Hi All! I love knowing how many people this guide has helped and enjoy offering it as a free online resource. That will never change. But, some might be in a position to give back and CircusWife and I would appreciate it if those who can afford it would send what they can as a show of support.
- XMR: 46yn37Go9nP66k2bCy9uQsgsXtJExcZzST78SL2QfU5w7JCZneZRNGCNtfJL49BfEfckPii9g23TiBTdnHct6AQL444nDpY
- ETH: 0x0Db3B49F4075eA688d89E86cbd5750890687F066
- NANO: xrb_3bqdyojqj1tg7ddd94nado8zcbbduwkzrrje8znuhgot9gacdcan5sspzzni
- BTC: 17gywiZLp7BPM2ZZu6FMW57ANswBGuxycJ
Troubleshooting / FAQ Index
If your here for troubleshooting let me start by affirming for you that I have received many confirmations that the guide works. You will get there and are probably close. Don't give up. More often then not when issues arise it is because someone has accidentally (or intentionally) skipped a step in the guide. Take a moment and ensure you didn't take a shortcut that might be causing your problem. If you have exhausted this FAQ and would like to contact me directly, the best way is to PM me on Reddit, I go by u/CircusDad.
My computer crashes with a TDR Failure (BSOD)!
- Click <here> if this crash is occurring when the miner is mining
- If you get a TDR failure spontaneously after booting or resetting the Vega's, and it is not even associated with starting the miner, then it might be that you have a gimpy Vega that can't handle the undervolting you tried to implement via the soft power play table. If this is the case then it is likely you already know exactly which Vega is giving you the problem based on it's behavior during boot-up (maybe it starts working hard for no reason?). You need to boot to safe mode and back out the soft power play table.
- Option 1:
- Boot the computer to safe mode
- Apply this registry file (Vega 56 / Vega 64) that will reset it to stock values
- You need to change file extension to .reg (vs. .txt) after downloading.
- Restart: If the computer is happy again, move to the second phase below...
- Option 2:
- Boot the computer to safe mode
- Double-click the registry file you exported after driver install
- If you did not save the export then you need to uninstall/reinstall the drivers
- Go to each Vega's index and delete the entry called "PP_PhmSoftPowerPlayTable"
- Check to ensure that EnableUlps and EnableCrossFireAutoLink are still set to zero
- Restart your computer... It is now like you never installed any soft power play tables.
- Install the soft power play tables again but this time skip the index number of the card that gave you the issue. (It will be the only one with stock settings)
- Proceed to Phase 2
- Phase 2
- Apply these less aggressive soft power play tables, 50% then if successful, 75% and 85% until you find how aggressive you can be without recreating the problem.
- 50% of the way to guide settings: (Vega 56 / Vega 64)
- 75% of the way to guide settings: (Vega 56 / Vega 64)
- 85% of the way to guide settings: (Vega 56 / Vega 64)
- These were created for your special case so please consider a donation if your in a position to do so :-)
- I left the MEM_P3 at stock settings in the files above to reduce variables (because you can control it directly with OverdriveNTool when you start the miner). The Stock MEM_P3 voltage is very high so I suggest you make a new profile in OverdriveNTool.ini for that "special" Vega and see if you can work the Mem_P3 values all the way to guide settings (950;900 for Vega 56 and 1100;905 for Vega 64 without causing problems (see the tuning page for clarification as required).
- Note: These Power Play Tables were created using a tool made by brnsted. I left his ETH/XMR address in each of the files incase you are feeling so inclined.
My computer crashes when I start the miner
- Reduce variables: We will add a command line to the Hash Monitor script so that xmr-stak will focus on your Vega's and disregard any config setting issues you may have with your CPU Nvidia (if you have one) hardware/config files.
- Open JJs_HashMonitor.ps1 with a text editor
- Scroll down until you see a line that says: (~line 168)
#$global:STAKcmdline = "--config config.txt"
- Remove the "#" in the front and replace the value in quotes per the text below
$global:STAKcmdline = "--noCPU --noNVIDIA"
- Now restart the miner using Run_HashMonitor.cmd. If your system still crashes you know it is not due to CPU miner / Nvidia mining issues.
- Confirm that your AMD Blockchain drivers are healthy:
- Open the OverdriveNTool Gui and use the drop down menu in the upper left to select your Vega(s). If you get the message, "This GPU is not supported" then sadly, that is a sign your driver is corrupted and you need to: (a) Revert to an earlier Windows restore point or (b) uninstall/reinstall the block-chain driver and repeat registry edits (click here).
- Are you sure you are not applying Vega 64 OverdriveNTool settings to a Vega 56? If you have mix of Vegas in your system then you need to double check that you have them right.
