EVGA

Helpful ReplyHelp can't seem to get dnetc to run on Windows 10

Author
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
2018/11/14 05:04:09 (permalink)
QuintLeo
The average window AC unit uses a similar amount of power as an average ASIC miner, or 2 6-GPU riser rigs.
Miners tend to have more then two rigs though and they're on 24/7 where a window A/C unit is rarely on 50% of the time over the course of a day even on the hottest days.
 
The BIGGEST Window A/C unit I'm aware of uses about 1.5 times the power draw of a Bitmain S9.
 
It's not ALL power companies that care - it's the ones that are in very low cost electric areas that are getting a LOT of cryptocoin miners relocating for the low power cost (Chelan County Washington being probably the FIRST) that are doing the "mining specific rate" stuff, as it's having a MAJOR impact on their planned "growth" budget and power availability.
 
The various documentation on just how MUCH power folks wanted to add to Chelan's grid in just a one year period is an eye opener - and makes it understandable that the PUD decided they "needed to DO something about it".
We're talking requests totalling HUNDREDS OF MEGAWATTS in the appx. 1 year before they imposed a "moritorium" on new large projects to research the issue and figure out what to do about it.
 
I don't like the result they came up with, but I can understand WHY they did it (and what they don't realize is that it CAN be worked around easily by small mining operations, fairly easily by medium, and even the LARGE operations could work around it with some effort).
 
 




Hi QuintLeo, sorry, I know this is not related but I see that you are crunching on distributed.net, was wondering if you can help me out here. I've a few amd gpu on a mining rig all connected via risers, can't seem to get dnetc to run on Windows 10.
Do I require to at least plug 1 of the gpu into PCI-E slot as primary gpu while keeping the rest of the cards on riser in order to get the dnetc running?
#1
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/14 11:58:00 (permalink)
I don't use 10 at all any more - suffered through it on my main gaming rig for about 3 months before I ripped it back out and rolled back to 7 - and at this point I refuse to even look at 10 period.
I can't think of any reason why dnetc would require a GPU on the motherboard though - several of my mining rigs are all risers but run dnetc on the APU just fine.
Are you sure you're using the correct version (the very first entry, Win32/OpenCL, is the one you want for use with GPUs on Windows)?
 
 
edit title Cool GTX after split thread
 
post edited by Cool GTX - 2018/11/14 12:12:20

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#2
Cool GTX
EVGA Forum Moderator
  • Total Posts : 30983
  • Reward points : 0
  • Joined: 2010/12/12 14:22:25
  • Location: Folding for the Greater Good
  • Status: online
  • Ribbons : 122
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/14 12:11:42 (permalink) ☄ Helpfulby resync 2018/11/17 06:06:07
Split this in to your own thread ---> this is the Best Way to get the Help your after

Learn your way around the EVGA Forums, Rules & limits on new accounts Ultimate Self-Starter Thread For New Members

I am a Volunteer Moderator - not an EVGA employee

https://foldingathome.org -->become a citizen scientist and contribute your compute power to help fight global health threats

RTX Project EVGA X99 FTWK Nibbler EVGA X99 Classified EVGA 3080Ti FTW3 Ultra


#3
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/17 06:13:21 (permalink)
Hmm.. yes, I'm using the Win32/OpenCL version.
 
OS: Windows 10 64-bit (1709)
AMD graphics card driver: Win10-64bit-radeon-software-adrenalin-edition-18.6.1-june13
dnetc client: Windows 32bit [x86/OpenCL] v2.9112.521
 
Rig Setup:
Intel iGPU enabled as primary display
2 x AMD RX580, 1 x Vega56, 1 x R9 290 - connected via PCI-E 1x risers
This is a crypto mining setup and I wanted to switch over to dnetc to crunch some stats but getting the below errors when running the dnetc client on Windows 10. Seems like the crunching is only working for the Intel iGPU.
I wonder if this is OS issue and if the same setup would work on Windows 7
 
