Version 1.1.4:  (Posted January 28, 2019)
  Fixed an issue where algegraic factors of the form (k^x)*b^n+/-1 were
  not removed where n%x = 0.
  Output "CPU processor time" when sieving is done.
  
Version 1.1.3:  (Posted August 27, 2018)
  Fixed an issue where the algebraic factor found is the same as the term
  which would remove all terms before sieving.
  
Version 1.1.1:  (Posted January 9, 2018)
  Fixed an issue with ABCD files when there are no terms for a sequence.

Version 1.1.0:  (Posted January 20, 2017)
  Rewrote code to find algebraic factorizations so that more can be caught.
  It will search for:
     GFNs -> where k*b^n+1 can be written as x^m+1
     Trivial -> where k*b^n-1 can be written as x^m-1
  which will remove all terms for the sequence.

  These algebraic factorizations are now written to algebraic.out so that
  they can be verified with pfgw:
     where k*b^n+1 can be written as x^q*y^r+1 and r%q=0 and q is odd
     where k*b^n-1 can be written as x^q*y^r-1 and r%q=0
     where k*b^n+1 can be written as x*2^m+1 and m%4=2
     where k*b^n+1 can be written as 4*x^z*y^m+1 and z%4=0 and m%4=0
     
Version 1.0.8:  (Posted November 1, 2013)
  Fixed build issues from 1.0.7.

Version 1.0.7:  (Posted April 27, 2013)
  In srfile, added -B to convert from one base to another.  Both the input
  and output bases must be integer powers of an integer.
  In srfile, added -E to extract one or or more comma-delimited k from the
  input file.

Version 1.0.6:  (Posted December 23, 2012)
  Now builds cleanly with both MinGW and *nix systems.
  
Version 1.0.5:  (Posted September 12, 2012)
  Rebuilt Win64 with latest MinGW.  Previous builds could report invalid
  factors (which are rejected by the code) due to a miscompile by MinGW.

Version 1.0.4:  (Posted April 9, 2012)
  Fixed console output issue for algebraic factorizations.  The remaining
  candidates are correct, but the console output will be incorrect if the
  power of the algebraic factors is > 1.

Version 1.0.3:  (Posted February 10, 2012)
  Missed more algebraic factorizations.  For example, if k=x^6 and
  c=+1, then n where n%3=0 can be factored algebraicly.  Modified
  the code to iterate from 2 to x-1 so that it looks for all
  algebraic factors for all numbers that evenly divide x.

Version 1.0.2:  (Posted February 2, 2012)
  Missed some obvious algebraic factorizations.  For example if
  c=-1 and k=x^6 for some integer x, then we know that when n%6=0 that
  k*b^n-1 is divisible by x*b^(n/6)-1.  What we also know is that 
  k=(x^2)^3, thus we can also conclude that k*b^n-1 is divisible by
  (x^2)*b^(n/2)-1.

Version 1.0.1:  (Posted January 10, 2012)
  Fixed issue from 1.0.0 where it was removing wrong n based upon
  algebraic factorizations.

Version 1.0.0:  (Posted November 22, 2011)
  0.9.0 was built on 0.6.13 (with 0.6.17 enhancements), which caused
  srfile to error valid factors.  1.0.0 re-implemented the changes
  made to 0.9.0 on top of clean 0.6.17 code.
  srfile now supports GFN factor removal (ABC $a^$b+1 format).

Version 0.9.0:  (Posted July 22, 2011)
  srsieve will remove numbers with algebraic factorizations before sieving.

Version 0.6.17:
-d switch for srfile now accepts a filename argument. This feature added by
 Mark Rodenkirch.

Version 0.6.16: (Posted 31 March 2009)
Fixed a bug that resulted in memory being freed twice when a sequence is
 removed from the sieve. Thanks to Jayson King for reporting this bug and
 locating the problem in the source.

Version 0.6.15: (Posted 8 March 2009)
Fixed a modular inverse bug that can cause the sieve to stop with an error
 when a candidate factor p is a divisor of k or c in k*b^n+c. Many thanks to
 Jayson King who found the bug and provided a source code patch.
The above bug results partly from a change made in version 0.6.4 to use the
 64-bit arithmetic routines factors below 2^32 in the 64-bit build. This
 change also had other much more serious effects for the 64-bit build, and
 has now been undone. See BUG64.txt for details.
Check factors by default, use --no-check to turn off checking.

Version 0.6.14: (Posted 17 February 2009)
Replaced all variable length automatic arrays with malloc()'d arrays.
Stop with an error if p_max is above the sieve limit.
For x86_64 Windows the FPU must be changed into extended precision mode when
 compiling with USE_FPU_MULMOD=1.
Added WARNING.txt for the srsieve-x86_64-windows.exe executable.

