Topic: CGWatcher 1.4.0, a GUI/monitor for CGMiner & BFGMiner to help minimize downtime - page 24. (Read 402544 times)

How to enable auto intensity?
It does not work on my 6970s and 5970s
What it does if I use auto-gpu then cgminer turns off overheating gpu for good.
if I don't use auto-gpu then gpu overheats and burns down
CGWatcher on one of my rig did auto adjust intensity and it was miracle, on rest of the rigs it ignores overheating problem
can't rely on it


--scrypt  -o stratum+tcp://stratum.give-me-ltc.com:3333 -I 16  --auto-fan --auto-gpu


Miner GPU Report:


--[Miner GPU0]--------------------------------------------------------------------------------------
AMD GPU
Index=0
Name=AMD Radeon HD 6900 Series
CodeName=Cayman
ID=0
Available=True
ADL=0
OpenCL=0
DeviceID=0
BusID=-1
Driver=opencl
DevicePath=
Identifier=
IsMining=True
Status=ALIVE
Vendor=Advanced Micro Devices, Inc.
VendorID=4098
OpenCLPlatformID=8791401843208
OpenCLPlatformIndex=0
OpenCLDetected=True
OpenCLVersion=OpenCL 1.2 AMD-APP (938.2)
OpenCLDriverVersion=CAL 1.4.1741 (VM)
OpenCLDeviceID=494659680
OpenCLMaxClock=880

--[Miner GPU1]--------------------------------------------------------------------------------------
AMD GPU
Index=1
Name=ATI Radeon HD 5900 Series
CodeName=Cypress
ID=1
Available=True
ADL=1
OpenCL=1
DeviceID=1
BusID=-1
Driver=opencl
DevicePath=
Identifier=
IsMining=True
Status=ALIVE
Vendor=Advanced Micro Devices, Inc.
VendorID=4098
OpenCLPlatformID=8791401843208
OpenCLPlatformIndex=0
OpenCLDetected=True
OpenCLVersion=OpenCL 1.2 AMD-APP (938.2)
OpenCLDriverVersion=CAL 1.4.1741 (VM)
OpenCLDeviceID=495712336
OpenCLMaxClock=880

--[Miner GPU2]--------------------------------------------------------------------------------------
AMD GPU
Index=2
Name=ATI Radeon HD 5900 Series
CodeName=Cypress
ID=2
Available=True
ADL=2
OpenCL=2
DeviceID=2
BusID=-1
Driver=opencl
DevicePath=
Identifier=
IsMining=True
Status=ALIVE
Vendor=Advanced Micro Devices, Inc.
VendorID=4098
OpenCLPlatformID=8791401843208
OpenCLPlatformIndex=0
OpenCLDetected=True
OpenCLVersion=OpenCL 1.2 AMD-APP (938.2)
OpenCLDriverVersion=CAL 1.4.1741 (VM)
OpenCLDeviceID=494983504
OpenCLMaxClock=880

--[Miner GPU3]--------------------------------------------------------------------------------------
AMD GPU
Index=3
Name=ATI Radeon HD 5900 Series
CodeName=Cypress
ID=3
Available=True
ADL=3
OpenCL=3
DeviceID=3
BusID=-1
Driver=opencl
DevicePath=
Identifier=
IsMining=True
Status=ALIVE
Vendor=Advanced Micro Devices, Inc.
VendorID=4098
OpenCLPlatformID=8791401843208
OpenCLPlatformIndex=0
OpenCLDetected=True
OpenCLVersion=OpenCL 1.2 AMD-APP (938.2)
OpenCLDriverVersion=CAL 1.4.1741 (VM)
OpenCLDeviceID=519653616
OpenCLMaxClock=880

--[Miner GPU4]--------------------------------------------------------------------------------------
AMD GPU
Index=4
Name=ATI Radeon HD 5900 Series
CodeName=Cypress
ID=4
Available=True
ADL=4
OpenCL=4
DeviceID=4
BusID=-1
Driver=opencl
DevicePath=
Identifier=
IsMining=True
Status=ALIVE
Vendor=Advanced Micro Devices, Inc.
VendorID=4098
OpenCLPlatformID=8791401843208
OpenCLPlatformIndex=0
OpenCLDetected=True
OpenCLVersion=OpenCL 1.2 AMD-APP (938.2)
OpenCLDriverVersion=CAL 1.4.1741 (VM)
OpenCLDeviceID=495352416
OpenCLMaxClock=880


