Topic: [ANN] ccminer 2.3 - opensource - GPL (tpruvot) - page 23. (Read 500257 times)

Sure, but some of the yiimp clone pools are using different exchanges and their estimated as well as actual BTC per 24 is different. If you could figure it out you could figure out which results are erroneous. Conversely if you could see estimated and actual coins per 24 (instead of BTC) that would be more helpful for figuring that out yourself. One more step, but BTC per 24 seems to be a legacy stat back from when yaamp was auto-exchange (almost none of the pools do this).

jr325: same as before


the main differences in the graphes is not related to the price but to the coin diff... Else only big ones in general

If you only have CUDA 9.1 installed, is there even a 'lib'?  Can you tell me the resulting value for LD_LIBRARY_PATH in your case?
Thanks

How do you figure out what exchange yiimp pools are using for their estimated 24 hour actual rates?

Yep, from my tests here, the first linux drivers compatible with cuda 9.1 have a very strange behavior.. Unsure on windows yet

I reported the issue to nvidia dev team

I started getting this error on one of my rigs everytime I run ccminer 2.2.2

cuda error in function "scanhash_lyra2v2" at line 144 : out of memory

That's the last message I get before windows shuts it down.

I'm not sure why this is happening as it was working fine last week.

Apparently something in my system changed and I'm just trying to figure out what would cause this.

Obviously it seem like an "out of memory" issue lol but what memory is it referring to? and how could I fix that?

Any help with this would be greatly appreciated...

If I cant figure it out soon it's system reset time lol :-)

I got CUDA 9.1 to work with the cuda-9 branch of ccminer.  I compiled CUDA 9.1 using the .run file and had it install the drivers included with the full .run package as the proprietary drivers did not install correctly using either the GUI (system setting route) or the PPA from the command line (at least for me, though I'm not an expert so I could very easily have been doing something wrong).  As I posted above thread, the key for me was to ensure that I had both lib and lib64 referenced in my LD_LIBRARY path before I compiled, otherwise, even though ccminer would compile without failing, it wouldn't do so correctly.

I have compiled the project, but, when run, I receive this error message%




and ccminers shut downs


can you help me?

Use the cuda-9 branch

is yiimp down? seems like a bit target for always getting attacked

It seems building the latest ccminer source against CUDA 9.1 fails, because the file device_functions_decls.h cannot be found.

Edit: On ubuntu 16.04. Of course an easy workaround is to use CUDA 9.0 libraries for now.

Someone left you Chinese christmas cheer at your pool Epsylon.

why the connection on yiimp go down every two three days? someone is ddossing the pool or what?

Several questions and things to try (though I'm on 16.04 not 17.04).  Do the CUDA example libraries compile and run for you (don't just try devicequery, also try some of the others like randomfog). If not, likely you have a problem with the install of either CUDA or the driver.  I had to work hard to get the right driver and the right libraries for the examples to compile.

Separately, i had exactly the same error and the way I eventually got it fixed was that I had to add lib and lib64 to my LD_LIBRARY path.  None of the documentation mentioned lib64, just lib.  Give that a try, but make sure you recompile ccminer first as if you didn't include lib64 in the LD_LIBRARY path it likely didn't compile correctly (I learned that one from bitter experience).  Good luck.

Anyone else having problems building and running cuda-9 branch with Ubuntu 17.04 and CUDA 9.1?  It compiles successfully but crashes within few seconds after running. I am not overclocking and have run these 2 cards extensively with ccminer with windows.  I've also tried equihash and that crashes as well.  Any ideas?

/ccminer$ ./ccminer  -r 0 -a Lyra2v2 -o stratum+tcp://hub.miningpoolhub.com:20593 -u xyz -p x 
*** ccminer 2.2.4 for nVidia GPUs by tpruvot@github ***
    Built with the nVidia CUDA Toolkit 9.1 64-bits

  Originally based on Christian Buchner and Christian H. project
  Include some kernels from alexis78, djm34, djEzo, tsiv and krnlx.

BTC donation address: 1AJdfCpLWPNoAMDfHF1wD5y8VgKSSTHxPo (tpruvot)

[2017-12-24 01:40:42] Starting on stratum+tcp://hub.miningpoolhub.com:20593
[2017-12-24 01:40:42] NVML GPU monitoring enabled.
[2017-12-24 01:40:42] 2 miner threads started, using 'lyra2v2' algorithm.
[2017-12-24 01:40:42] Stratum difficulty set to 256 (1.00000)
[2017-12-24 01:40:42] GPU #0: Intensity set to 20, 1048576 cuda threads
[2017-12-24 01:40:42] GPU #1: Intensity set to 20, 1048576 cuda threads
[2017-12-24 01:40:49] Stratum difficulty set to 16 (0.06250)
Cuda error in func 'scanhash_lyra2v2' at line 118 : unknown error.
Cuda error in func 'scanhash_lyra2v2' at line 118 : unknown error.
Segmentation fault (core dumped)


ccminer$ nvidia-smi
Sun Dec 24 15:19:44 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 387.34                 Driver Version: 387.34                    |
|-------------------------------+----------------------+----------------------+

Parameter passing from Awesome Miner to ccminer-tpruvot possibly?

Try ccminer alone with full parameters and see if works correctly on it's own. That should determine where the issue is - and ccminer-tpruvot being what it is, I highly doubt it's the miner. It's probably the parameter passing from one application to the miner.

#crysx

Started having this issue where I fail to connect a lot. Different pools, different alogs. I'm pretty sure its not the pool.

I'm using Awesomeminer with ccminer 2.2.3

Any ideas?

https://github.com/tpruvot/ccminer/pull/34/files

hello

why do I have this error?

Wow, Debian 9 is proving to be a real headache. Maybe its time to stop using Debian.

Anyway, is support for cuda 9.1 at least functional if not complete? Apparently CUDA9 doesn't have some of the header files that it expects when compiled.

Installing Cuda 8 is complicated by the fact that Debian 9 ships GCC 6 instead GCC 5, again breaking compatibility.

Maybe I should just reinstall with something other than Debian.