Bitcoin Forum

Hi People,

After testing a few of Claymore's releases, I have found out that some variables are better than other, with some giving an increase in profit while others offering some potential profit in the future, if these coins are kept "in the vault".
All tests have been documented and I want to share these results with you, in the following lines:

Claymore's Dual Ethereum+Decred AMD GPU Miner
Power Consumption


The following test shows the approximate power consumption (from wall) for each of 5 steps in the Claymore’s Dual Ethereum+Decred Miner using Catalyst 15.12.
This test only approximates the power usage of 1 rig and may not be a valid way of calculating your exact power consumption, due to component variables, Volt input (test was made on 230V line), overclock and component life.

In order to demonstrate how this mining test will perform, all clock/core speeds are set to stock, only temperatures and power consumption will be monitored during this test.
Temperatures are high because I usually under volt Core and under clock Memory for GPU1 (r9 290)
(-100mV/1000Mhz – Temp ~79*C/~69*C/65*C)
Also, Ambient temperatures are at around 17*C, but cards are not very well ventilated (working on it)

Miner: Claymore's Dual Ethereum+Decred AMD GPU Miner v4.1 - POOL Catalyst 15.12
GPU:
GPU2:r9 290
GPU1: r9 280x
GPU0: r9 280x

PSU: AX1200i – 80+ Platinum  - Power in/output reference

Mobo:ASROCK H77 pro4/MVP
CPU:  Intel Core i3 2100
Ram: 2x2GB Kingston Value Ram 1333Mhz
HDD: Toshiba MK2565GSXN  Laptop Drive 256gb 5200RPM

*.Bat Setup:
EthDcrMiner64.exe -epool eth-d00dfe289cbf79d8.dwarfpool.com:8008 -ewal 0xd32576bba69ab4b67dc15dd8b32a8b5adcdf4a0a -epsw Geth -dpool http://dcr.suprnova.cc:9111 -dwal cianura.reth -dpsw reth

Environment Variables:
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_SINGLE_ALLOC_PERCENT 100

Virtual Memory:
20000MB

Begining test:

TEST 1: MODE 1 –Decri 0

Power consumption:    
From wall: ~775W/h

Power Output:      
ETH - Total Speed: 70.023 Mh/s, Total Shares: 13, Rejected: 0, Time: 00:03
ETH: GPU0 19.646 Mh/s, GPU1 20.283 Mh/s, GPU2 30.094 Mh/s

GPU Temperature:
GPU2: r9 290   - 82*C
GPU1: r9 280x – 61*C
GPU0: r9 280x – 56*C

Profit:    
HashRate: 70.023Mh/s
Price / Difficulty ETH $9.73: 34.25 Trillion
Power Cost per Month (0.12USD/kWh): $66.96
Income per Month ETH (Minus Power cost): $186.13


TEST 2: MODE 0 –Decri 30

Power consumption:    
From wall: ~835W/h

Power Output:      
ETH - Total Speed: 69.781 Mh/s, Total Shares: 1284, Rejected: 0, Time: 11:01
ETH: GPU0 19.454 Mh/s, GPU1 20.350 Mh/s, GPU2 29.976 Mh/s
DCR - Total Speed: 1046.709 Mh/s, Total Shares: 2322, Rejected: 387
DCR: GPU0 291.811 Mh/s, GPU1 305.255 Mh/s, GPU2 449.643 Mh/s

GPU Temperature:
GPU2: r9 290   - 89*C
GPU1: r9 280x – 62*C
GPU0: r9 280x – 65*C

Profit:    
HashRate ETH/DCR: 69.781 Mh/s / 1046.709Mh/s
Price / Difficulty ETH $9.73 / DCR $1.49: 34.25 Tril. / 612306
Power Cost per Month (0.12USD/kWh): $72.14
Income per Month ETH/DCR (Minus Power cost): $180.07/ DCR: $28.161 = $208.231



TEST 3: MODE 0 –Decri 50

Power consumption:    
From wall: ~970W/h

Power Output:      
ETH - Total Speed: 67.255 Mh/s, Total Shares: 2, Rejected: 0, Time: 02:01
ETH: GPU0 19.684 Mh/s, GPU1 20.250 Mh/s, GPU2 27.322 Mh/s
DCR - Total Speed: 1681.378 Mh/s, Total Shares: 14, Rejected: 0
DCR: GPU0 492.098 Mh/s, GPU1 506.241 Mh/s, GPU2 683.039 Mh/s