STATUS=I,When=1374447825,Code=32,Msg=GPU 0 set new intensity to 0,Description=cgminer 3.3.1,

If I get this correctly it switches to most profitable with restart of cgminer right? Would it be possible to do by just changing pool(same tipe of coin)? My cgminer sometimes dosen't start right but when it is running it is stable... I would like to avoid restarts

Just tried the new version, and it is not starting my version 3.1.1 of bfgminer.  Going to try the updated version of bfg and report back to see if watcher is able to start it.

*EDIT*

Okay, for some reason I dont know if it is bfg, or watcher, but something no longer likes config files. So I had to dump all my old settings and go with command lines.  Seems to be working now.

Version 1.2.3 is available and should have most of the features that had been requested, most notably:

• Restart computer if accepted or total share count doesn't change for N minutes.
• Config files can now be set to URLs in Profile Manager.
• New scheduled action: change intensity for specified GPU.
• Switch to most profitable, most profitable (adjusted), most profitable (average), lowest difficulty, highest reward, lowest network hashrate, lowest network hashrate (average), or next profile.
• For each of the conditional profile changes, you can set 1st, 2nd, 3rd, etc. to get 1st most profitable, 2nd most profitable, 3rd most profitable, etc. of the coins you mine (have linked to a profile)
• New scheduled action frequencies: Run on specific date at specific time, run when computer becomes idle, run when computer becomes no longer idle, run when specified program starts, run when specified program stops. These can trigger any scheduled action.
• Allow customization of FPGA and ASIC devices to correct the miner reporting them incorrectly (e.g. BFGMiner reporting all ASICs as FPGAs. You can even create custom codes and names for devices.
• Performance improvements, including CGWatcher startup.
• Auto-update option with included utility to automatically update to new versions.

Assuming I didn't break anything too bad, the CGRemote beta will be next.

In case CGWatcher will be able to shut down when  the driver crash. Or in case of a power outage . You can do as follow:

1) Configure BIOS Settings.
Power on your computer, and press the “delete” key a few times immediately after power on. You should end up in the BIOS configuration area. Do the following, then save & exit:
Change power options so that the computer automatically turns itself on whenever power is restored. The reason for this is two-fold: first, it’ll make sure that your miner automatically starts up after a power outage.

2) Configure automatic login
We want our mining rigs to boot up and start mining without any intervention on our part. We’ll need to enable auto-login for that to happen.
Click on the Start Menu and type “netplwiz” (without the quotes) into the search box, and click on it when it appears in the programs list.
Uncheck the box that says “Users must enter a user name and password to use this computer”.
Click “Apply”. You’ll be prompted to enter your password to confirm the change.

3) Configure cgminer and/or  CGWatcher to start automatically
We want cgminer to automatically start mining whenever the rig is powered on. That way, we keep mining losses to a minimum whenever a power outage occurs, and we don’t have to worry about manually starting it back up in other situations.
In your cgminer folder, right-click on your new mine_litecoins.bat file, and click “Copy”.
Now, click on the Start Menu, then right-click on “All Programs”. Finally click on “Open All Users”.
Double-click “Programs”, then double-click “Startup”. Right-click anywhere on the empty background of the Startup folder and select “Paste shortcut”.
You should see a shortcut to your mine_litecoins.bat batch file appear in the startup folder. Windows should automatically execute it upon bootup.
At this point, we’re done with the essentials. If you’ve done everything correctly, you should be able to reboot and watch your rig automatically start mining shortly after the Windows desktop loads. 

4) Until CGRemote beta will start configure remote administration
Configuring remote administration will allow you to disconnect the monitor, keyboard, and mouse from your mining rig and manage it from another computer  over the internet, if you wish. Install VNC like TightVNC or Google Chrome Remote Desktop. Both are lightweight, simple, and free.