DNETC Log (don't mind the date/time, lazy to grab new logs, c/p from previous post on dnet fb page):
[Jul 19 13:10:33 UTC] size_t on device 0: 64 bits, host: 32
[Jul 19 13:10:33 UTC] size_t on device 1: 64 bits, host: 32
[Jul 19 13:10:33 UTC] size_t on device 2: 64 bits, host: 32
[Jul 19 13:10:33 UTC] size_t on device 3: 64 bits, host: 32
[Jul 19 13:10:33 UTC] size_t on device 4: 64 bits, host: 32
[Jul 19 13:10:33 UTC] Recovered 5 checkpoint packets
[Jul 19 13:10:33 UTC] Automatic processor detection found 5 processors.
[Jul 19 13:10:33 UTC] Loading crunchers with work...
[Jul 19 13:10:34 UTC] Automatic processor type detection did not
recognize the processor (tag: "Intel(R) HD Graphics 4600 ")
[Jul 19 13:10:34 UTC] RC5-72: Running micro-bench to select fastest core...
[Jul 19 13:10:55 UTC] RC5-72: using core #3 (CL 4-pipe).
[Jul 19 13:10:55 UTC] RC5-72 : Loaded D8:284A6B40:00000000:64*2^32 (63.80% done)
[Jul 19 13:10:55 UTC] RC5-72 : Loaded D8:2FD556CE:00000000:50*2^32 (26.00% done)
[Jul 19 13:10:55 UTC] RC5-72 : Loaded D8:2FE3F841:00000000:64*2^32
[Jul 19 13:10:55 UTC] RC5-72 : Loaded D8:2FE4D071:00000000:64*2^32
[Jul 19 13:10:55 UTC] RC5-72 : Loaded D8:2FE4FEC0:00000000:64*2^32
[Jul 19 13:10:55 UTC] RC5-72: 97 packets (6020.00 stats units) remain in
buff-in.r72
Projected ideal time to completion: 0.16:03:12.00
[Jul 19 13:10:55 UTC] RC5-72: 0 packets are in buff-out.r72
[Jul 19 13:10:55 UTC] 5 crunchers ('a'-'e') have been started.
.....10%.....20%..b..30%.....40%.....50%.....60%.a|
[Jul 19 13:10:55 UTC] Detected IP-link on 'eth0/lan0'...
[Jul 19 13:10:55 UTC] Error creating command queue on device 4
[Jul 19 13:10:55 UTC] Error code -6, message: Out of host memory
[Jul 19 13:10:55 UTC] Error creating command queue on device 2
[Jul 19 13:10:55 UTC] Error code -6, message: Out of host memory
[Jul 19 13:10:55 UTC] Error creating command queue on device 3
[Jul 19 13:10:55 UTC] Error code -6, message: Out of host memory
[Jul 19 13:10:55 UTC] Error creating command queue on device 1
[Jul 19 13:10:55 UTC] Error code -6, message: Out of host memory
 
post edited by resync - 2018/11/17 06:23:13
#4
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/18 03:48:35 (permalink)
It's picking up your on-CPU gpu and trying to use it but you apparently don't have openCL drivers installed for that - which throws everything else off.
If you can get OpenCL installed for that, it should start working correctly.
Workaround if you can't is to run one client per GPU pointing it at each specific GPU, running each copy of the client in it's own subdirectory.
 

