Topic: [Awesome Miner] - Powerful Windows GUI to manage and monitor up to 200000 miners - page 599. (Read 703396 times)

I want to mining....but how..? .I have cor I5 laptop... But i don't   know  how to start....??

after update on version 3.0.3 managed miner stop counting time

Good development. Looks like most of my points you're already addressing. Dynamic IP allocation is pretty standard. So while maybe you haven't gotten a request for adding computer names (hostnames) instead of IP, it could also be because a lot of miners currently use work arounds instead of reporting issues. IE static IP allocation. Hammering out that post for instance took a decent amount of time (as do your responses).

Seems as though you haven't fully implemented and integrated Afterburner and a lot of my issues stem from that.

"No. Do you want this for all mining software, or only specific? The reason why I'm asking is if this could be a global setting to control this behavior."

It's not specific. Sometimes, depending on the algo, depending on the miner you'll run into weird cases of CPU utilization and you have to adjust it. This isn't predictable. Affinity is another option that also would be useful and I do use (assigning the miner to certain cores).

Another issue I noticed is running the MSI remote process is a chore. You can't set it to run on startup without a lot of work arounds, I don't know why it's not integrated into Afterburner as a option, but it's not. If there is a way that AM service could automatically run this process it would be very helpful. I know this is kind of a pointing fingers thing as Afterburner really should take care of this, but it doesn't. Tangent to the above and not a high priority, but if you could set a list of software you want to have running on the PC and have AM start it up and make sure the process is running would be good as well.

Afterburner monitors a lot of this information (I assume you're eventually going to pull CPU information from it), but if you're looking for system information that's pretty accurate, HWinfo covers most of it and comes with it's own API. https://www.hwinfo.com/sdk.php

Going back to the discrepancies between Afterburner reported GPU, System reported GPU, and miner reported GPU stats (which should be noted clearly somewhere), why does AM wait till the miner is running before reporting GPU stats? Is this because some of the information is pulled from the miner? It would be nice if it was reported all the time even if you can't report things like hashrate till the system is fully mining. Sometimes you'll be using a miner that isn't supported by AM as well (which is why being independent from mining software is a good thing).

I haven't take time to play around with this yet, but when running a completely foreign miner (IE nothing AM is built for), I assume AM will monitor the process and while it can't see what's happening inside of it as it doesn't have a API (black box), it should be able to tell when the process hangs according to the system and restart it. Additionally adding the ability for the miner to be restarted every X number of minutes would be useful. Currently a miner unsupported by AM I'm mining with requires me to change the priority on it and since I can't do that, I can't use AM with it.

This is a small niggle, but every time I update AM it breaks my desktop shortcut. XD






Yeah, you want me to pay 50k for someone to manage hostnames? *clap*

Devices change networks, they aren't always on the same network. If you're building your entire system based around nothing changing, when it does, everything shits the bed. There is dynamic IP allocation for a reason. Unless you're talking about a static part of the network that never changes (like a router or managed switch), this is silly. Once you get past 10 rigs you'll start to figure that one out on your own.

I have further investigated this issue and found an obvious bug:
Claymore's software is listening to port 3333 by default and AM was using the same port for pulling miner info in the previous versions, but now it looks like AM is trying to connect to Claymore through port 4028 and as I remember it is default port of CGminer.
After setting port 4028 to Claymore, everything works as it should!

I hope this information will help you to fix it and them who are using Claymore with Awesome miner.

I agree with some of your ideas, but what's about IP allocation, using computer names in large farms instead of DHCP IP allocation is a fools errand, nobody does that.
I suggest to hire a network administrator if you don't have the right skills for it.

Awesome Miner version 3.0.3 is an update with the following improvements:
- GPU Voltage boost can be configured using the MSI Afterburner integration
- Profit switcher more customizable by introducing device profile settings per dual mining algorithm when using Claymore's Ethereum Dual miner
- New setting for Offline Detection rule to include detection of disconnected Remote Agent connections
- Improvements to GPU clocking and GPU mapping features
- Stability corrections

First of all, thanks for putting all this feedback together. I will comment on some of the items below, but keep in mind that I do take notes of all suggestions - so some of your points will be, or already are, planned for the future even if I don't add any specific comment.