GPU Temperature:
GPU2: r9 290   - 89*C
GPU1: r9 280x – 64*C
GPU0: r9 280x – 68*C

Profit:    
HashRate ETH/DCR: 67.255 Mh/s / 1681.378 Mh/s
Price / Difficulty ETH $9.73 / DCR $1.49: 34.25 Tril. / 612306
Power Cost per Month (0.12USD/kWh): $83.81
Income per Month ETH/DCR (Minus Power cost): $156.82/$45.594 = $202.394


TEST 4: MODE 0 –Decri 70

Power consumption:    
From wall: ~1000W/h

Power Output:      
ETH - Total Speed: 63.329 Mh/s, Total Shares: 5, Rejected: 0, Time: 01:03
ETH: GPU0 19.076 Mh/s, GPU1 19.481 Mh/s, GPU2 24.772 Mh/s
DCR - Total Speed: 2216.530 Mh/s, Total Shares: 28, Rejected: 2
DCR: GPU0 667.651 Mh/s, GPU1 681.844 Mh/s, GPU2 867.035 Mh/s

GPU Temperature:
GPU2: r9 290   - 90*C
GPU1: r9 280x – 67*C
GPU0: r9 280x – 71*C

Profit:    
HashRate ETH/DCR: 63.329 Mh/s / : 2216.530 Mh/s
Price / Difficulty ETH $9.73 / DCR $1.49: 34.25 Tril. / 612306
Power Cost per Month (0.12USD/kWh): $86.40
Income per Month ETH/DCR (Minus Power cost): $140.18/ DCR: $59.46 = $199.64


TEST 5: MODE 0 –Decri 100

Power consumption:    
From wall: ~1040W/h

Power Output:      
ETH - Total Speed: 53.884 Mh/s, Total Shares: 1, Rejected: 0, Time: 01:01
ETH: GPU0 16.724 Mh/s, GPU1 16.926 Mh/s, GPU2 20.235 Mh/s
DCR - Total Speed: 2694.201 Mh/s, Total Shares: 9, Rejected: 6
DCR: GPU0 836.177 Mh/s, GPU1 846.295 Mh/s, GPU2 1011.729 Mh/s

GPU Temperature:
GPU2: r9 290   - 91*C
GPU1: r9 280x – 68*C
GPU0: r9 280x – 72*C

Profit:    
HashRate ETH/DCR: 53.884 Mh/s / 2694.201 Mh/s
Price / Difficulty ETH $9.73 / DCR $1.49: 34.25 Tril. / 612306
Power Cost per Month (0.12USD/kWh): $89.86
Income per Month ETH/DCR (Minus Power cost): $102.93/ DCR: $73.308 = $176.238

Test Finished



Conclusion:
Please note that these tests have been conducted with a fee of 2% in mind. Shares have not been calculated due to the fact that they can be influenced by internet provider, proxy and other factors.

The command line –decri 100 does not provide my rig with any beneficial output over ETH when mining DCR, while draining 1040W/h from the wall, it only provides an additional ~470Mh/s for DCR, but at the cost of minus ~10Mh/s ETH.
Also, mining on regular ETH could not be that beneficial also, since profits drop by ~13-15%.

Calculations per Month have been made using ETH calculator:
http://karldiab.com/EthereumMiningCalculator/
And DCR calculator:
http://decred-calc.cryptohub.info/

If you plan on building a new rig, please take into consideration the following calculations for your GPU/PSU, using more than one reference for GPU power draw on full load (https://forum-en.msi.com/faq/article/printer/power-requirements-for-graphics-cards):
Eg. Radeon R9-290 - 31A and a 550W psu minimum
Meaning: ~Full load r9 290 may draw 31A

31A*12V=372W from the 12V rail.

1A for the 12V rail (PCI-e power) means 12W / 1A*12V=31A
2*r9 290 might draw up to ~744W
3*r9 290 might draw up to ~1116W
4*r9 290 might draw up to ~1488W
5*r9 290 might draw up to ~1860W
6*r9 290 might draw up to ~2232W
~ Depending on clock speed, core speed, voltage and temperature

For this example (31A on 6+8pin PCI-e connector) you will need a good power supply that can provide a minimum of 16A for each rail.

Other Graphic Cards can differ greatly, drawing more power from one 8pin connector
(Asus Strix 380), which draws 28A from one rail, with a minimum recommended PSU/1GPU of 500W
[/b]

Please use the following guide (or similar) as reference: http://www.tomshardware.com/reviews/power-supply-specifications-atx-reference,3061-12.html

With these in mind, please note that using Claymore’s dual Miner can affect your Amper (A) power draw and may lead to permanent PSU damage!!!