Version 0.6.13: (Posted 7 December 2008)
Fixed a buffer overflow that could occur when printing messages with long
 file names. Thanks Chuck Lasher for reporting this bug.
In priority.c, use PRIO_MAX=10 if not defined in sys/resource.h
Changed the way ABC file headers are written from $a*2^$b+$c to $a*2^$b$c.
 Both types of header will be accepted as input.

Version 0.6.12:
Added x86-64 powmod from sr5sieve 1.7.10.
Added x86-64 memset_fast32 from sr5sieve 1.7.10.
Allowed for _WIN64 sizeof(uint_fast32_t)==4 in clear_hashtable().
Added k8-gcc430 target to avoid possible bugs in GCC 4.3.0.
Built srsieve-0.6.12-windows64-k8.zip for testing using mingw64-gcc 4.3.0.

Version 0.6.11:
Use `addc' instead of `adde' instruction in ppc64 mulmod.

Version 0.6.10: (Posted 27 December 2007)
Just set thread affinity, not process affinity, for Windows.
Allow X > 2^32 for the minimum reported factor switch `-m --minimum X'.

Version 0.6.9: (Posted 7 December 2007)
Added `-A --affinity N' switch to set affinity to CPU N.

Version 0.6.8: (Posted 3 December 2007)
Fixed another typo in srfile.c.
Added `-z --idle' and `-Z --no-idle' switches. Idle priority is default.

Version 0.6.7: (Posted 21 October 2007)
Testing reveals that the x86-64 SSE2 mulmod can fail with modulus between
 2^51 and 2^52. Reduced limit for this arch to 2^51.
`Lavalamp' found a serious bug in the filtering code that can cause some
 factors for sequences with both odd and even exponents to be missed when
 the subsequence base b^Q has an odd Q. Affected versions 0.5.0 to 0.6.6.

Version 0.6.6: (Posted 13 March 2007)
The sieve limit for ppc64 is 2^63 when using assembler mulmods, not 2^52.
Added x86-64 assembler powmod (powmod-k8.S)
Fixed a bug in write_newpgen_files() that caused the wrong number of terms
 to be reported when writing multiple files.
Added -Wall to default CPPFLAGS. Thanks jasong.

Version 0.6.5: (Posted 31 January 2007)
Fixed typo in srsieve.c, thanks michaf.
srfile -c now takes an argument: `srfile -c X infile.txt' will now report
 the number of terms in each congruence class modulo X (i.e. the number of
 terms in each subsequence when sieving in base b^X). Thanks masser for
 the suggestion.

Version 0.6.4: (Posted 14 January 2007)
Fixed a serious and longstanding bug in remove_eliminated_subsequences()
 that corrupted the sequence list when all subsequences for more than one
 sequence were removed from the sieve during the same event. Thanks Mark for
 reporting this.
Added Mark Rodenkirch's improved ppc64 assembler mulmod() function.
Use the 64-bit sieve for primes below 2^32 on 64-bit machines.
Added x86-64 assembler mulmod64()/sqrmod64() functions.
Added build for linux64-k8 binary.

Version 0.6.3:
Fixed a bug in mod64_init/fini() that may have prevented the FPU control
 word being correctly restored on i386.

Version 0.6.2:
Removed some core64() debugging stuff that should only print when DEBUG=yes. 

Version 0.6.1: (Posted 30 November 2006)
Fixed trial factoring in core64() which caused extremely long init times for
 sequences with large prime k. Thanks Citrix for reporting this bug.
Save and restore r2 register in expmod-ppc64.S, fixes segfault on PPC/Linux.
 Thanks Ed for the report and Mark for the fix.
Fixed mpz_get/set_uint64() in srtest to work on 32-bit MSB systems.
The srtest program is now run as a part of 'make check'.
Fixed a typo in arithmetic.h which was preventing the sqrmod code being used.
Test sqrmod64() in srtest too.
Made assembler pre2_* functions into PRE2_* macros. Benefits GCC 4.1.
Better FPU instruction scheduling in i386 assembler mulmod64()/powmod64().
Test for __ppc64__ or __powerpc64__ to recognise PowerPC 64.

Version 0.6.0: (Posted 22 November 2006)
New i386 assembler mulmod64() and powmod64() functions give correct results
 for all primes up to 2^62. Testing against GMP revealed that all earlier
 versions could give occasional incorrect results for primes as low as 2^46.
Added sqrmod64() function for use by generic powmod64().