Hope it helps

Yes, I checked on the pool to see it is not mining and once I get to the computer to throw the switch, the GPU's are blowing cold air which is a pretty safe bet it's not mining
I talked to a friend of mine who has some server maintenance experience and he told me to wait at least 15 minutes after I tried the shutdown command since windows takes it's time shutting down all unresponsive services (even with /f since this only applies to programs running). So I took his advise and as it seams the system really does reboot if you wait long enough after issuing the command.

So a feature to restart the computer if accepted shares don't change for a long time would be perfect for fixing crashed video drivers.

Thank you for taking your time addressing this problem and I'm really looking forward to future versions and definitely cgremote
Keep up the good work. Obviously I will send over some coins for your help.

I checked and found I didn't have any logging for computer restarts, so I've added some in 1.2.3 so you can check the logs to see if a restart was attempted. By default I set it to not use the force close flag, though I've also changed this in 1.2.3 so it always will. It gives a 30-second countdown before restarting, as well as a Cancel button, so I don't think the force close flag is too intrusive.

CGWatcher also disables the WER UI on starting which prevents the "Not responding" prompt from appearing, which could also cause a halt. CGWatcher restores the original value on close. While it has done this for many previous versions, in 1.2.3 it saves the original value to the INI file in case it is not restored before the computer is restarted, allowing it to be restored to the correct original value the next time CGWatcher closes.

Aside from you losing the remote display, how have you verified that it stops mining? By checking the pool? If this is the case, an option to restart the computer if accepted shares stop increasing for X minutes or something similar may work. I plan on adding several triggers like this to the schedule, so you could create a scheduled action to restart the computer if accepted shares don't increase for 10 minutes, for example. My goal is to provide enough scheduling options that no matter what condition or action you want, you can schedule it. The difficult part is figuring out how to keep the process of creating these actions as simple as possible. Along with creating actions to run at a certain day or time, there will be actions that are triggered by events or circumstances.

And the CGRemote beta will be starting soon, which assuming the computer is not frozen, would allow remote access to CGWatcher. It can also access the miner directly for non-Windows systems, but that method is somewhat limited to control because if the miner stops responding or its API stops working, you have no control. Future plans involve creating a small Linux and OSX "listener" app that is lightweight compared to CGWatcher but is able to kill and launch the miner using commands sent to it from CGRemote. I'm not sure how soon I'll be able to get to that though. So for Windows-based miners running CGWatcher, as long as CGWatcher is running and responding, you can initiate a computer restart (or many other actions) using CGRemote.

Yes, I have set this option but it doesn't work for me.
Ofcourse I can't say for sure that it's the display driver crashing that is causing the problem but when it stops mining and I can't get a remote desktop but ping still works then that's the only reason I can think of since there is nothing else running on the PC that can crash it (ofcouse it's still windows so everything is possible).

Maybe cgwatcher is trying to restart the PC when the driver crashes but windows just refuses to shut down.
I have done a test today with shutdown.exe to remotely command the problematic PC to reboot. It does accept the command and starts shutting down but it never finishes shutting down even if I use the /f (force programs to shut down) parameter. It's probably just waiting for something indefinitely

Maybe it's time to invest in one of these: http://www.digital-loggers.com/220.html

Have you tried the 'Restart computer when sick or dead GPUs are detected' monitor option? In my experience, display drivers crashing always cause a GPU to become sick. So if this is always the case, this monitor option would work. But... this might not always be the case so if you have seen a driver crash that hasn't resulted in a sick GPU (it doesn't happen right away, it may take a minute or two), let me know. If that is the case I would have to figure out a different way of detecting driver crashes... which I wouldn't say is impossible, but isn't something I've looked into.

I've spent most of today adding additional schedule options and changing the way they are triggered to make sure none are ever skipped, so I'll try to get the additional profitability options worked in. I may have to redesign the scheduled action form because it is growing more and more complex, at least from my point of view. For the user it should still be comparable to building a sentence.

And thank you for the donation.

+1, yes please.


And as a pie in the sky request...
When doing 'most profitable' coin jumping, it is easy to spot the masses jumping in and out of the highest percent coin of the hour, which leads to excessive stales and orphans. If you could have an option to mine the "2nd" or "3rd" most profitable coin instead of #1. Doing things manually with this strategy works for me, but would like to see it built into CGWatcher.


