WildRig Multi 0.47.9 multi-algo miner with low dev-fee (AMD/NVIDIA/INTEL)

[buraz11]

is it possible to get bc-250 supported?( gfx1013)

[fenomenyaa]

Hello,

G3930Tv8gb ram,240gb ssd ,page files 160gb Win 10 Cpu usuage %100 off all my rigs,

Cpu usuage %100 off all my rigs,amd rigs works but all nvidia rigs have low hashrate or no hashrate, i have all gens of gpus, 3080,4060ti,5070ti

[andrucrypt]

WildRig Multi 0.43.7
- improved qhash on different gpu's
- reduced CPU usage for qhash

[fenomenyaa]

WildRig Multi 0.43.7
- improved qhash on different gpu's
- reduced CPU usage for qhash

Good job ,cpus can breath now.,Thx.

[andrucrypt]

WildRig Multi 0.43.8
- improved qhash for AMD, Intel and NVIDA(4000/5000 series only)
- improved auto-tune for qhash and added parameter --qhash-blocks

NOTE:
- on weak gpu's auto-tuning may take few minutes
- CPU load can be still high for NVIDIA gpu's

[fenomenyaa]

WildRig Multi 0.43.8
- improved qhash for AMD, Intel and NVIDA(4000/5000 series only)
- improved auto-tune for qhash and added parameter --qhash-blocks

NOTE:
- on weak gpu's auto-tuning may take few minutes
- CPU load can be still high for NVIDIA gpu's

Hello,agaib qhash high cpu usage and worse perf. on windows 10 nvidia gpus..

[andrucrypt]

WildRig Multi 0.43.8
- improved qhash for AMD, Intel and NVIDA(4000/5000 series only)
- improved auto-tune for qhash and added parameter --qhash-blocks

NOTE:
- on weak gpu's auto-tuning may take few minutes
- CPU load can be still high for NVIDIA gpu's

Hello,agaib qhash high cpu usage and worse perf. on windows 10 nvidia gpus..

Unfortunately nothing can be done with current implementation, maybe in 0.44.0 I will fix the issue

[andrucrypt]

WildRig Multi 0.44.0
- improved qhash for AMD RDNA1, RDNA2 and RDNA3 architectures
- significantly improved qhash for NVIDIA Ampere and older architectures
- fixed low hashrate on CMP 170HX for qhash(use parameter --qhash-kernel 2)
- up to 20% improvements for NVIDIA Ada Lovelace and Blackwell
- CPU usage decreased again

[Nigh7man]

WildRig Multi 0.44.0
- improved qhash for AMD RDNA1, RDNA2 and RDNA3 architectures
- significantly improved qhash for NVIDIA Ampere and older architectures
- fixed low hashrate on CMP 170HX for qhash(use parameter --qhash-kernel 2)
- up to 20% improvements for NVIDIA Ada Lovelace and Blackwell
- CPU usage decreased again

problem on windows... Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG

[andrucrypt]

WildRig Multi 0.44.0
- improved qhash for AMD RDNA1, RDNA2 and RDNA3 architectures
- significantly improved qhash for NVIDIA Ampere and older architectures
- fixed low hashrate on CMP 170HX for qhash(use parameter --qhash-kernel 2)
- up to 20% improvements for NVIDIA Ada Lovelace and Blackwell
- CPU usage decreased again

problem on windows... Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG

this error means not enough memory to initialize DAG. Also it can be low Virtual Memory on Windows, so increase it.

[Nigh7man]

WildRig Multi 0.44.0
- improved qhash for AMD RDNA1, RDNA2 and RDNA3 architectures
- significantly improved qhash for NVIDIA Ampere and older architectures
- fixed low hashrate on CMP 170HX for qhash(use parameter --qhash-kernel 2)
- up to 20% improvements for NVIDIA Ada Lovelace and Blackwell
- CPU usage decreased again

problem on windows... Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG

this error means not enough memory to initialize DAG. Also it can be low Virtual Memory on Windows, so increase it.

