v1.6.1

Recompiled using cuda 9.2 - drivers will need to updated to latest versions to use.

x32 no longer built as nVidia are removing support in the near future.

v1.5.1

Add --logfile option based on KlausT

v1.5.0

Before the imminent release of cuda 9.2.

• More speed squeezed out (slightly more than 1%) - now >5% faster than nevermore imo.
• Hopefully fixed the newlines in the windows batch files
• Extended the benchmark-hash command to include _80 variants (I'm probably the only one who cares about this).

v1.4.1

People seem to prefer > .0 releases :-)

Fix for intensity setting on multi-gpu rigs and other small tweaks including some compiler options on Windows.

Probably a good idea to upgrade.

v1.4.0

Now with added pectin.......

Adds a bit more performance and displays a stable hashrate which is faster and uses less memory if the api is disabled (also corrects a potential issue with keccak80).

New options:

• "--benchmark-hash ": lets you benchmark specific algorithms
• "--no-simple-hr": if you disable the api via "-b 0" then the miner will calculate the hashrate based on a rolling average which means you don't have store a large number of samples to stabilse the displayed hashrate. This option disables that behaviour.
• "--autotune": enables automatic tweaking of the grids built for each algorithm based on the underlying hardware. This is very much in ALPHA state (here be dragons) but the framework is in place. Once again - this IS NOT for production use.

The code base has also undergone major changes - all of the algorithms now sit in individual files nad some of the function names have been changed.

Enjoy.

v1.3.1

Hygiene release.

Added scripts for each of the releases to start the miner ;
Put in a proper fix for stabilising the display of the hash rate on the machine (may take > 100 submissions before it is really noticeable) ;
Tweaked the heuristic for setting the default intensity ;

v1.3.0

Known internally as the 'Byte Magazine/Lemon Song' release.

4% improvement for Groestl ; 8% improvement for SIMD ; 14% for JH ; 23% for Blake ; 12% for BMW ; 2% for Skein.

Overall >3% faster than Nevermore 0.2.2 & >2.5% faster than Suprminer 1.6.

Only cards > sm_50 (i.e. 750ti and newer) supported.

1% dev fee incorporated in binary builds but can be disabled by passing '--donate=0' on the command line.

v1.2.0

More cleaning up - no longer supports < sm_50.

poolparty v1.1.0

Immediate upgrade recommended.

Fixes issue with hamsi80 share submission.
Small speedup in keccak.
Turn off hash display in quiet mode.

poolparty v1.0.0

Initial release.

Faster than nevermore 0.2.0.

Merges the nevermore/suprminer code with a version of ccminer which has had (all/most) of the non-raven code stripped out and various local optimisations.

There is a 1% dev donation fee coded in for continued development but this can be disabled by passing '--donate 0' on the command line (or you can increase it through the same parameter if you want :-) )

Instructions for running in docker can be found in DOCKER (linux-only)