LTC Tip sent for the great work so far.

Is there a way to have cgwatcher reboot the computer when the display driver crashes?
I sometimes have the driver crash and I can't access the system anymore by remote desktop since there is no desktop anymore (no GPU's = no monitors = no desktop). The system is still running though since I can ping it but there is no way for me to get on it to reboot other than to drive over and flip the switch.

Thanks. When you say it is failing to auto adjust intensity, are you referring to overheat protection lowering intensities? Or adjusting intensity when idle... or adjusting intensity on a schedule? Debug reports will be helpful any time there is a problem... you can create one in the Tests tab (Create Debug Report button). Then email me the report. The log entries are particularly helpful here, and the Debug report shows the last 50 I think, so it is important to do it either around the time that the problem occurs or to also email me the CGWatcher.exe.log file, as it contains all log entries.

Good programme, but on some mining rigs CGWatcher is not auto adjusting intensity, same rigs, same 6970s and 7950s
Don't know why

That makes sense. I was thinking in terms of one exchange not having a buy order and had forgotten that coinchoose looks at many exchanges. So if there are no buy orders, things definitely do not look good for the coin, at least in short term. If it is resurrected from the dead then coinchoose should pick it back up and its profitability will be updated as usual.

I'll do it this way and set profitability to zero for all three profitability fields. I should have an update either today or tomorrow. There are more scheduled actions I plan on adding so might try to get one or two new ones in this update.

Thanks.

I think setting to 0 is a better option, any coin with no buy orders is almost certainly going to die, and mining it at that point is probably a waste of time, but it would be nice to keep the profile/settings intact just in case it makes a comeback.

Taking the infinitecoin above, there is a huge sell wall at the lowest possible price, so even if a buy order does appear it's going to be instantly filled so the program would be unlucky to check during that small window, and if that wall does dissapear then a switch would occur.

What would be the expected behavior for coins that have dropped off of coinchoose? Should their profitability data remain the same (until they are back on coinchoose or until you close CGWatcher, which loses profitability data)? I would think that would be better than setting profitability to zero, since a buy order could appear at any time causing it to appear back on coinchoose and if I did set profitability to zero, it could potentially switch profiles to mine another coin during this period depending on what you have scheduled.

So I think it is best to ask what users would expect or want it to do before finishing the fix. Thanks in advance for suggestions.

Thanks for the quick feedback will keep an eye on future developments

Thanks for the info. I didn't realize coins with no buy orders would drop off completely. I hadn't really thought about it, but I guess I would expect their profitability to be unknown but for them to remain on the site. That is my misunderstanding, I'll be addressing this issue today.

Right now the timers for scheduled actions running every N hours start when CGWatcher starts. They are reset if either a) the action was disabled and is re-enabled, or b) the action is modified (Save button clicked in Schedule Action window). So currently those are the only ways to sync them. I can add an option to reset the timer at a certain time, but I have to think about it a little more so it might be a few days.

I'll also keep the percentage thing in mind as I will be adding more types of actions and triggers soon so this would be a good time to add that option.

Great program, only been trying it for a day but love it. I'm a newbie so maybe I'm doing something wrong, but think I might have found a problem

I have profiles set up for individual coins, and it swaps as instructed according to the schedule, but it seems that when a coin drops off the coinchoose site it sends the entries funny.

EG, I had a profile for InfiniteCoin



Because there are no buy orders, it has dropped off the coinchoose site and changed the coin selection to Feathercoin (which is next up in the list) and so when the time comes to swap to FTC it ends up mining on InfiniteCoin instead

Edit:
Question while I'm posting, currently have 4 mining rigs and each one is running a different coin due to the way the timers appear to work (I have it checking every xxx minutes). What's the easiest way to get them all to sync up so the checks are done at the same time.

Edit 2:
I think a good feature would be to allow the user to set a percentage in relation to switching pools, when it checks for the most profitable coin if it's within the figure set it doesn't swap, with most of the pools using last N shares I think sometimes it would be more profitable to stick where you are and would cut down on pool hopping when the difference is very small.

IE
Percentage 5%
Coin currently being mined is at 200%
To swap best alt coin would need to be 210%

It looks really nice! Thanks!