All About BTC, LTC, ETH mining as well as other alternative crypto currencies
A quick update of the public ccMiner fork by SP of his Nvidia Maxwell optimized miner in the form of a Windows binary compiled from the latest Git source code of the ccMiner 1.5.79-git SP-MOD fork (source). The latest release comes with some small fixes and slight improvements in some algorithms. Have in mind that is a CUDA 7.5 compiled release, the latest CUDA 6.5 release of SP’s fork of ccMiner was version 74. Do note that the SP-MOD fork of ccMiner is designed for Nvidia Maxwell GPUs such as the already available GTX 750, 750 Ti as well as the newer GTX 960, GTX 970, GTX 980 and GTX 980 Ti and GTX Titan X. The Windows binary release we have made available here is 32-bit with support for Compute 5.0 and Compute 5.2 GPUs or with other words only for Maxwell-based Nvidia video cards compiled with CUDA 7.5 and VS2013.
Alexis Provos has updated his fork of ccMiner (source) with improved performance for mining Vcash and Decred Blake256-based crypto coins and we have compiled a Windows binary to try it out. His latest release brings the same performance or even a slightly higher for Vcash (former Vanillacoin) as the specially optimized CUDA 6.5 Compute 3.5 release of ccMiner using the latest CUDA 7.5 and the Compute version supported by the specific GPU you have. The Decred (DCR) hashrate is also higher when compared to the latest tpruvot release that includes the optimizations submitted by pallas. Below you can find a quick comparison of the hashrate you can expect from this release on GTX 980 Ti, the archive below includes a 32-bit and 64-bit binaries compiled with CUDA 7.5 and VS2013 for Compute 2.0 – Compute 5.2 GPUs, though we have not tested how it handles on older GPUs.
Do note that the focus of the ccMiner fork from Alexis Provos is on the Blake256 algorithms and most specifically on Vcash and his release brings support for very few algorithms. There are some optimizations available for the other supported algorithms as well, so you might want to compare them as well in terms of performance. It is not a full ccMiner fork with support for all of the algorithms available in other releases, so with this one you can mine only the following algorithms:
– blake for Blake256-14rounds (SFR)
– decred for Blake256-14rounds (DCR)
– blakecoin for Blake256-8rounds (BLC)
– vcash for Blake256-8rounds (XVC)
– whirlpoolx for WhirlpoolX (old VNL algo)
– keccak for keccak256 (Maxcoin)
– lyra2 for LyraBar
– lyra2v2 for VertCoin
Since this fork is not based on the latest code by tpruvot there might be some bugs that have been resolved in the newer codebase of tpruvot’s ccMiner. We have observed something strange already, namely the very fast increase of the difficulty to a very high level on the Suprnova pool for Decred using Stratum over getwork, though it works just fine on YIIMP. In the end the difficulty issue appeared to be temporary and apparently something poolside and not related to the miner itself. Below you can find some performance comparison results using an Nvidia GTX 980 Ti GPU:
This release with GTX 980 Ti intensity 31 Decred, 30 Vcash
Vcash – 4453 MHS
Decred – 2394 MHS
1.7.4 alexis78 release Compute 3.5 CUDA 6.5
Vcash – 4418 MHS
Latest 1.7.6-git tpruvot with Pallas tweaks
Decred – 2218 MHS
Another update with a new Windows binary compiled from the latest ccMiner 1.7.6-git fork from tpruvot (source) with fixes that make the Decred (DCR) performance improvements by pallas work properly under Windows as well as Linux. The release below includes two versions of ccMiner, one is 32-bit binary for Windows that is compiled with CUDA 7.5 and with support for Compute 2.0 or newer Nvidia GPUs, and another that is 64-bit Windows binary compiled with CUDA 7.5 for Compute 3.5 or newer Nvidia GPUs. Do note that if you have a newer Nvidia video card you should use the 64-bit CUDA 7.5 release as it is providing slightly better hashrate than the 32-bit CUDA 7.5 binary.