GPU #0: FATAL: failed to allocate device memory for long state

[ghost]

i am running a evga GTX 750 Ti i keep getting this error, i think it has something to do with the .bat file but not sure, i think i have to set or change the blocks and threads but not sure, can anyone help me. tsiv-xmr [2017-11-11 19:33:35] Using JSON-RPC 2.0 [2017-11-11 19:33:35] 1 miner threads started, using 'cryptonight' algorithm. [2017-11-11 19:33:35] Starting Stratum on stratum+tcp://xmr-us-east1.nanopool.org:14444 [2017-11-11 19:33:35] GPU #0: GeForce GTX 750 Ti (5 SMX), using 40 blocks of 8 threads [2017-11-11 19:33:35] GPU #0: FATAL: failed to allocate device memory for long state

[mountassir]

Same issue here with GTX1070.

[auroracoin]

@darrellschwass -l 8x30 is what you should set your parameters for the 750ti. I get 280 h/s. But this error I don't know about. Sorry. I'm actually using a different app. but will be trying tsiv app soon. :-)

[ghost]

does this look correct, this is my bat file. ccminer -q -o stratum+tcp:// xmr-us-west1.nanopool.org:14444 -u 46PAEU91ZYbDPVeemUtAk35uUuhqprfD5fKmg9c3V92LMW9gGQNtNEoZ6cthDGCNybDKBC6eS7aUQYNiWePtLjJ2MSVxtKn.YOUR_PAYMENT_ID.GTX750ti/YOUR_EMAIL -p x --launch 8x30

On Sun, Nov 12, 2017 at 8:15 AM, tbear notifications@github.com wrote:

@darrellschwass https://github.com/darrellschwass -l 8x30 is what you should set your parameters for the 750ti. I get 280 h/s. But this error I don't know about. Sorry.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/tsiv/ccminer-cryptonight/issues/31#issuecomment-343740244, or mute the thread https://github.com/notifications/unsubscribe-auth/AgB_ViPoIhbHg6yUK7U5A5KjI1GEX4i_ks5s1v2SgaJpZM4QatSk .

[will2022]

I also have that issue, i'm mining on Nvidia GTX 1080 8 GB Strix. I think cpu mining is the alternate solution on my Ryzen 7

[dstana]

Same issue GTX 1060.

[Honowski]

I've got the same error from both ccminer-cryptonight and tsiv-xmr. I've got GTX1060s

[ghost]

I am also getting this issue when using the updated 388.13 Drivers for NVidia, but I was able to resolve it by downgrading drivers to 372.90, although I am not sure what is causing the probelm.

[Dutchy-79]

Same problem here [2017-12-08 09:51:15] Using JSON-RPC 2.0 [2017-12-08 09:51:15] 1 miner threads started, using 'cryptonight' algorithm. [2017-12-08 09:51:15] Starting Stratum on stratum+tcp://monerohash.com:3333 [2017-12-08 09:51:15] GPU #0: GeForce GTX 960 (8 SMX), using 40 blocks of 8 threads [2017-12-08 09:51:15] GPU #0: FATAL: failed to allocate device memory for long state

[Azanigin]

1080ti same problem :(

[NoSubstitute]

1070 same problem, drivers v388.59

[Jertheunited]

Same with GTX1070ti v388.59

[ghost]

if anyone finds the solution for this please post it here or post a link, thank you

On Tue, Dec 12, 2017 at 12:34 PM, Jertheunited notifications@github.com wrote:

Same with 388.59 [image: image] https://user-images.githubusercontent.com/34487454/33901848-58316f2c-df30-11e7-83da-07a2bfea78b9.png

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/tsiv/ccminer-cryptonight/issues/31#issuecomment-351145054, or mute the thread https://github.com/notifications/unsubscribe-auth/AgB_VsxMkuDhbl7K2ENv_ZqXcxx4dZKsks5s_sccgaJpZM4QatSk .

