on September 22, 2016, 06:00:26 AM

Quote from: Claymore on September 21, 2016, 12:43:23 PM

Quote from: tutorialevideo on September 21, 2016, 12:19:28 PM

on ubuntu i get this syslog 50GB

-rw-r----- 1 syslog adm 51086061568 sep 21 20:17 kern.log
drwx--x--x 2 root root 4096 sep 21 14:27 lightdm
-rw-r----- 1 syslog adm 51086168064 sep 21 20:17 syslog
-rw-rw-r-- 1 root utmp 2304 sep 21 14:43 wtmp
-rw-r--r-- 1 root root 369812 sep 21 14:45 Xorg.0.log
user@user-To-be-filled-by-O-E-M:/var/log$ tail -f kern.log
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730901] amdgpu 0000:04:00.0: VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0E044001
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730902] VM fault (0x01, vmid 7) at page 135284862, read from 'TC5' (0x54433500) (68)
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730915] amdgpu 0000:04:00.0: GPU fault detected: 147 0x03d0c401

"-etha 2" will help you. Within 24 hours I will release an update, this option will be set automatically on Linux+gpupro.

-etha 2 reduce hash rate 0,2Mh on each RX480 card - 1Mh in 5 card rig.

That's the price we have to pay for Linux+gpupro drivers. Better use Windows.

GandalfG

sr. member

Activity: 259

Merit: 250

Dig your freedom

Quote from: Claymore on September 21, 2016, 12:43:23 PM

Quote from: tutorialevideo on September 21, 2016, 12:19:28 PM

on ubuntu i get this syslog 50GB

-rw-r----- 1 syslog adm 51086061568 sep 21 20:17 kern.log
drwx--x--x 2 root root 4096 sep 21 14:27 lightdm
-rw-r----- 1 syslog adm 51086168064 sep 21 20:17 syslog
-rw-rw-r-- 1 root utmp 2304 sep 21 14:43 wtmp
-rw-r--r-- 1 root root 369812 sep 21 14:45 Xorg.0.log
user@user-To-be-filled-by-O-E-M:/var/log$ tail -f kern.log
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730901] amdgpu 0000:04:00.0: VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0E044001
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730902] VM fault (0x01, vmid 7) at page 135284862, read from 'TC5' (0x54433500) (68)
Sep 21 15:35:12 user-To-be-filled-by-O-E-M kernel: [ 4060.730915] amdgpu 0000:04:00.0: GPU fault detected: 147 0x03d0c401

"-etha 2" will help you. Within 24 hours I will release an update, this option will be set automatically on Linux+gpupro.

-etha 2 reduce hash rate 0,2Mh on each RX480 card - 1Mh in 5 card rig.

scryptr

legendary

Activity: 1797

Merit: 1028

A GOOD RULE-OF-THUMB--

Place no more than 2 GPU's powered risers on a single SATA cable. Placing the powered riser for GPU 0 on a single SATA cable, along with a HDD or other low powered device, may also be good policy. --scryptr

elmanager

sr. member

Activity: 306

Merit: 250

Cryptobulgarian ninja

Quote from: armyman on September 21, 2016, 04:34:41 AM

Guys I am trying to set up the following rig:

Asrock H81 pro btc
16 gb ram DDR3
6 x XFX RX480 8gb
6 x PCI 1x 16
SSD external memory
Intel Pentium G3260
1200 w PSU
WIN 10 64 bit

I would like to install the latest version of Claymore's dual miner on it, the problem is that the PC can not even run the 6 GPU's all the same time apparently, without anything installed on it yet, but Windows (I am helping which is setting it up and we are about 2000 miles far from each other, so this doesn't help).

Windows starts normally if only 3 GPUs are connected, but a black screen will come off if we try to connect all the 6 GPU's.

Any idea?

Also which drivers would you recommend for the XFX RX480 8gb? Is the latest crimson driver ok?

Any help would be deeply appreciated, thank you!

Try like separate the main GPU razer with only one SATA, HDD + 1 gpu - sata, 2x gpu-sata, 2x gpu-sata

xeridea

sr. member

Activity: 449

Merit: 251

Quote from: wll1rah on September 21, 2016, 08:02:51 PM

Quote from: xeridea on September 21, 2016, 10:41:37 AM

Quote from: citronick on September 21, 2016, 04:41:00 AM

Quote from: armyman on September 21, 2016, 04:34:41 AM

Guys I am trying to set up the following rig:

Asrock H81 pro btc
16 gb ram DDR3
6 x XFX RX480 8gb
6 x PCI 1x 16
SSD external memory
Intel Pentium G3260
1200 w PSU
WIN 10 64 bit

I would like to install the latest version of Claymore's dual miner on it, the problem is that the PC can not even run the 6 GPU's all the same time apparently, without anything installed on it yet, but Windows (I am helping which is setting it up and we are about 2000 miles far from each other, so this doesn't help).

Windows starts normally if only 3 GPUs are connected, but a black screen will come off if we try to connect all the 6 GPU's.

Any idea?

Also which drivers would you recommend for the XFX RX480 8gb? Is the latest crimson driver ok?

Any help would be deeply appreciated, thank you!

Did you power the molex connector on the ASrock mobo?

For multi GPU setup, you need this for extra power going to the GPUs in addition to powered USB risers.

Plus... you also need to set minimum 16GB virtual windows memory.

I don't know why lots of people think you need to power the molex on mobo. If using the 1x-16x powered USB risers, they are completely unnecessary. The risers are all powered, there is 0.00000w going through the slots on the motherboard. The molex on mobo are for if you are using non powered risers. I have that board, and it will run fine with 6 cards, nothing plugged into mobo.

Bios throws error telling you to plug in the 4-pin molex power on the motherboard.

You can tell it to not show that message. I am currently mining with that board with 5 cards, all powered risers, no molex plugged into the board.

dolores12

sr. member

Activity: 420

Merit: 250

Eth dcr mining is stable now with good temp Wink

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 1113. (Read 6590718 times)