Creating one big buffer for the DAG Allocating/mapping single buffer failed with: clCreateBuffer(-61). GPU can't allocate the DAG in a single chunk. Bailing. clEnqueueWriteBuffer(-38)

[danderson198]

Receiving this error. Wont mine.

[danderson198]

C:\Users\dande_000\Desktop\ethereum-mining-windows\ethminer-opencl-0.9.41>setx G
PU_MAX_ALLOC_PERCENT 95

SUCCESS: Specified value was saved.

C:\Users\dande_000\Desktop\ethereum-mining-windows\ethminer-opencl-0.9.41>setx G
PU_USE_SYNC_OBJECTS 1

SUCCESS: Specified value was saved.

C:\Users\dande_000\Desktop\ethereum-mining-windows\ethminer-opencl-0.9.41>setx C
L_DEVICE_MAX_MEM_ALLOC_SIZE 2147483648

SUCCESS: Specified value was saved.

C:\Users\dande_000\Desktop\ethereum-mining-windows\ethminer-opencl-0.9.41>ethmin
er.exe -G -F http://eth-us.suprnova.cc:3000/danderson198.1/2 -t 1
Found suitable OpenCL device [Cayman] with 2147483648 bytes of GPU memory
miner 20:24:51|main Getting work package...
miner 20:24:51|main Grabbing DAG for #526ee6e6…
i 20:24:52| Loading full DAG of seedhash: #3db89d40…
miner 20:24:52|main Got work package:
miner 20:24:52|main Header-hash: 0a07b510faebd9f72c7e079ebaa83bc85d552bfea98
41b917df4f01720d12678
miner 20:24:52|main Seedhash: 526ee6e688a9b3d466fe4a7355197a89898593509d75c5
70942cc73a5ca8299d
miner 20:24:52|main Target: 00000023ca98ce50e9b5ca0a331813ab66d9e2177017787a
f44b4f4215dd7041
i 20:24:52|gpuminer0 workLoop 0 #00000000… #526ee6e6…
i 20:24:52|gpuminer0 Initialising miner...
miner 20:24:53|main Mining on PoWhash #0a07b510… : 0 H/s = 0 hashes / 0.5 s
miner 20:24:53|main Got work package:
Using platform: AMD Accelerated Parallel Processing
i 20:24:53| Full DAG loaded
Using device: Cayman(OpenCL 1.2 AMD-APP (1642.5))
miner 20:24:53|main Header-hash: e2f0ec89552088f7beb8296b1597873f189321d0fea
5ae75c241a68e6b6ff94a
miner 20:24:53|main Seedhash: 526ee6e688a9b3d466fe4a7355197a89898593509d75c5
70942cc73a5ca8299d
miner 20:24:53|main Target: 00000023ca98ce50e9b5ca0a331813ab66d9e2177017787a
f44b4f4215dd7041
Printing program log

Creating one big buffer for the DAG
Allocating/mapping single buffer failed with: clCreateBuffer(-61). GPU can't all
ocate the DAG in a single chunk. Bailing.
clEnqueueWriteBuffer(-38)

[LefterisJP]

Your GPU does not have enough memory to mine.

[dagelf]

I had the same error even on cards that do have enough memory, when mining on multiple cards. Mining both I got the above error, but only one at a time all worked fine. Mining both on geth worked fine, and ethminer on Windows worked fine.. just ethminer on Ubuntu that failed as per above.

Ubuntu 14.04.1
AMDAPPSDK-2.9-1
fglrx-updates- 2:15.201-0ubuntu0.14.04.1

*Fixed with: apt-get install fglrx *

It seems fglrx-updates gets the two cards mixed up.

[bobsummerwill]

Thanks for the update, @dagelf! Re-opening the issue.

[dagelf]

I can provide access to the rig if needed, or if I can see how it's fixed and what is looked for I might be drawn into ethereum dev... Seeing as that 14.04 will be supported through 2019, it might be worth figuring it out or at least documenting workarounds.

[gregbkr]

Hello, I got the same DAG error on windows8.1. After lot of investigation and reboot, I think the problem comes from graphic card driver & soft (for me catalys amd 6900 series).