[Honowski]

I found the solution. Download https://github.com/fireice-uk/xmr-stak. Works on windows and Linux on both AMD and Nvidia cards. It also utilizes your CPU. I haven't had any issues with it.

[yonicsurny]

Same issue :-( [2017-12-21 01:02:30] Using JSON-RPC 2.0 [2017-12-21 01:02:30] 1 miner threads started, using 'cryptonight' algorithm. [2017-12-21 01:02:30] Starting Stratum on stratum+tcp://xmr.pool.minergate.com:45560 [2017-12-21 01:02:30] GPU #0: GeForce GTX 970 (13 SMX), using 40 blocks of 8 threads [2017-12-21 01:02:30] GPU #0: Warning: block count 40 is not a multiple of SMX count 13. [2017-12-21 01:02:31] Pool set diff to 1063 [2017-12-21 01:02:31] Stratum detected new block [2017-12-21 01:02:34] GPU #0: FATAL: failed to allocate device memory for long state

[tousifhasanzai]

downgrade your cuda from 9 to 8

[Kastor83]

Same issue on 1070Ti and Titan V with current drivers.

Can you explain 'downgrade cuda from 9 to 8' please?

[SilverCory]

@Kastor83 https://developer.nvidia.com/cuda-80-ga2-download-archive I'm just doing this I'll edit this comment if it works.

[MADeveloper]

Had this error on Win10, GTX 1060 with Cuda 9. I installed Cuda 8 and it corrected it. The miner is now working

[SilverCory]

I discovered this maintained fork. https://github.com/KlausT/ccminer-cryptonight

It has cuda 9 support. I had trouble installing cuda 8.

[ArtemkaKun]

Hi gues! If You have the Fatal Error, do next steps:

1. Install https://www.microsoft.com/pl-pl/download/details.aspx?id=48145
2. Install https://www.visualstudio.com/downloads/ (CUDA NEED IT!!! WITHOUT VS 2013(or better) CUDA DOESN'T INSTALL!!!)
3. Install https://developer.nvidia.com/cuda-80-ga2-download-archive
4. If You have troubles with step 3 - download latest drivers for your card and install it with CLEAR INSTALLATION
5. Reboot your PC

6. PROFIT!111!

[Kahless77]

Im downloaded and installed CUDA8, and this ccminer its just not working, same error message before CUDA8, any trick?

[ArtemkaKun]

@Kahless77 , try to install CUDA 9

[ash-predictiva]

Same Issue, tried all suggestions, nothing!

Installed VS Installed CUDA 8 Installed CUDA 9 Updated Drivers

Nothing works!!

[ash-predictiva]

Here's the solution after long trials:

1- Download VS 2- Download CUDA 9 3- Download this version of CCminer (ccminer-x64-2.2.5) from here https://github.com/tpruvot/ccminer/releases

4- Format your patch as follows

ccminer-x64.exe -a cryptonight -o YOUR POOL ADDRESS -u YOUR WALLET ADDRESS -p x

[auroracoin]

If your running windows, get the latest updates. May help.

[milnomada]

Hi,

I'm using:

• cuda-6.5
• nvidia geforce GS 9300M
• compute_30 settings

The following command works:

./ccminer -a monero -o stratum+tcp://xmr-eu1.nanopool.org:14444 -u x -p {addr} -l 4x20 --bfactor=2

Having a look at the settings:

-l, --launch=CONFIG   launch config for the Cryptonight kernel.
                        a comma separated list of values in form of
                        AxB where A is the number of threads to run in
                        each thread block and B is the number of thread
                        blocks to launch. If less values than devices in use
                        are provided, the last value will be used for
                        the remaining devices. If you don't need to vary the
                        value between devices, you can just enter a single value
                        and it will be used for all devices. (default: 8x40)
--bfactor=X    Enables running the Cryptonight kernel in smaller pieces.
                        The kernel will be run in 2^X parts according to bfactor,

to tweak these values to smaller numbers to fit your card features.

[2021-05-25 17:32:16] GPU #0: GeForce 9300M GS, 93253.05 H/s