If you need and other information, please feel free to contact me at: Adrian.botoncea@gmail.com

Also, If you think that this guide helped, you can tip me with some ETHs at:
0xd32576bba69ab4b67dc15dd8b32a8b5adcdf4a0a
Or some DCRs at:
DsekkBhNHfYAzmde5yywg43e9Vdqb7MxRUC


Disclaimer:
Please note that these tests have been conducted with a fee of 2% in mind. Shares have not been calculated due to the fact that they can be influenced by internet provider, proxy and other factors.

Neither Claymore or I will not be held responsible for misleading information, component damage or other problems that may occur due to following this guide!
It is your duty to research more than one source and test your gear in a cautious manner!


Date: 5/3/16 - Adrian

That is good effort. What are the frequency of the graphics?

372 W for a R9 290 is quite high. What are the voltage an frequency?

your wattage numbers are way too high you need to post pics of a wattmeter or your PDU

that calcualator for Eth is not a good one either, the best one is the one on etherscan.io

Power consumption very depends on GPU temperature.
89C is too high for good efficiency.

For example I measured for Sapphire R9 290X (original AMD design):
(only eth minining, not Claymore, but qtminer on ethpool.org)
if 70C - aprox. 232W
if 80C - aprox. 242W
if 90C - aprox. 258W
(from the wall, EVGA Gold 1000W PSU 230V, only GPU - deducted consumption for MB etc.)

For Asus Strix R9 380/MSI Gaming R9 380 I have usually only 180-185W (only eth minining, 65C)
for MSI Gaming 3G 280X between 200-230W (only eth mining, 65-75C)

Good work, but You missed finding maximum profitability laying between -dcri 30 and 50 (my assumption 42-45).

Thank you for the detailed information.

I have been mining only Ethereum with Claymore's miner so far, I will check for a Decred pool it seems worth it!

Thanks, this is very interesting! Keep in mind that you can also tune the clock speed and on some cards the GPU voltage, which has a strong effect on the power consumption. I have made some tests with the hashing rate and power consumption of r9 390, and could bring my system from 28.9Mh/s with 312W to 30Mh/s with 285W

That is great work. I noticed that your power consumption is from the wall. So undervolting makes a big difference.

Nice Info,

I have 4x XFX R9 390's and I had to undervolt and underclock them to keep their temps below 72C with just ETH mining. I don't have decred enabled.

I have them each configured independently but in total they run about 102MH with Claymores miner.... not sure how much power they are drawing but at stock speeds/volts I was seeing temps well into the 80C+ range with an ambient temp of 23-24.5C (74-76F) and I was getting about 115ish MH.

For my money its better to undervolt and underclock to keep the temps down and save a ton of power/heat.

How do you do the undervoltage? Do you use MSI afterburner or the GPU mod programme such as OP mentioned?

I used MSI Afterburner to undervolt the maximum -100mv after applied the unlock voltage controls setting. You can't use the slider to go past zero but you can manually type -100 in the field and apply it.

I plan on eventually doing a Bios mod as I'm familiar with that type of mods as I have experience with them from my gaming systems.

Edit: I also used the same software to underclock the clock speed of the GPU's using the same -value approach (example -45). I also use GPU-Z for monitoring when testing new settings. I did not modify the memory speeds.

If your Core clock is less than 1100 MHz, the memory speed can be lower than 1250 MHz, 1200 MHz is fine.

That doesn't make any sense to me.  AMD GPUS  have memory  timings set at 125MHz intervals, always ending at a multiple of 125MHz.  You get the best random access speed at the high end of each range, like 1250 or 1500.

That is because the memory speed is fast enough or 1100MHz, if you use the 1150 MHz (Core), you have to use 1250 memory frequency, otherwise the memory controller will be 100%.

How did you undervolt those cards? I have two I would love to undervolt and cut some temps too.

The 290/390/290X/390X must be undervolted to be profitable, period.

When i was mining, i used the Stilt's modded bioses.  He had managed to disable bits and pieces of the chip that were not necessary for mining.  I remember my 290s being right near 200w power usage @ stock clocks / volts / 70c max temps.

I also recall that the bios repo was hard to get at, but perhaps enough google fu can resurrect a link for them.

You can use the MSI Afterburner to under volt the XFX by -100 mV. You can use the Sapphire Trixx to undervolt by -200 mV.

Did you check all your results vs siacoin bro  ?
Nice info here thanks !