i try increase .. no effect (

[21:30:27] kernel 1 compiled for sm86
[21:30:27] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG
[21:30:29] kernel 1 compiled for sm86
[21:30:30] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG
[21:30:32] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:33] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:35] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:36] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:37] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'

[andrucrypt]

WildRig Multi 0.44.0
- improved qhash for AMD RDNA1, RDNA2 and RDNA3 architectures
- significantly improved qhash for NVIDIA Ampere and older architectures
- fixed low hashrate on CMP 170HX for qhash(use parameter --qhash-kernel 2)
- up to 20% improvements for NVIDIA Ada Lovelace and Blackwell
- CPU usage decreased again

problem on windows... Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG

this error means not enough memory to initialize DAG. Also it can be low Virtual Memory on Windows, so increase it.

i try increase .. no effect (

[21:30:27] kernel 1 compiled for sm86
[21:30:27] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG
[21:30:29] kernel 1 compiled for sm86
[21:30:30] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when initialize DAG
[21:30:32] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:33] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:35] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:36] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'
[21:30:37] Error CL_MEM_OBJECT_ALLOCATION_FAILURE when calling clEnqueueNDRangeKernel for kernel 'search'

what coin is that? And exact cards

[entryboy]

Hello andrucrypt,

are you planning to optimize qhash for graphics cards with GCN 3.0?
In each of the latest versions, the hashrate climbs on my 7800XT (currently 220 MH/s @ 275W v.0.46.4).
With the Fiji's (R9 Fury), it stagnates at 22 MH/s.

Thank you for your service.

Best regards from germany.

[andrucrypt]

Hello andrucrypt,

are you planning to optimize qhash for graphics cards with GCN 3.0?
In each of the latest versions, the hashrate climbs on my 7800XT (currently 220 MH/s @ 275W v.0.46.4).
With the Fiji's (R9 Fury), it stagnates at 22 MH/s.

Thank you for your service.

Best regards from germany.

hi, better math for qhash didn't suite well for older cards in previous versions. But try 0.46.5, it should improve significantly them.

[fenomenyaa]

Hello,
On Windows 10 LTSC 22H2 with either a G3930 or G4560 CPU, 8GB RAM, and a 160 MB page file, using the qhash algorithm, we’ve tested multiple driver versions (currently running the latest).

With the NVIDIA Blackwell series GPUs, rigs running 10–12 GPUs show high CPU usage and sometimes no mining at all, or performance drops to about one-quarter of expected hashrate.

In contrast, NVIDIA 3000 and 4000 series rigs run perfectly fine under the same conditions. The issue also doesn’t occur on Blackwell rigs with up to 8 GPUs.

Other miners, such as Rigel and Lolminer, do not show this problem, so it appears to be specific to this miner. We also tested with a stronger CPU (i7-2600K),also 16gb ram tested, but the issue persists, so it doesn’t seem related to CPU performance.

<a href="https://ibb.co/9mQJCLxn"><img src="https://i.ibb.co/9mQJCLxn/miner.jpg" alt="miner" border="0"></a>

[andrucrypt]

Hello,
On Windows 10 LTSC 22H2 with either a G3930 or G4560 CPU, 8GB RAM, and a 160 MB page file, using the qhash algorithm, we’ve tested multiple driver versions (currently running the latest).

With the NVIDIA Blackwell series GPUs, rigs running 10–12 GPUs show high CPU usage and sometimes no mining at all, or performance drops to about one-quarter of expected hashrate.

In contrast, NVIDIA 3000 and 4000 series rigs run perfectly fine under the same conditions. The issue also doesn’t occur on Blackwell rigs with up to 8 GPUs.

Other miners, such as Rigel and Lolminer, do not show this problem, so it appears to be specific to this miner. We also tested with a stronger CPU (i7-2600K),also 16gb ram tested, but the issue persists, so it doesn’t seem related to CPU performance.

<a href="https://ibb.co/9mQJCLxn"><img src="https://i.ibb.co/9mQJCLxn/miner.jpg" alt="miner" border="0"></a>

try set intensity higher like -i 27, but that stale shares looks strange.

