I'm having the same issue with nvOC, so we can rule out smos being the culprit. https://www.nicehash.com/miner/3HQMKEJT88ZdGLtJAZyRYA6EacknELcbbjI threw this little table together using dstm stats and some custom commands within nvOC. Handy reference for the various flavors of current gen nvidia GPUs since I run a mix of just about all of them. The big looser is the 1050Ti's for sure.  Check out the max power limit on that Zotac 1080Ti ArcticStorm in miner05. 432 watts  No wonder I have to run it as 230W just to get it to "wake up". |
|
|
|
i am running it inside smos/linux pointing at nicehash, I don't like that it shows 8 miners on an eight card board not a sum of all eight cards https://www.nicehash.com/miner/1JdC6Xg3ajT3rge3FgPNSYYFpmf53VbtjeYes Equihash 348.15 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 464.21 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 565.75 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 464.21 Sol/s 14.51 Sol /s (3.03%) 256 112 USA Yes Equihash 381.81 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 420.69 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 449.7 Sol/s 0 Sol /s (0%) 256 112 USA Yes Equihash 536.74 Sol/s 0 Sol /s (0%) 256 112 USA I plan to point five 8 card boards this will mean 40 workers I do not like this can this be fixed? ebwf does not do this it shows 1 worker for the 8 card mobo. I realize it could be nicehash smos dstm so I will start with dstm since nicehash smos ebwf shows 1 board for 8 cards I'll second that. Having the same issue running dstm on 8 nvOC rigs. So we can rule out smos. https://www.nicehash.com/miner/3HQMKEJT88ZdGLtJAZyRYA6EacknELcbbj |
|
|
|
|
So are we at a crossroads?
On one side we have poisonxa's 19-1.5, and on the other side we have papampi's 19-2 (with contributions from several other members here).
I have been doing a lot of testing with stubo and know a lot of tweaks to the watchdog script have really helped with stability. Very few of these tweaks are in the current 1.5 poisonxa code and several bugs that have been discovered are still present there as well.
Isn't this the "official" thread for nvOC?
Why did some run off to the discord channel? Was some need not being met here?
Hopefully fullzero will be back at some point to "get the band back together".
|
|
|
|
It would be really nice if there was a autostart feature built in the miner so we don't have to sit and baby sit it.
I am running Ubuntu 16.04 LTS with two Nvidia 1070s
There is, it's called nvOC (Ubuntu based) and has a build in watchdog script. Sample email from watchdog script: Sun Dec 3 09:49:53 EST 2017 - Debug: NUM_GPU_BLW_THRSHLD=8, COUNT=0, RESTART=0, REBOOTRESET=6
Sun Dec 3 09:49:53 EST 2017 - Problem found: See diagnostics below:
name, pstate, temperature.gpu, fan.speed [%], utilization.gpu [%], power.draw [W], power.limit [W]
GeForce GTX 1070 Ti, P8, 34, 50 %, 0 %, 11.80 W, 130.00 W
GeForce GTX 1070 Ti, P8, 30, 50 %, 0 %, 11.76 W, 130.00 W
GeForce GTX 1070 Ti, P8, 31, 50 %, 0 %, 11.90 W, 130.00 W
GeForce GTX 1060 3GB, P8, 30, 50 %, 0 %, 4.53 W, 85.00 W
GeForce GTX 1060 3GB, P8, 26, 50 %, 0 %, 4.44 W, 85.00 W
GeForce GTX 1060 3GB, P8, 30, 50 %, 0 %, 4.63 W, 85.00 W
GeForce GTX 1060 3GB, P8, 31, 50 %, 0 %, 7.26 W, 85.00 W
GeForce GTX 1060 3GB, P8, 30, 50 %, 0 %, 4.14 W, 85.00 W
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
# GPU0 connected to: us-east.equihash-hub.miningpoolhub.com:20570
SSL_connect failed r:2 error:140750DD:SSL routines:ssl23_connect:ssl23 doing session id reuse
Sun Dec 3 09:49:53 EST 2017 - Unknown miner error detected and a miner restart did not work, so restarting system in 10 seconds
|
|
|
|
Unfortunately those 25MM msata drives are too short. There is only a screw down stand off for 40MM units. This is the case on both the Onda and Colorful boards. |
|
|
|
|
I started out with NH like a lot of other newbies and switched to NHLM when it came out.