Version 0.5.3:
Exchange %ebx,%ebp in powmod64().
Set CONST_EMPTY_SLOT in hashtable.h to use a constant value to mark empty
 hashtable slots. This requires one extra 16-bit register/constant comparison
 in lookup(), but avoids a 64-bit register/memory comparison when the slot is
 in fact empty.

Version 0.5.2: (Posted 31 October 2006)
Removed redundant call to fill_bits() in prime_sieve().
Replaced memset32_8() in hashtable.h with memset_fast32() in memset_fast32.h.
Assembler powmod64() for i586/i686.

Version 0.5.1: (Posted 22 September 2006)
Added a seperate code path for the 32 bit sieve when all sequences k*b^n+c
 have k,|c| < 2^16.
Pass a pair (mod,map) instead of a filter function to the prime sieve, check
 test_bit(map,p%mod) instead of calling filter(p).
Reduced default hashtable maximum density from 0.8 to 0.65.

Version 0.5.0:
For each subsequence k*b^n+c, check whether -ck (or -bck if n is odd) is a
 quadratic residue with respect to p before adding it to the BSGS list.

Version 0.4.8: (Posted 12 September 2006)
Allocate ck64,bd64 arrays on the stack instead of the heap.

Version 0.4.7: (Posted 11 September 2006)
Use uniformly sized bitmaps for each subsequence so .m_low and .m_high
 fields are no longer needed.
Added umod(a,p), and add an extra branch to lmod(a,p) to try to avoid a
 division if possible.

Version 0.4.6: (Posted 6 September 2006)
Increased minimum short/long hashtable size to 2^10/2^9.
New invmod() functions taken from Jason Papadopoulos's Msieve 1.10.
Removed unused submod() functions.

Version 0.4.5: (Posted 1 September 2006)
Relaxed asm constraints on i386 memset32_8(). Added x86-64 assember version.
A small improvement to powmod() saves one mulmod per call.
Removed unused sieve_high variable.
Added function invmod32_64(a,p) used instead of invmod64(a,p) when a < 2^32.

Version 0.4.4: (Posted 27 August 2006)
Unrolled clear_hashtable() loop, optimising seperately for 32 or 64 bit
 machines, with inline assembler version for i386.
Don't compile pentium4 or athlon binaries, i686 is just as good now that
 base-specific code is no longer used.

Version 0.4.3:
Removed the base-specific mulmods as these hardly ever get used except in
 bsgs_small() when first starting a new sieve. Removed the setup() functions
 optimised for the Q=1 case for the same reason.
Limit --nmax to 2^32-2, not 2^32-1.

Version 0.4.2: (Posted 12 August 2006)
Added quadratic residue filters for sequences where each term is of the form
 r*A^2+B^2 or r*A^2-B^2 with r <= 6.
Don't employ per-subsequence filtering unless at least 20% of subsequences
 have filters installed.
Fixed a problem introduced in 0.4.1 that could prevent the special form of
 some sequences being recognised if Q was not a power of 2.
Don't cast from uint32_t to void * on 64 bit machines. (srfile.c).
Adjusted the formula for choosing b^Q to favour slightly lower Q.

Version 0.4.1:
Changed the way that Q is chosen for sieving in base b^Q. This should result
 in better choices eventually, but needs tuning.
`srsieve --force-base X' forces b^X to be the sieving base.

Version 0.4.0:
Extended k,c in k*b^n+c to 64 bit variables: 0 < k < 2^64, 0 < |c| < 2^63.
 This results in a small slowdown for the 32 bit sieve, but a seperate code
 path means no slowdown for the 64 bit sieve if all k,|c| values are < 2^32.
Changed pre2_mulmod64_init() to make use of stack created by mod64_init().
Removed asm-i386-msc.h because I can't test it and it is doubtful whether
 the rest of the program compiles with MSC anyway.
Avoid use of ceil(x) where the integer part of x might be large (> 2^31?)
 because of a bug in GCC 4.1.0. [ceil(15^8) returns rubbish in GCC 4.1.0].
 (A work-around would be to add -fno-builtin-ceil to CFLAGS).

Version 0.3.18: (Posted 4 August 2006)
Fixed a sign bug reported by `Citrix' that caused bad results for sequences
 k*b^n+c with 2^31 <= k < 2^32 when sieving with 0 < p < 2^32.

Version 0.3.17: (Posted 3 August 2006)
Removed the 64-bit mode K8 assembler mulmod64() because it is slower than
 the generic C version. Thanks to the mersenneforum.org members who tested
 this for me.

