Release Notes

• Add docker build and run
• Log difficulty of solutions
• Help corrections

Release Notes:

• Dropped AMD binary kernel support. No suitable kernels.
• Dropped effective hash rate calculator. Not working in stratum v2.
• Added API Prometheus http metrics support.
• Fix API HTTP response RFC deviations.
• Minor bug fixes.

Release Notes

• CRITICAL: fix Windows build error
• Add -F option to load parameters from a file instead of on command line

Release Notes

• add timestamp to console log
• Fix simulation and benchmark for low epochs
• add --cl-split option to force split DAG mode for AMD GPUs. May improve performance on older GPUs.
• display GPU memory temperature if available. Also add to API.

Release Notes

• Avoid splitting DAG unnecessarily.
• Fix missed connection delay when --retry-max is 0
• Minor bug fixes (display formatting)
• Add --seq option to sequentially load DAG one GPU at a time.

Release Notes

• Fix unknown console log entry in Windows
• Update Windows CUDA toolkit to cuda_11.2.2_461.33

Release Notes:

• Lower compiler optimizations to safer level
• Effective hash rate (option -v4) calculations working for stratum2.
• Fix nvml.dll search for NV GPUs.
• Use single DAG instead of split DAG on AMD drivers that allow greater than 4GB allocation.
• Handle f2pool out-of-spec. messages.

Release Notes:

• Use board name rather than architecture for AMD

Release Notes:

• Fix skipping of 1st pool when connecting with multiple pools specified.
• Fix appearance of 'ghost' GPU metrics on epoch change.
• Remove specific path for nvml.dll. It needs to be included in the library search path.
• Prevent Windows from sleeping while mining.

Release Notes:

• Allocate per GPU light cache to avoid potential GPU crash on rapid epoch transitions in stratum2.
• More details on GPU memory requirements.
• fix simulation and benchmark

Release notes:

• Detect potential NVIDIA memory allocation failure that happens if too much memory is used by the Windows desktop.

Release Notes

• Fix potential buffer overun in Nvidia GPU code.
• Rename --farm-recheck option to --getwork-recheck. It applies only to getwork protocol.
• Rename --farm-retries option to --retry-max. Applies to both getwork and stratum protocols.
• Set --retry-max to 0 for unlimited retries.
• Abbreviate PCI addresses in API responses to provide compatibility with Ethminer and mining OS integration.

Release Notes:

• close potential Nvidia race window.

Release Notes:

• Fix missing telemetry (fan temperature & power)
• Restore PCI ordered GPUs and device selection (--devices)

Release Notes:

• Hung GPU detection with optional exit or reboot action
• Multi-line status update console display (--multi)
• Standardize OpenCL and Nvidia console output
• Simplify OpenCL object management
• Reintroduce binary kernel support for AMD GPUs. No binaries yet.
• Fix paused GPU crash.
• Update various documentation items.
• Upgrade to G++-17 compile.
• Disallow unsecure SSL protocols. only TLS 1.2 is supported for SSL.
• Support multi PCI domain platforms.
• Remove locale dependencies.
• Add ability to specify miner base nonce via parameter (--nonce) or API.
• Dropped Unbuntu 16.04 release builds.
• Miscellaneous code improvements.

Nicehash mining support

Update CUDA to version 11.2 in Windows build Added OpenCl enumeration diagnostics, for undetected GPUs Miscellaneous code improvements

Upgraded all release builds to CUDA 11.2 Upgrade opengl bindings to 2.2 Miscellaneous code improvements

Add Ubuntu 16.04 build to release

Minor cleanup and addition of Cuda 10 version for Linux

Automatically tune Nvidia grid size, and AMD group multiplier. Reinstate Nvidia block size and stream parameter. Reinstate AMD group size parameter. Optimized job switch time vs. performance balance.