- While you are beside your computer open the OverdriveNTool GUI and set the minimum fan speed of each Vega to 0 (it will auto correct to the minimum when you hit apply (I think 1200 rpm).
- Go back to the first Vega in the list and change min speed to 4200rpm and hit apply. Look at you unit... note if it was a Vega 56 or a Vega 64.
- Repeat until you have identified every Vega 56
- Update your OverdriveNTool.bat file with the proper card sequence
- r0 is the first card in the OverdriveNTool Gui Dripdown
- Re-run OverdrivenNTool.bat in order to rest the fan speed on all Vega's
- Do you have a monitor plugged into your Vega? This is not directly forbidden but you are using optimized miner settings that do not include a GPU intensity allocation to serve your monitor (even if your monitor is a remote desktop). Because your Vega's are so good at mining, it is best to have another card (or even better the iGPU if your MB/CPU has one) serve the display. That way you can leave the Vega's to focus on mining. This is explained in the main guide but if you must use the Vega to serve your display then 1932/1932 is too much intensity for the Vega to do both. Artifacts followed by a possible Windows crash are a good tell that this is your problem.
- Try reducing the thread intensity to 1800 per thread for the Vega that is serving your monitor (the other Vegas can stay the standard 1932 / 1932).
- I have a dedicated miner and use intensities of 1908 and 1908 for the one Vega I use to serve the "display" (actually an HDMI dongle). I do get some artifacts from time to time but it preserves me some hash rate and I never use the display so I can live with the artifacts so long as windows is stable.
- Ensure your PSU is has the power to support the full mining load of your cards (this has been the problem in some cases). It sucks to pay for the big PSU but the Vega's need power (~180/190w with the stock settings from the guide, 130/150w if you optimize power over hash rate). The cards do not draw full power when you start and use the computer. If you have inadequate power you would not know it until you draw full load with the miner... but it is not actually a software/settings problem.
- You should size your PSU to be running at about 80% power utilization.
- I ran mine for about two weeks at 95% utilization as measured by a kill-a-watt. This was not a workable long term solution. It will reduce life of your PSU and you should expect instability/crashes. I have since joined two PSU's together to get to the desired 80% utilization.
- FYI - Because my personal miner is located remote to where I am, I wish I had a monitor I could view remotely like this... note that it is not a personal endorsement as I have not used it.
- If you need to combine two PSU's together in order to run your system, you need a way for BOTH PSU's to know when the power is turned on (and off). This is the cable I used... some people use an add2PSU but I haven't
- All you do is plug both PSU's into the cable and plug the cable into the Motherboard.
- The PSU that went to the connector with more then two wires is the PSU that is powering the motherboard.
- Power the CPU from the PSU that is powering the motherboard.
- Be specific with your Vega connections as you want each cards to be tied to a specific PSU. That means for each Vega you want the riser and the Vega to connect to the same PSU (if your Vega is in the MB vs a riser, then you want that one to be attached to the PSU that goes to the MB).
- Note: Vega's are efficient at mining cytptonote currency (and cryptonight-light used in AEON). If you think you might someday want to mine other algorithms, design your system to allow more power per card.
- Did it Blue Screen of Death (BSOD) with a TDR Failure?
- You need to raise the virtual memory. Revisit this part of the guide <HERE>. Maybe you forgot to hit "set"?. The number listed there definitely works but you might need to increase the memory if you are pushing 8 Vega's.... I don't have experience beyond 5.
- If it is happening at bootup and is not even associated with starting the miner, then it might be that you have an outlier Vega :-( that can't handle the undervolting you tried to implement via the soft power play table. If this is the case then it is likely you already know exactly which Vega is giving you the problem based on it's behavior during bootup (maybe it starts working hard for no reason?). If that is the case then I suggest booting to safe mode and backing out the soft power play table.
- Boot the computer to safe mode
- Double-click the registry file your exported after driver install
- Go to each Vega's index and delete the entry called "PP_PhmSoftPowerPlayTable"
- Double check that EnableUlps and EnableCrossFireAutoLink are still set to zero
- Restart your computer... It is now like you never installed any soft power play tables.
- Install the soft power play tables again but this time skip the index number with the card that gave you the issue.
- Consider making a special OverdriveNTool.ini profile for that "special" vega.
- Bad PCI riser? I might be reaching a little here as I have not had this experience myself but I hear that risers can be bad from time to time so if you have the ability to swap some out it may be worth trying.
- Change OverdriveNTool settings. My intent was to post safe numbers that would work for most any system... but everyone would then have to self optimize from their. It is possible that my "safe" values aren't as safe as I thought. If you have Vega 64 try upping the GPU_P7 to 950 and then eventually 1000.