I can plan to add a setting for this when you do the network scanning. Although using DHCP is common, most routers assign the same IP to the same computer (MAC address) all the time. However, a router can fail and be replaced, so I do see your point. I've actually never had a request on this in the past, and there are people using Awesome Miner with very large numbers of miners.

Isn't that the network scanning you refered to above? Some people use it to add hundreds of miners at a time.

When you add multiple miners using the network scan, you don't have to create the hosts. They are created automatically. I do understand that you don't want to use this feature now because of the IP vs hostname point above, but this is still just a result of not using the network scan.

I just investigated this one and actually found a bug introduced in the recent 3.0 release, where you sometimes when adding new miners run into this. I correction will be available soon. Performance mode is intended for much more miners than that.

Different mining software reports the GPU's in different orders, and sometimes you don't even assign all GPU's on the system to the same mining software. The same mining software can also report GPU orders differently depending on custom command line parameters. In addition to all that, you can run Ethereum mining on two GPU's and Zcash mining on the other three for example. The only way to support this is to force the user to do some mapping. At first I was thinking of trying to match simliar fan speeds and temperatures to automatically do the mapping, but that's not stable enough.

That one is only available from some mining software, like sgminer. So why don't Awesome Miner bring in this value just like GPU and memory clock? The monitoring library currently used to get GPU information can actually not read the powertune value.

This also brings us to another point about how Awesome Miner get the GPU information. First of all it uses the information from the mining software. When using the mapping feature, you can bring in the GPU monitoring information that Awesome Miner itself (with help of a monitoring library) can get. By doing this all users of Awesome Miner can get much more GPU information without even running MSI Afterburner.

At the moment, MSI Afterburner is only used to set GPU parameters (the GPU clocking dialog), not to display them on the GPU tab. This is subject to change, because if you do run Afterburner, it would of course be fully possible to use the display values from there. There are however fewer users that will run Afterburner, so it will not be available for everyone. In the future there will probably be some concept where Afterburner data will be used if available, and than fallback to the GPU monitoring that is always available in Awesome Miner.

This is to force the user to look at the list of GPU's and consider the mapping. If I would just bring in the GPU monitoring information right away, it would be so many questions from users about GPU statistics not being displayed correctly.

CPU monitoring information is unfortunately only available if you run a process as Administrator. If you would run Awesome Miner Remote Agent as administrator, you would probably see more info.

MSI Afterburner always runs as Administrator and will of course have access to this information. This relates to the point earlier that Afterburner isn't used for monitoring yet.

The default is to show the highest temperature in the miner list. You can change to show all temperatures using the Options dialog, General section, "Temperature display".

Yes, I've recently implemented that and it will be released today. I was also going crazy on this when I was doing GPU clocking.

Please try again in the next release, because I've made some changes to this after getting a few similar reports the last week.

No. Do you want this for all mining software, or only specific? The reason why I'm asking is if this could be a global setting to control this behavior.

Thanks!

Is there a way to start a miner with a lower process priority? Normally you do something like this with a batchfile 'start /low ccminer COMMANDS'

Looking pretty good, bought a large license so threw some money your way today.

Still needs some polish, specifically with adding devices.

-When you add a device through scanning the network it shouldn't add the device based on IP, it should be based on computer name. Almost all miners will have dynamic IP allocation and adding IPs is a fools errand.

-You should be allowed to mass add devices through the network.

-Adding a 'host' still shouldn't be a thing, the device name should be the hostname. Perhaps allowing you to custom overide it. Right now you have to click through adding a hosts dialogue box which is a waste of time.

-Installing the remote service should immediately start it, for some reason you have to manually start it the first time it installs.

-If you add more then 10 devices it should display a warning and tell you that you should probably enable performance mode (AM basically becomes unresponsive at that point).

-With the addition of GPU OCing through AM via Afterburner there needs to be a distinction between what the miner is reporting and what the system sees as devices. As you know a CUDA or OCL device may not be directly named the same as the way the system sees it. I see that you allow people to change the order of the devices through map to system monitoring. I'm not sure what's being reported by the miner, through CUDA/OCL, and through the system. Trying to figure out what connects to what is very convoluted.

--With some trial and error AM could auto-detect what's connected to what with some simple testing. For example having Afterburner set the TDP to minimum for 20s and see what GPU loses hashrate, then adjust it accordingly.

-For some reason powertune stats aren't being displayed on the GPU tab.