I have since switched to nvOC and was mining ZEN for a good while and holding. Then sold when it hit $13 (yes I should have waited until it hit $26 for double the profit, but I pulled the trigger to soon). At any rate, I still did well.
Since then, I have been locked in on mining ZEC exclusively and about once every week or two, I exchange my ZEC holdings for BTC. I try to time it to when ZEC is peaking, but it can be tricky.
I have 7 rigs that between them have:
21x 1080Ti
9x 1070
3x 1070Ti
5x 1060 3GB
7x 1050Ti
I'm pulling in about 0.35 ZEC per day on average.
After reading this thread, I'm tempted to try NH again so see if I'd do better. I know I'm missing out on profitable alts for short bursts when they happen, but as others have mentioned, I have a day job and don't have time to switch 7 rigs around to mine whatever is the most profitable at any given time.
|
|
|
|
Looks good! Would certainly help clean up the rats nest I currently have:  Any pics of the fan support beam with fans installed? I'm having a hard time visualizing where they would mount exactly. |
|
|
|
|
The funny thing is that I purchased all those cards used on eBay for between $340 and $390 with free shipping over the last 4 months. The only card I purchased new was the FE that I got directly from nvidia the week it was released.
Yeah, the 2400W Delta is overkill, but it is nice and quiet with only ~1,100W being drawn from it. I have a 8x1080Ti rig that draws 1800W from a Delta, and it is pretty loud. A 2nd 8x 1080Ti rig is in the living room, and I had to put a 2nd Delta on it to quiet it down.
The Delta's are a lot cheaper than the EVGA 1200P2, let alone the 1600P2 or the Corsair AX1500i (I run them all in various rigs).
|
|
|
|
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?
That could potentially give you another 10 - 15% more net power by reducing losses. I know they do things like that in large datacenters sometimes. Unfortunately, 12V PSUs are still not mass produced and therefore expensive. However, compared to high end consumer PSUs and solar panels, it may be very profitable. You could perhaps implement it by using some properly dimensioned battery bank like an UPS and then add some custom wiring to the GPUs.
Interesting idea! One would need some sort of DC to DC converter to get to a constant 12V output. And also some way to switch back to AC power at night when the sun doesn't shine. Here's what the DC output of a typical solar panel looks like during the day:  Solar inverters are quite efficient at 97% typically, and good Platinum PSUs are around 94% efficient. So there would be a little gain to be had for sure. |
|
|
|
|
Very cool. I also have a residential solar system to assist with power cost for mining. I pay a flat rate of $0.12 per kWh year round. Currently have 84 260W panels and am in the process of adding an additional 48 280W panels. The 84 panels produced the following numbers for me this summer:
March 2,391 kWh
April 2,121 kWh
May 2,268 kWh
June 2,578 kWh
July 2,846 kWh
August 2,475 kWh
September 2,172 kWh
October 2,152 kWh
My usage last month was 6,644 kWh, so the 48 additional panels will help, but I'll need another 100 panels at least to "break even". And since I also constantly add miners, I don't think I'll ever "catch up".
I do all the work myself, so the cost is pretty close to $1 per watt and I get 30% Federal tax credit, so that makes it closer to $0.70 per watt.
|
|
|
|
One of my rigs has 7 EVGA 1070's of varying vintages, along with a 1070FE. Here's the rig:  Left to right I have: 0 1st gen SC 1 1st gen SC 2 1st gen SC 3 2nd gen SC 4 2nd gen SC2 5 FTW 6 FTW2 7 FE They all share these settings: POWERLIMIT_WATTS=120 CORE_OVERCLOCK=130 MEMORY_OVERCLOCK=1100 (550 in windows) And they average 478 Sol/S and just a tick above 4 Sol/W.  |
|
|
|
I'm running the Onda D8P with nvOC with no issues. I got 1070Ti's in GPU slots 0 and 2, and 1060's in slots 3-7. I have monitor connected to GPU in slot 0. I don't believe I changed any of the defaults in the BIOS. It just worked.  I installed nvOC onto a little $20 Msata SSD. |
|
|
|
Hi. Thx for the miner.