Miner Does Not Run
- "Error CL_MEM_OBJECT_ALLOCATION_FAILURE" at miner start
- You need to raise the virtual memory. Revisit this part of the guide <HERE>. Maybe you forgot to hit "set"?.
- The number listed in the guide definitely works for 5 Vega's but you might need to increase the memory if you are pushing 8 Vega's.... I don't have experience beyond 5.
- I know one data-point where 20000 was working for one Vega but this error occurred when moving to 2 Vega until virtual memory was bumped to 300000
Hash Rate Not As Expected:
- I'm only getting 1000ish h/s on my Vega 64 (930ish h/s on my Vega 56)
- You forgot to add TWO threads per Vega in your amd.txt file. Open amd.txt. Ensure there are two lines for each thread index. Restart
- I'm only getting 1600 h/s to 1700 h/s per Vega
- These are "hash dropped" numbers. Make sure there is no other program running that is causing the Vega's to "hash drop" instantly after you reset them. For instance, GPUz causes hash rate so reliably that I used it to force hash drop when @TheJerichoJones and I were testing the JJs_HashMonitor script. If you have it open then you would expect it to always mine at 1600h/s to 1700h/s.
- Note: It is not a hard and fast rule that no program can be open. I use my Chrome browser and notepad++ from time to time while it is mining (i.e. when I am recording optimization iterations). Still, since you are still in troubleshooting... just close everything.
- I am sorry this is vague as I don't know all the details yet but Windows 10 imposed a forced (hard reboot) update the week of 11/12 and one person reported that they couldn't get back to full speed until they reverted back to the pre-update windows version. What a pain. That is an unlikely root cause in your case so try everything else first.
- So, rule out an xmr-stak configuration issue by giving CastXMR a shot and see if it gives the same result. (I will add instructions on how to use CastXMR here eventually)
'Application has been blocked from accessing Graphics hardware' Error
- Sadly, this is a sign your driver is corrupted and you need to: (a) Revert to an earlier Windows restore point or (b) uninstall/reinstall the block-chain driver and repeat registry edits (click here).
Hash Monitor script doesn't start the miner
- Make sure you put all the files in the same directory (like the guide said ;-). The script will only look for xmr-stak in the directory the .ps1 file is located in (this is true for OverDriveNTool too)
- Did you rename the xmr-stak executable? This is actually OK to do provided you updated the program name variable in the script with the new name. If the names are not the same... a miner that did not start would be the expected result.
Hash Monitor never detects xmr-stak and keeps resetting:
- JJ's detects the presence of xmr-stak by locally accessing the xmr-stak web data. There is likely something preventing this access.
- The guide said this but just checking... did you give a port ID in your config.txt file? Did you update PreciousGuardian_Config.txt with the same value?
- The config file says this but just checking... you didn't turn on login authorization in xmr-stak config file did you? Precious Guardian needs access and does not support authorization. Your xmr-stak config.txt file should have:
- "http_login" : "",
- "http_pass" : "",
- When windows asked if your miner could access the network, did you say yes?
- If not, copy the entire directory to a directory with another name, say "xmr-stak2" and re-run. Windows will ask you again. Say yes this time.
- Still no luck? Direct Message TheJerichoJones on the slack channel. He will hook you up.
System Hash Drop Is Repeatable/Predictable
- Hash drop can be random but it should not happen often. If you are noticing any pattern in your hash drop you should perform this step
- Hit windows start key and type: "Task Scheduler"
- Click on, "Task Scheduler Library" in the left hand menu
- Look at the list of scheduled tasks.
- Look down the "Last Run Time" column. Any correlation with your hash drop issue?
- Look at the "Triggers" column, anything scheduled to run that you don't really need?
- If you identified a task that is causing you trouble, right click on it and select "disable" (assuming of course it is something you can afford to disable).
- Note: If you have an Nvidia card in your system, you may need to disable the Telemetry tracking using the method above (see here).
Why do you use XMR-Stak vs. the Vega Optimized CastXMR?
- I actually wrote a pretty in depth review of my reasons as they relate to performance
- There are other non-performance related reasons why you might make a different choice. You can read some of them here.
How much power do these settings use?
- I will start by reminding you that the guide settings are intended to be safe for most people so they can get up and running at the expense of power optimization. Now that you are up you should play with your GPU_P6, GPU_P7 and MEM_P3 values to tweak for your system
- I do not know the idle component of a Vega's power. I do know that my system with 2x Vega 64's and 3x Vega 56 draws 219W from the wall when I close my miner and have a blank desktop. Again, I don't know but I suspect it is 40W for MB, Hard Drive and Ram... That would make it about 35w for an idle Vega?
- When I start my miner with no CPU mining and only one Vega 64 mining, it draws 347w from the wall. 347w - 219w + 35w = ~163w
-- Copyright Vega.MiningGuides.com 2017 --