-For some reason you have to click on map to system monitoring > display system monitoring data in the GPU tab in order for the GPU tab to actually become active (why is this a option?)

-CPU monitoring doesn't seem to work. That can be monitored through Afterburner as well. AB has built in monitoring for that.

-In the compact list view when you're looking at GPUs it only displays the temperature of one GPU, instead of this you could simply seperate all the GPUs with a comma so you can display all the GPUs from a system there for easy to read access.

-Uploading software is still a PITA. There should be a upload manager for all systems, where it simply syncs the upload folder with the systems that are managed by AM. You could manage this on a per rig level with overides if need be.

-There needs to be a warning in case Afterburner is offline or the remote service is offline so you restart it.

-With GPU overclocking you need to be able to manually enter in a number instead of using sliders. Text entry next to the slider?

-For some reason clicking on 'GPU settings' under 'Tools' gives me the AB/AB service is not installed instead of displaying AB for all systems.

--With a bunch of rigs running it's impossible to see where AB remote is down which goes back to the suggestion earlier.

-Managing dual mining with Claymore is a huge PITA to the point at which it doesn't seem to be manageable by AM. This is due to the way AM manages miners, pools, and mining software. There is way too much overlap between categories. My post a couple months ago outlines how to deal with the logistics of managing all of the above.

-Modifying a template that is currently in use should present a warning and apply it to all existing hosts. Those that are in use should have it applied and restart.

-GPU Overclocking does not display Thermal Throttle Temp. This is important for adjustment (may need to add tic box for locking TDP and Thermal Temp).

Already tried many times, but no changes. Everything is the same for all servers, not showing hashrate/income, "Interface Offline" and "Mining".
Maybe it is not even trying to pull info from the miner software? Is there an option to disable direct Afterburner control and switch back to using miner software for temperatures like before? I tried to use different versions of Claymore too, no changes.

Thanks for the update. I problem right now is that Awesome Miner only grabs statistics from the two main sources whattomine.com and coinwarz.com. None of them list any Skein based coins. Because of this, it will not be much for Awesome Miner to show in terms of coin statistics and profitability.

The only ones that I know of are Digibyte - Skein and Myriad - Skein

Thanks!  i can work with that!

i'll use the reboot computer command and put the other stuff in its startup.

cheers and thanks!

The rules are only executed locally where Awesome Miner main application is running. So unfortunately you cannot execute any command on the remote computer. Even when you define your bat-file locally, it will also only run locally.

However... If you want to reboot a remote system, you can add an action of type "Miner Command", and then select the command "Reboot computer". That should solve the scenario you are trying right now with remote reboots.

If there is a need to run commands (other than reboot) on a remote system, I can of course add support for that.

Thank you. Yes, I know that the download exists, but it is very inconvenient since every time you change the algorithm or currency, the downloads are reset and when you return to that currency you have to do everything again (((

Thanks for the info on 'dead devices' - I am now writing my rule to work if the hashrate drops more than 20mh/s

i need one other piece of information need on Rules;

see image:   https://snag.gy/dxA4eO.jpg

I am writing this rule for a remote machine that awesome miner monitors. My rule action is to run a batch file.
Is awesome miner going to look for the batch on the awesome miner machine or run a batch file on the REMOTE system?

i.e. I need to know where it is going to look for the batch file so I can set the correct path - locally or remote?

thanks in advance

Hi,
Yes, the rules can be quite powerful.

Dead Device Detection is however only supported in some mining software. It's supported by almost all ASIC's (Antminers, ...) running cgminer but also on some GPU mining software like sgminer and EWBF Zcash miner. It's not supported by the Claymore miners. Awesome Miner is simply reacting to what is reported over the API interface in these cases.

Thanks for the suggestion. I'm always trying to add miners that are popular, and I will keep an eye on this one as well.
Until officially supported, I hope that you do know that you can use the Upload Software feature to upload any mining software to all your remote computers in a single operation:
http://www.awesomeminer.com/help/uploadsoftware.aspx

Hi Patrike,

I just discovered the 'rules' section - very cool!  using it to write reboot on gpu crash scripts now 

I need a clarification - in the 'dead device detection' rule - what parameter is being looked at to determine a device is 'sick'?

Can you please try to stop a single Managed Miner, then start it again. Any change?