Small suggestion about API: can you please add GPU model to JSON-RPC responce? Thx!
This was requested already. It seems like this information isn't available on linux. It's on my notes, I'll take a closer look at it later. You should be able to do something like this to at least get the vendor added: #!/bin/bash
DEVICE_IDS=$(nvidia-smi --query-gpu=pci.sub_device_id --format=csv,noheader,nounits)
i=0
for ID in $DEVICE_IDS
do
VENDOR_ID=${ID:6:5}
# GPU Query, 3842=EVGA, 1462=MSI, 10DE=Nvidia, 19DA=Zotac, 807D=Asus, 1458=Gigabyte
case $VENDOR_ID in
3842) VENDOR_ID="EVGA";;
1462) VENDOR_ID="MSI";;
10DE) VENDOR_ID="NVIDIA";;
19DA) VENDOR_ID="ZOTAC";;
807D) VENDOR_ID="ASUS";;
1458) VENDOR_ID="GIGABYTE";;
esac
MODEL=$(nvidia-smi -i $i --query-gpu=name --format=csv,noheader,nounits | tail -1)
echo "$i,$VENDOR_ID,$MODEL"
i=$(($i + 1))
done
|
|
|
|
|
Does Awesome Miner support remote controlling miners running nvOC (Ubuntu)?
My farm of nvOC based miners is growing and being able to centrally manage 1bash, 3main and the watchdog scripts (or perhaps in case of the watchdog script) be a replacement for it, would be very nice.
I'm running DSTM ZM Miner 0.5.6 on all rigs. I know that Awesome Miner recently added support for it, but it is not clear to me if Awesome Miner is only capable of interfacing with the ZM miner's API, or if it can manipulate GPU clock rates, power level, coins mined, etc, which are all normally selected within the 1bash script on each rig.
|
|
|
|
|
So 12 months ago it was at $733. Today it is at $9,685. So a 13x increase. Suppose that rate of increase will be sustained:
11/26/2018 = 13 x $9,685 = $125,905
11/26/2019 = 13 x $125,905 = $1,636,765
11/26/2020 = 13 x $1,636,765 = $21,277,945
What are the indications that bitcoin will NOT continue to grow in 2018 and beyond at the same rate it has been doing in 2017?
|
|
|
|
Can you really fit 13 cards inside? Be very curious how they cram 13 cards into it. I have 5 1080Ti's in mine and it is pretty tight, then then I'm water cooled with dual pumps/res.  |
|
|
|
|
At less than 40% load, not bad at all (GPU's make more noise generally speaking). Once you get around 70%, those tiny fans start getting up there in rpm where you notice them, but they don't really start to "scream" until you get past 80%.
|
|
|
|
Could someone please write clock settings for 1080ti for getting over 750sols? I have asus strix gaming 1080ti.
750 Sols on air is pushing it I think. I primarily run EVGA SC's and SC2's and they do around 710 Sol/s at modest OC settings (190W PL, +120 Core and +600 Memory). Here's one of them:  If you're water cooled, you can push things further. Here's a rig with 4x MSI Sea Hawks EK's and 1x Zotac ArcticStorm. The Sea Hawks are running 190W PL, +155 Core and +700 memory. The Zotac is running 230W PL, +230 Core and +700 memory.  |
|
|
|
for the onda b250, what are the front panel connector colors for power and reset? i could just start shorting them and find out but figured i would ask.
Green - Power LED Red - HD LED Blue - Reset Switch Black - Power Switch |
|
|
|
|
Show Posts