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#5
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/20 06:57:58 (permalink)
Installed Intel HD4600 drivers, seems it's able to find the iGPU now but hangs at the stage below.
----------------------------------------------------------------------------------------------
dnetc v2.9112-521-CTR-16021317 for OpenCL on Win32 (WindowsNT 6.2).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at

[Nov 20 14:40:52 UTC] size_t on device 0: 64 bits, host: 32
[Nov 20 14:40:52 UTC] size_t on device 1: 64 bits, host: 32
[Nov 20 14:40:52 UTC] size_t on device 2: 64 bits, host: 32
[Nov 20 14:40:52 UTC] size_t on device 3: 64 bits, host: 32
[Nov 20 14:40:52 UTC] size_t on device 4: 64 bits, host: 32
[Nov 20 14:40:52 UTC] Automatic processor detection found 5 processors.
[Nov 20 14:40:52 UTC] Loading crunchers with work...
[Nov 20 14:40:52 UTC] Automatic processor type detection did not
                      recognize the processor (tag: "Ellesmere (Radeon RX 580 Series)")
----------------------------------------------------------------------------------------------
 
Running dnetc -gpuinfo, gives the following output:
Not sure what the "size_t on device 0: 64 bits, host: 32" and "Device address bits: 64 - NOT MATCHED -" means..
----------------------------------------------------------------------------------------------
dnetc v2.9112-521-CTR-16021317 for OpenCL on Win32 (WindowsNT 6.2).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at

[Nov 20 14:42:14 UTC] size_t on device 0: 64 bits, host: 32
[Nov 20 14:42:14 UTC] size_t on device 1: 64 bits, host: 32
[Nov 20 14:42:14 UTC] size_t on device 2: 64 bits, host: 32
[Nov 20 14:42:14 UTC] size_t on device 3: 64 bits, host: 32
[Nov 20 14:42:14 UTC] size_t on device 4: 64 bits, host: 32

Device #0:
------------
Automatic processor identification tag: 0
        name: Ellesmere (Radeon RX 580 Series)
Estimated processor clock speed (0 if unknown): 1335 MHz
Number of processors detected by this client: 5
Number of processors supported by this client: 512

Platform info:
--------------
                 Platform Name: AMD Accelerated Parallel Processing
               Platform Vendor: Advanced Micro Devices, Inc.
              Platform Version: OpenCL 2.1 AMD-APP (2580.6)
           Platform extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_am ...

Device info:
--------------
                          Type: GPU
                          Name: Ellesmere
                 Max clockrate: 1335
             Max compute units: 36
      Global memory cache size: 16384
      Global memory cache type: Read/Write
      Unified memory subsystem: No
                 Image support: Yes
             Local memory size: 32768
            Max workgroup size: 256
     native vector width (int): 1
   native vector width (float): 1
              OpenCL C version: OpenCL C 1.2
  Device timer resolution (ns): 1
                 Device vendor: Advanced Micro Devices, Inc.
              Device vendor id: 0x1002
                Driver version: 2580.6
           Device address bits: 64 - NOT MATCHED -

Device #1:
------------
Automatic processor identification tag: 0
        name: Hawaii (AMD Radeon R9 200 Series)
Estimated processor clock speed (0 if unknown): 1500 MHz
                          Type: GPU
                          Name: Hawaii
                 Max clockrate: 1500
             Max compute units: 40
      Global memory cache size: 16384
      Global memory cache type: Read/Write
      Unified memory subsystem: No
                 Image support: Yes
             Local memory size: 32768
            Max workgroup size: 256
     native vector width (int): 1
   native vector width (float): 1
              OpenCL C version: OpenCL C 1.2
  Device timer resolution (ns): 1
                 Device vendor: Advanced Micro Devices, Inc.
              Device vendor id: 0x1002
                Driver version: 2580.6
           Device address bits: 64 - NOT MATCHED -

Device #2:
------------
Automatic processor identification tag: 0
        name: Ellesmere (Radeon RX 580 Series)
Estimated processor clock speed (0 if unknown): 1335 MHz
                          Type: GPU
                          Name: Ellesmere
                 Max clockrate: 1335
             Max compute units: 36
      Global memory cache size: 16384
      Global memory cache type: Read/Write
      Unified memory subsystem: No
                 Image support: Yes
             Local memory size: 32768
            Max workgroup size: 256
     native vector width (int): 1
   native vector width (float): 1
              OpenCL C version: OpenCL C 1.2
  Device timer resolution (ns): 1
                 Device vendor: Advanced Micro Devices, Inc.
              Device vendor id: 0x1002
                Driver version: 2580.6
           Device address bits: 64 - NOT MATCHED -