I can mine with default driver, but very low perf (I think) =5MH

ethminer --list-devices 
CL_DEVICE_GLOBAL_MEM_SIZE: 2 147 483 648
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 1 877 999 616  --> 1.9GB (DAG ok)

With catalyst drivers, seems like whatever option I setx (+ some reboot), it is stuck on 500MB, so I can't create DAG. I understand somehow catalyst soft hardfixe that value.

ethminer --list-devices 
CL_DEVICE_GLOBAL_MEM_SIZE: 2 147 483 648
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 536 870 912 -->500MB (DAG FAILS)

I tried to install lastest amd drivers only without catalyst soft, but couldn't get the card detected, or just got the value hard stuck.

I will be glad to get any suggestion....

Thank you!

[bobsummerwill]

See also ethereum/libethereum#203

[tomafc330]

@dagelf I'm getting the same issue on 15.10. I tried you suggestion but I'm still getting an error. Did you have to do anything after apt-get install fglrx?

[dagelf]

It's very driver specific - my guess is that the default 15.10 fglrx driver
is too new. See if you can force the exact driver that 14.04 uses, or
download and install it manually from AMD's site or a cache (not sure if
they serve old versions)

Sent from my phone - +27737721223
On 26 Mar 2016 10:53 AM, "Tommy Chan" notifications@github.com wrote:

@dagelf https://github.com/dagelf I'm getting the same issue on 15.10.
I tried you suggestion but I'm still getting an error. Did you have to do
anything after apt-get install fglrx?

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#2988 (comment)

[DrengorDraxlon]

I fixed this problem with a different environement variable.

GPU_SINGLE_ALLOC_PERCENT

I've got mine set to 90.

[novice987]

I am getting same error "GPU can't allocate the DAG in a single chunk. Bailing. clEnqueueWriteBuffer(-38)" with the following device list.

ethminer --list-devices

Listing OpenCL devices.
FORMAT: [deviceID] deviceName
[0] Intel(R) HD Graphics 5600
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 5109950055
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 1277487513
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
[1] GeForce GTX 960M
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 2147483648
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 536870912
CL_DEVICE_MAX_WORK_GROUP_SIZE: 1024

How can I fix it?

[l29ah]

Got this on HD7870 w/ 2GB RAM.
[OPENCL]:Using device: AMD PITCAIRN (DRM 2.43.0 / 4.6.3-grsec+, LLVM 3.8.1)(OpenCL 1.1 Mesa 12.1.0-devel (git-cfbf956))
CL_DEVICE_TYPE: GPU
CL_DEVICE_GLOBAL_MEM_SIZE: 2147483648
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 1503238553
Nothing else is occupying the GPU. I wonder what's the problem here.

[l29ah]

Huh?

[Qalander]

I have NVIDIA 720M GTx and i tried to mine Ether, but got error ''GPU can’t allocate the DAG in a single chunk'' what should I do here.

[chfast]

How much VRAM?

[Qalander]

I set it at 12 GB min and max at 32 GB but after researching found that I need latest drivers and vram should be 16 GB so I have to set it like that let me check if it works.

[shadowvex]

Setting GPU_SINGLE_ALLOC_PERCENT to 90 worked for me too.

[killerpollo236]

sgminer 5.5.5-gm - Started: [2019-07-26 23:37:08] - [0 days 00:00:08] -------------------------------------------------------------------------------- (5s):0.000 (avg):0.000h/s | A:0 R:0 HW:0 WU:0.000/m ST: 2 SS: 0 NB: 2 LW: 7 GF: 0 RF: 0 Connected to eu1.ethermine.org (stratum) diff 4G as user 0x36ee666186474615778f82e3798a688dd56bf8d6 Block: 9ce38d2b... Diff:0 Started: [23:37:16] Best share: 0 -------------------------------------------------------------------------------- [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit GPU 0: 44.0C 0RPM | OFF / 0.000h/s | R: 0.0% HW:0 WU:0.000/m xI:1024 -------------------------------------------------------------------------------- [23:37:06] Started sgminer 5.5.5-gm [23:37:06] * using Jansson 2.7 [23:37:06] WARNING: GPU_USE_SYNC_OBJECTS is not specified! [23:37:06] Probing for an alive pool [23:37:07] Startup GPU initialization... Using settings from pool eu1.ethermine.org. [23:37:07] Startup Pool No = 0 [23:37:07] Initialising kernel ethash.cl with nfactor 10, n 1024 [23:37:13] Error -61: Creating the DAG buffer failed. [23:37:13] Error: clSetKernelArg of all params failed. [23:37:13] GPU 0 failure, disabling!

[killerpollo236]

Mine is not working. I have the virtual memory config 3000-30000. What could it be?

[killerpollo236]

It is a rx580 radeon amd and even cudominer is not working with the gpu but it detects it.