Skip to content

DynexSolve Version 2.2.5
Compare
Choose a tag to compare
@dynexcoin dynexcoin released this 07 Mar 08:54
· 77 commits to main since this release
DynexSolve-v225
c3677e1
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

DynexSolve PoUW CPU & GPU Miner

  • Mining Software Supporting DynexSolve algorithm
  • CUDA (Nvidia GPUs) (AMD and Intel are being supported by commercial miners)
  • Windows, Linux & Hive
  • Requires libcurl (Ubuntu: sudo apt install libcurl-dev)
  • Require open ports 80 (http) and FTP
  • Requires 1 CPU core for each GPU, preferably newer versions with AES support

Hive installation path: https://jobs.dynexcoin.org/dynexsolve-2.2.5hive.tar.gz

An overview of pools supporting DS225 is available here: https://dynexcoin.org/mining-pool-certifications-status/

Release notes (Version 2.2.5):

  • Fix: Last GPU with lower chips
  • MALLOB: INVALID TIMESTAMP improvement
  • Kernel check and hang GPU hang detection
  • Stability improvements

Release notes (Version 2.2.4):

  • continuous power draw (no more power cycles)
  • distributed hashing over multiple CPU threads (pool hash-rate matching reported hash-rate)
  • performance increase over previous versions (HR)
  • significantly reduced rejected shares, more stability
  • improved ODE integration scheme, now showing how close to the solution reward the miner is (LOCAL MINIMA = 0 = solution reward)
  • new command line parameter "-intensity" (default: 8192) which sets the number of CUDA blocks
  • fixed download of computation file CNF
  • Windows logging colors
  • streamlined logging
  • general stability fixes and code cleanup
  • miner stops after 20 consecutive rejected shares
  • significantly improved hardening against tampering or malicious usage
  • AMD Support: will be offered by commercial miners
  • Slow CPU support: this version will be as fast as it gets as for CPU hashing, the only way to support slow CPUs further is to use GPU for hashing, too (but it will need some of your memory). This will be offered by the commercial miners
  • Further hash-rate improvements will potentially be seen by the commercial miners, too

Release notes (Version 2.2.3):

  • Up to 15% more hashrate
  • Supporting the new FastMallob (decentralized) mallob system
  • parameter -adj for each GPU
  • Security improvements

Release notes (Version 2.2.2e):

  • Fix: some rejected shares not showing in console
  • New command line option "-stratum-interval" (default value: 100): determines number of hashes until Stratum refreshes; higher value = higher hashing speed, but also less frequent get_job updates
  • Command line option "-adj" new minimum of 0.8; default set to 1.3
  • Chip memory requirement optimisation, fitting more chips - increased HR

HiveOS only:

  • parameter "-adj" per GPU
  • new *.cnf problem file loader
  • memory optimisations

Release notes (Version 2.2.2d):

  • Improved hashing function
  • Stale shares (err#499) fix

Release notes (Version 2.2.2c):

  • Improved Stratum protocol handling (<1ms)

Release notes (Version 2.2.2b):

  • Improved mallob connectivity
  • Eliminated first share reject
  • Eliminated duplicate share reject
  • computation file: removed FTP entirely and enabled automatic download
  • HR fixes when disabling GPUs
  • Graceful exit fixes

Release notes (Version 2.2.2):

  • New GPU embedded double proof-of-useful-work (PoUW) scheme
  • Local RPC Service for pool operators for faster and more efficient share verification, can be clustered
  • Option for pool operators to run their own dedicated Mallob servers for workers
  • Added multiple additional tampering and security measurements
  • Mallob resource usage significantly lowered and network id improvements
  • Improved hashrate on low end cpu (~15%)
  • Hashrate on pool now exactly matches reported
  • Updated json stats to contain GPU hashrates
  • Added stratum reconnect on errors
  • Added Mallob error reporting
  • Added sync option for CUDA blocking sync (reduce cpu usage) (new command line option)
  • Added adjust option for memory usage - more Chips possible (default 1.5, new command line option)
  • Added option not to create GPU_x.bin file (new command line option)
  • Update capacity request moved to before filling GPU to reduce startup time on Mallob errors
  • Exit on atomic expiring or 20 failed attempts
  • Improved logging, stability and overall code cleanup
Assets 5
Loading