Device #3:
------------
Automatic processor identification tag: 0
        name: gfx900 (Radeon RX Vega)
Estimated processor clock speed (0 if unknown): 1400 MHz
                          Type: GPU
                          Name: gfx900
                 Max clockrate: 1400
             Max compute units: 56
      Global memory cache size: 16384
      Global memory cache type: Read/Write
      Unified memory subsystem: No
                 Image support: Yes
             Local memory size: 32768
            Max workgroup size: 256
     native vector width (int): 1
   native vector width (float): 1
              OpenCL C version: OpenCL C 1.2
  Device timer resolution (ns): 1
                 Device vendor: Advanced Micro Devices, Inc.
              Device vendor id: 0x1002
                Driver version: 2580.6 (PAL,HSAIL)
           Device address bits: 64 - NOT MATCHED -

Device #4:
------------
Automatic processor identification tag: 0
        name: Intel(R) HD Graphics 4600
Estimated processor clock speed (0 if unknown): 1150 MHz

Platform info:
--------------
                 Platform Name: Intel(R) OpenCL
               Platform Vendor: Intel(R) Corporation
              Platform Version: OpenCL 1.2
           Platform extensions: cl_intel_dx9_media_sharing cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_ ...

Device info:
--------------
                          Type: GPU
                          Name: Intel(R) HD Graphics 4600
                 Max clockrate: 1150
             Max compute units: 20
      Global memory cache size: 262144
      Global memory cache type: Read/Write
      Unified memory subsystem: Yes
                 Image support: Yes
             Local memory size: 65536
            Max workgroup size: 512
     native vector width (int): 1
   native vector width (float): 1
              OpenCL C version: OpenCL C 1.2
  Device timer resolution (ns): 80
                 Device vendor: Intel(R) Corporation
              Device vendor id: 0x8086
                Driver version: 20.19.15.5058
           Device address bits: 64 - NOT MATCHED -

----------------------------------------------------------------------------------------------
 
 
Question:
What OS ver are your rigs with PCI risers on?
AMD Graphics driver version?
 
I found that by disabling the AMD OpenCLin registry, iGPU 4600 works fine being the only GPU detected.
Registry: HKLM\SOFTWARE\WOW6432Node\Khronos\OpenCL\Vendors
DWord Key: C:\WINDOWS\System32\DriverStore\FileRepository\c0329759.inf_amd64_379f06e83dab0e48\B329716\amdocl.dll
Value: Changed to 1 (0 to enable, 1 to disable)
Result:
----------------------------------------------------------------------------------------------
dnetc v2.9112-521-CTR-16021317 for OpenCL on Win32 (WindowsNT 6.2).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at

[Nov 20 14:59:33 UTC] size_t on device 0: 64 bits, host: 32
[Nov 20 14:59:33 UTC] Automatic processor detection found 1 processor.
[Nov 20 14:59:33 UTC] Automatic processor type detection did not
                      recognize the processor (tag: "Intel(R) HD Graphics 4600 ")
[Nov 20 14:59:33 UTC] RC5-72: Running micro-bench to select fastest core...
[Nov 20 14:59:54 UTC] RC5-72: using core #3 (CL 4-pipe).
[Nov 20 14:59:54 UTC] RC5-72: Loaded D8:284A6B40:00000000:64*2^32 (81.70% done)
[Nov 20 14:59:54 UTC] RC5-72: 101 packets (6262.00 stats units) remain in
                      buff-in.r72
                      Projected ideal time to completion: 3.08:00:52.00
[Nov 20 14:59:54 UTC] RC5-72: 0 packets are in buff-out.r72

----------------------------------------------------------------------------------------------
 
