ChangeLog

xptMiner changes:

b16 (2014-09-01)
	- Fixed critical bug that was breaking b15 on windows.
	Bug does not seem to manifest on Linux, but I'm pushing
	b16 just in case.

b15 (2014-04-26)
	- Major internal architectural overhaul.  Sieving and
	primality testing are now divided among all threads
	instead of having each do a single operation.

	The current consequence of this is a good speedup on
	modest-core architectures while using substantially less
	memory.  4-16 core machines should be particularly happy with
	this upgrade.

	Sieves can now be up to -s 4100000000 (4 billion) in size, though
	this does not appear to be a particularly useful setting from
	a performance perspective.

	Single-core machines may suffer a 5-10% slowdown.  If this is
	prohibitive, let me know, but for now I plan to let it stay
	that way.

	Very large, slow core machines (e.g., 64 core AMD) are running
	MUCH slower.  Please either continue to use b14 or run multiple
	copies of the miner, one per physical CPU, using taskset.

	Windows users must use at least Vista (2006, NT 6.0) or later.  XP and
	Windows Server 2003 are no longer supported.

b14 (2014-03-29)
	- Providing binaries linked against gmp6.  Tested and
	  validated.  Decent speedup for newer Intel CPUs.

	- Optimization for sieving to large numbers (-s #).
	  Works properly up to around 2^31-1, but to be safe, use at
          most 2.1 billion (-s 2100000000).  This optimization uses an
	  idea from a00k's miner - thank you!  If the offset of
	  a prime in the sieve is greater than the maximum amount the
	  sieve can be extended within the Riecoin nonce, it is discarded;
	  this helps save some space with large sieves and lets the miner
	  be more efficient when extending.

	- Note that super-large sieve values are not necessarily the
	  best.  Benchmark and evaluate for your own CPU.  In general,
	  higher difficulty calls for a larger -s, but doing so
	  increases the amount of memory used per thread and
	  increases the think-time at the start of every block.
	  A slower CPU or a lower-memory CPU wants a smaller -s
          value.  Some good values to try:
	     Small CPU:  200000000
             Default:    500000000
             Good CPU:   900000000
             Fast CPU:  1500000000
	
b13 (2014-03-23)
	- Calculate instead of guess the maxiter (it was set too high)
	  based upon the riecoin nonce limit, the primorial, and the
	  size of the sieve.
	- Switch target recomputation to every 3 minutes.
	
	- Minor code optimizations for initial sieve offset computation,
	  which now occupies a bit more CPU time than it did before because
	  of b12 changes to recompute target.

	- Continued internal code cleanup and clarification.
	
b12:
	- Re-generates target and nonce time every 60 seconds.
	  This may be too aggressive - I'm going to back it out a bit
	  in b13 and see if people continue to have the invalid share
	  problem.

b11:
	- Import reconnect-hang fix from clintar/xptMiner.

b10:
	- MAJOR CHANGE:  Update to the definition of a "share"
	  from a 4-chain of primes (first four all prime) to
	  "any four".  This change results in substantially more
	  shares per hour, but addresses fairness problems resulting
	  from targeting 4-chains vs 6-chains.

	  A consequence of this change is that 2ch/3ch/4ch numbers
	  are all increased and should not be compared vs previous 
	  versions.  No substantive changes to actual search process,
	  so blocks per hour should remain the same.