I'm happy to open source my solvers in exchange for a 20% community contribution to the Cuckoo Cycle Bounty Fund, which will offer up to 100 BTC in bounties for disproving my claims of the Cuckoo Cycle proof-of-work being optimization free. See
for details on the existing bounties. The remaining 80BTC, if claimed, will be at my personal expense. The fund allows for significantly raising the reward for a performance doubling from the current $500 to a much more enticing 10BTC, and will scale with the 2-logarithm of the actual improvement, so that a 1024x speedup (or memory reduction) can claim the maximum 10*10BTC payout.
The bounties would be guaranteed for at least 4 years, plenty of time to encourage some serious attacks on Cuckoo Cycle.
While you're all no doubt keen on supporting advances in crypto-currency research, you may also be curious to see some well optimized Equihash solvers, to which end I offer the following in return.
On a 4Ghz Intel Core i7, one of my CPU solvers reaches 2 Sol/s single-threaded, 4.3 Sol/s 4-threaded, and 6.7 Sol/s 8-threaded, while another solver reaches 5.14 Sol/GBs single threaded for peak memory.
On an Nvidia GTX980, my CUDA solver reaches 17.6 Sol/s.
Open sourcing these solvers will take place if donations reach 20 BTC by noon Oct 14 (with a possible one day extension if the goal looks within reach at that time). If the goal is not met then donations will be returned (unless you're happy to unconditionally support Cuckoo Cycle bounties).
There are multiple ways to make donations. The escrow options are detailed in cryptogoth's companion post at
while those who trust me with their donations can contribute directly to
the Cuckoo Cycle Bounty Fund bitcoin address