i need to reup that video as I mentioned above
but you said you got it working right?
If that still doesnāt work, try booting up with 1 gpu first. Install the driver, shutdown, plug in the rest of the gpus + install driver again.
What follows will seem like a retarded question, but keep in mind Iāve been a Mac user since 1984, and this is my first foray into the world of PCs. So ā¦ 1) How do you change the primary display adapter from onboard to PCI-e when I donāt have a function monitor / GUI? 2) I downloaded the AMDGPU Pro driver, but didnāt actually install it. I figured I couldnāt until I had the R9ās installed; otherwise it would be looking for GPUs that didnāt exist. Mistake on my part?
@mrb Feature Request: log to a file and ā-v 0ā means no logging. I tried running silentarmy in the background and redirect the console output to a file, but the logging doesnāt show up there, so I canāt run in the background without console output continuing to output to the screen. Not very helpful on 20 headless rigs with 6 GPUs each. Thanks for your great work on this!
So ā¦
- is this just a matter of unplugging one of the R9ās from the power supply or
- do I have to also pull it out from the card slot?
well u said it worked when no GPUs are in, does it still work without any GPUs in?
if so, then boot in like that, change it, save, then shutdown and stick your cards back inā¦
if it still doesnāt work, clear your CMOS and back to before
otherwise I dont know, try what chtat12 said
I have 6 GPUs so Iām doing:
echo 150 > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/hwmon/hwmon1/pwm1
echo 150 > /sys/devices/pci0000:00/0000:00:1c.0/0000:02:00.0/hwmon/hwmon2/pwm1
echo 150 > /sys/devices/pci0000:00/0000:00:1c.1/0000:03:00.0/hwmon/hwmon3/pwm1
echo 150 > /sys/devices/pci0000:00/0000:00:1c.2/0000:04:00.0/hwmon/hwmon4/pwm1
echo 150 > /sys/devices/pci0000:00/0000:00:1c.4/0000:06:00.0/hwmon/hwmon5/pwm1
echo 150 > /sys/devices/pci0000:00/0000:00:1c.5/0000:07:00.0/hwmon/hwmon6/pwm1
Fan speed set to around 55%
not sure what the problem your having is, but would an HDMI switch that has many hdmi ports in it to hook up a bunch of rigs to one monitor help?
Well if changing the primary display to pcie doesnāt help, try unplugging the riser of 2nd Nano from mobo and then try rebooting again. My mobo wonāt boot with > 1 GPU when thereās no driver installed. I had to boot & install driver with 1 GPU first before plugging in the rest.
Edit: I assume you just plugged your monitor into the Nano without changing the primary display first? If so, you plug your monitor back to the motherboard, change primary display to pcie, and then plug your monitor to the Nano & reboot.
1 Like
Thanks for the input. Will do as you suggested and see what happens.
yes, i did!!! <3 only maxing out at like 9.5-9.7ish with an i5-4690k. how do i monitor everything, so that at the end of the day i know how much i mined, and whether or not iām getting paid out correctly and all that jazz? ALSO, one big question, how do i correctly stop the miner? lol
do you know how to look at claymores linux files? what encoding do you need to view it?
Heads up: I committed an optimization from nerdralph: improve perf by 5% (OPTIM_FOR_FGLRX) Ā· mbevand/silentarmy@a33a906 Ā· GitHub
In my tests, I notice a +5% performance increase in some cases. Or a -15% perf decrease in others. Itās a bit inconclusive, but worth a try. It seems to depend on the driver (fglrx vs amdgpu-pro) and on the hardware. To enable the optimization, edit param.h, set OPTIM_FOR_FGLRX to 1, and recompile.
More details:
My test machine with an R9 Nano can dual boot into either:
- Ubuntu 14.04 64-bit with the Ubuntu-packaged fglrx driver 2:15.201-0ubuntu0.14.04.1
- Ubuntu 16.04 64-bit with AMDās official amdpro-gpu driver 16.40
Performance results in sol/s, running 1 instance of āsa-solver --nonces 100ā:
- 26.8 / 27.7 (fglrx / amdgpu-pro) OPTIM_FOR_FGLRX=0 (default)
- 28.3 / 23.4 (fglrx / amdgpu-pro) OPTIM_FOR_FGLRX=1
It seems to improve performance by +5% with fglrx, but degrade performance by -15% with amdpro-gpu. However some users have reported either no change, or a +10-15% increase.
I just ssh into each rig from my laptop. The rigs are on another floor in my building. I have not need for a GUI interface to my mining rigs. For ETH, I run the miners as background processes and redirect the console logging output to a file. I canāt get silentarmy to do that, which I have seen off and on with other python apps. If I just put silentarmy in the background, the console output continues to come out on the screen, so itās useless for doing any work.
Working great for me, thanks! about a 5% increase. (AMDPRO drivers)
Is there any way to compile it without all the openCL goodies just so i can mine on the CPUs?
@dlehenky Output doesnāt immediately show because by default it is buffered and flushed every 4kB or so. But ok, Iāll consider implementing logging (if someone submits a pull request, thatās even better!)
In the mean time, redirect both stdout and stderr, and accept that it will only flush to the log file every 4 kB:
$ silentarmy >log 2>&1
marc, someone just said that claymore is using your code in his miner.
how can I view the claymore files to see his code?
and can you confirm this for me please?
did he use your code and not give you credit?
and can you please release a windows version of your miner before you release NVIdia
if you release NVidia, claymore is just going to take your code again and not give you credit and rake in tons of cash
if you can release a windows version firstā¦ you can gain some branding ground
AND if you can put a real copyright on your NVidia miner, you can get better exposure as people will rush to you, and claymore will not be able to impliment your code
you can then change the copyright aftera few weeks so others can use your code
but that few weeks will give you exposure and you will gain brand name acceptance from all minersā¦ chances like this do not come around very often in lifeā¦ I think that you should protect your code now so that you can benefit from it the most - it is not fair if someone just sits around, lets you do all the hardwork, and then takes credit for your work because they are well known in the mining community.
this could also raise questions on if his ethminer contains code from someone elseā¦ which might legally be ok, but morally its wrong to not give credit.
Thank you for your work, i want it to be properly appreciated
"mickr1h cryptomined
he
says he locks up his miner so tight so people wont copy his codeā¦ well
MAYBE he locks it up so tight so NO ONE WILL SEE HE IS NOT USING HIS
OWN CODE!
Looks like he didnāt pack / encrypt the linux version, just looks like a standard elf file.
If you open the linux version in a hex or text editor, youāll see strings such as ā¦
miner/blake.cpp
n > k hash_len <= 64 ZcashPoW msg_len <= 128 outlen <= 64
void zcash_blake2b_init(blake2b_state_t*,
uint8_t, uint32_t, uint32_t)
zcash_blake2b_update(blake2b_state_t*, const uint8_t*, uint32_t, uint32_t)"
can you teach me how to view his code so that I can teach others what is going on?
I really think people should know this as it is being hidden from them and he is really taking credit for something that is not his
and i noticed about a 3% increase in performance with nerdralph
You could use it like this.
unbuffer ./silentarmy --use=1 -c $STRATUM -u $USER -p $PASS >& silentarmy.log &
Just run āstringsā on the binary file. It will show every ascii string in there. That should give you a decent idea.
how do i do that?
whats strings?
can marc confirm that claymore is using his code?