 
By disabling IntelOpenCL and leaving the above AMD key enabled, iGPU obviously not detected but it ain't helping the AMD cards either, hangs
Dword Key: IntelOpenCL32.dll
Value: Changed to 1 (0 to enable, 1 to disable)
Result:
----------------------------------------------------------------------------------------------
dnetc v2.9112-521-CTR-16021317 for OpenCL on Win32 (WindowsNT 6.2).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at

[Nov 20 15:01:22 UTC] size_t on device 0: 64 bits, host: 32
[Nov 20 15:01:22 UTC] size_t on device 1: 64 bits, host: 32
[Nov 20 15:01:22 UTC] size_t on device 2: 64 bits, host: 32
[Nov 20 15:01:22 UTC] size_t on device 3: 64 bits, host: 32
[Nov 20 15:01:22 UTC] Automatic processor detection found 4 processors.
[Nov 20 15:01:22 UTC] Loading crunchers with work...
[Nov 20 15:01:22 UTC] Automatic processor type detection did not
                      recognize the processor (tag: "Ellesmere (Radeon RX 580 Series)")

----------------------------------------------------------------------------------------------
 
Kinda lost now.. not sure if this is driver related.. should I install older AMD drivers for dnetc?
post edited by resync - 2018/11/20 07:09:03
#6
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/20 10:13:20 (permalink)
You can't go too old on the AMD drivers or you lose support for the Vega entirely.
I'm also not sure how far back support for your Hawaii card goes - they stopped support for at least a year possibly 2 after the 15.12 version before they added it back in again.
What driver version are you running?

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#7
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/20 16:56:13 (permalink)
OS: Windows 10 64-bit (1709)
AMD graphics card driver: Win10-64bit-radeon-software-adrenalin-edition-18.6.1-june13
dnetc client: Windows 32bit [x86/OpenCL] v2.9112.521
#8
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/21 12:48:14 (permalink)
I prefer 16.9.2 in general for a machine that is just crunching, but that version does not support Hawaii at all.
I'd suggest trying a version newer than your current one, in the hope that AMD actually fixed Hawaii support - I remember them adding it fairly recently but not which version.
A better overall idea would be to split the Hawaii (and your "older than RX and Vega" cards) into it's own machine, run LINUX and the 15.12 drivers on that machine.
 

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#9
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/21 22:59:16 (permalink)
I thinking it may not be hawaii issue cause by disabling the r9 290 the rest of the cards still can't start crunching w the dnetc client.

Sounds like an plan. I'd probably run nix on usb thumbdrive and see if the cards would work with dnetc client else remove r9 290 and try again.

Thanks :)
#10
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/11/22 00:39:52 (permalink)
Actually, I was getting Hawaii (R9-290/290x) and Tahiti (R9 280/280x) confused again.
It's Tahiti they took forever getting support for in the drivers, not Hawaii - Hawaii just took a while.
 
On the other hand, mixing generations does tend to cause issues - and adding that Intel into the mix just makes it worse.
 

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#11
resync
New Member
  • Total Posts : 7
  • Reward points : 0
  • Joined: 2018/11/14 04:44:38
  • Status: offline
  • Ribbons : 0
Re: Help can't seem to get dnetc to run on Windows 10 2018/12/05 01:45:59 (permalink)
Finally, dnetc recognizes all of the cards and crunches without any issues under Ubuntu.

Begone Win7/Win10, nix still rox.

Thanks :)
#12
QuintLeo
SSC Member
  • Total Posts : 946
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 3
Re: Help can't seem to get dnetc to run on Windows 10 2018/12/06 21:30:43 (permalink)
I was just thinking - and part of the issue might have been the long-running Windows problem with not handling more than 4 GPUs in a single system without extreme workarounds and lots of issues - and occasionally even with 3 GPUs in a system.
 
 

Now that vorsholk has stopped his abuse, I'm returning to folding.
 I no longer MOO due to abuses by certain "whales" in the Gridcoin community - so I now work the Distributed.net project directly again.
 
#13
Jump to:
  • Back to Mobile