[fenomenyaa]

Hello,
On Windows 10 LTSC 22H2 with either a G3930 or G4560 CPU, 8GB RAM, and a 160 MB page file, using the qhash algorithm, we’ve tested multiple driver versions (currently running the latest).

With the NVIDIA Blackwell series GPUs, rigs running 10–12 GPUs show high CPU usage and sometimes no mining at all, or performance drops to about one-quarter of expected hashrate.

In contrast, NVIDIA 3000 and 4000 series rigs run perfectly fine under the same conditions. The issue also doesn’t occur on Blackwell rigs with up to 8 GPUs.

Other miners, such as Rigel and Lolminer, do not show this problem, so it appears to be specific to this miner. We also tested with a stronger CPU (i7-2600K),also 16gb ram tested, but the issue persists, so it doesn’t seem related to CPU performance.

<a href="https://ibb.co/9mQJCLxn"><img src="https://i.ibb.co/9mQJCLxn/miner.jpg" alt="miner" border="0"></a>

try set intensity higher like -i 27, but that stale shares looks strange.

It seems to Higher intensity fix the problem need to wait 24 hours more to check the stability, Maybe better to set default  i 27 to Blackwell series

[GW_ICO]

Hi,

Starting from version 46.7 I am getting these errors:

Error UNKNOWN_ERROR when calling cudaMemcpy to fill input buffer on GPU #1
[00:46:02] GPU #0: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:03] GPU #7: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:04] GPU #2: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:05] GPU #6: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:06] GPU #7: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:07] GPU #3: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:08] Error UNKNOWN_ERROR when calling cudaMemcpy to fill output buffer on GPU #1
[00:46:09] GPU #1: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:10] GPU #5: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:11] GPU #0: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'

Right now it is just cycling between these and similar errors and never starts mining.

On 46.6 everything works fine.

I was using the latest Hive OS, then I tried with 0.6-222@230511 combined with 525.116.04 drivers, but still no Luck.

The rig is having 8x3060Ti paired with Pentium G5420 and 16G of ram, H310M mobo. Hive is installd on 32Gb usb drive, but I also tried on 128GB ssd, no any luck as well.
Also I can't set intensity - whatever I put in the extra arguments on the flightsheet It runs with intensity 23. Please advise the correct syntax to adjust the intensity.

[andrucrypt]

Hi,

Starting from version 46.7 I am getting these errors:

Error UNKNOWN_ERROR when calling cudaMemcpy to fill input buffer on GPU #1
[00:46:02] GPU #0: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:03] GPU #7: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:04] GPU #2: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:05] GPU #6: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:06] GPU #7: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:07] GPU #3: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:08] Error UNKNOWN_ERROR when calling cudaMemcpy to fill output buffer on GPU #1
[00:46:09] GPU #1: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:10] GPU #5: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'
[00:46:11] GPU #0: Error UNKNOWN_ERROR when calling cudaLaunchKernel for kernel 'qhash'

Right now it is just cycling between these and similar errors and never starts mining.

On 46.6 everything works fine.

I was using the latest Hive OS, then I tried with 0.6-222@230511 combined with 525.116.04 drivers, but still no Luck.

The rig is having 8x3060Ti paired with Pentium G5420 and 16G of ram, H310M mobo. Hive is installd on 32Gb usb drive, but I also tried on 128GB ssd, no any luck as well.
Also I can't set intensity - whatever I put in the extra arguments on the flightsheet It runs with intensity 23. Please advise the correct syntax to adjust the intensity.

Reduce your OC, right now algorithm is pretty core intensive. Also update drivers to at least 535(they are minimum required for RTX 40x0, not sure about 30x0)

[GW_ICO]

I removed all OC, still the same.
I even tried to downclock the cards, no luck.
Right now I am using the Hive OS 0.6-229@251119 with 550.142 drivers.
To me it looks like 3060TI cards are unsupported after v 0.46.6 as there is no any load, the cars are stable, i just get a bunch of error loops, but the machine is stable, I can chnage the flightsheet without any screen freeze or other instability.
Did anyone with 3060TI mine successfully using newer version than 0.46.6?