Version 0.3.16: (Posted 1 August 2006)
Added the --mod=M,a,b,... command line switch to install a global filter for
 primes of the form a,b,... (mod M). e.g. `--mod 10,-1,1' will only test
 potential factors that are 1 (mod 10) or -1 (mod 10).
Added per-subsequence prime filters. Srsieve will now filter for special
 form factors even if not all subsequences have the same special form. e.g.
 if terms of one sequence are of the form A^(2^x)+B^(2^x) and another is of
 the form C^(2^y)+D^(2^y) then the first sequence will be left out of the
 sieve when p=K*2^x+1 and the second will be left out when p=K*2^y+1.

Version 0.3.15:
Added ARCH=k8 entry in Makefile for Athlon64 in native 64 bit mode.
Added assembler mulmod32/64() functions for x86-64. These are untested, and
 may be slower than the C versions. Try compiling with ARCH=k8 ASSEMBLER=no
 and then with ARCH=k8 ASSEMBLER=yes to see which are fastest.

Version 0.3.14: (Posted 28 July 2006)
Split sr5sieve into a seperate package sr5sieve-1.0.0.tar.gz

Version 0.3.13:
Extended the generalised Fermat filter to allow sequences of the form
 a^(2^y)+b^(2^y) instead of just x^(2^y)+1.
Fixed a bug introduced in 0.3.12 that prevented some sequences with square
 bases being properly recognised as having the form x^(2^y)+1.
Added a switch `srsieve --no-filter' to disable filtering of the sieve
 primes (for debugging only).
Check that k*b^n+c satisfies (b,c)=1.
Throw an error if an invalid sequence is input instead of just deleting the
 sequence with a warning.

Version 0.3.12:
Sequences in which every term is of the form x^(2^y)+1 can only have factors
 of the form z*2^(y+1)+1. Srsieve will now recognise these sequences and
 filter candidates from the prime sieve appropriately. Thanks to `Citrix'
 for this idea. Currently all sequences in the sieve must have factors of
 the same form for this to work -- removing this limitation is a TODO item.

Version 0.3.11: (Posted 23 July 2006)
`srfile --known-factors FILE' will now also accept GMP-ECM factors in FILE.
 Thanks `konrad127123' for the suggestion. Large Prime95 and GMP-ECM factors
 are accepted without verification.

Version 0.3.10:
Fixed a bug introduced in version 0.3.0 that could sometimes cause an extra
 term to be added to the end of a sequence when the sieve file was written.
Fixed init_prime_sieve() to handle very small sieve limits (p < 9).
Ensure small_sieve_limit is not less than the greatest prime factor of |c|.

Version 0.3.9: (Posted 20 July 2006)
`srfile --pfactor X' sets the last field of the worktodo.ini entry to X (the
 number of PRP tests saved by finding a factor). In previous versions this
 was incorrectly set to 0. Thanks to `michaf' for noticing this.

Version 0.3.8: (Posted 18 July 2006)
`srfile --known-factors results.txt' will now also accept Prime95 factors in
 results.txt for removal from the sieve. The factors are not checked.
Fixed a bug introduced in version 0.3.0 that sometimes caused srsieve to
 report the wrong number of terms when writing NewPGen format files.

Version 0.3.7: (Posted 17 July 2006)
Fixed a bug introduced in version 0.3.0 that sometimes caused terms k*b^n+c
 for all n>0 to be lost when the term k*b^0+c was prime and in the sieve.

Version 0.3.6: (Posted 16 July 2006)
sr5sieve now writes factors for the range <pmin>,<pmax> to factors<pmin>.txt
 instead of srfactors.txt.
Fixed for_each_term() so that terms are always processed strictly in order
 of increasing n, then increasing k, then increasing c. This showed up where
 --prp would sometimes print terms with equal n in the wrong order.

Version 0.3.5: (Posted 13 July 2006)
sr5sieve --newpgen K N0 N1 writes a NewPGen format file K.txt for PRP/LLR
 testing K*5^n+/-1 in the range N0 <= n < N1.
Fixed raising of base after small sieve.

Version 0.3.4: (Posted 8 July 2006)
Fixed typo: pre2_mod64_fini --> pre2_mulmod64_fini in asm-ppc64.h.
Removed nested function in srfile.c.

Version 0.3.3:
Use pre2_mulmod64 in bsgs setup64, although the gain is very small.
Don't raise the sequence base after the small sieve finishes except for new
 sieves. Don't raise it after the 32bit sieve at all.
By default srsieve will make one pass at splitting base b sequences into
 base b^Q subsequences. If --limit-base X is supplied with a value larger
 than the resulting Q, then it will continue splitting the subsequences into
 higher bases until either the base reaches b^X or the base doesn't increase
 from one pass to the next. I don't know if this is useful or not.

Version 0.3.2: (Posted 7 July 2006)
Removed a redundant trailing mulmod from the giant steps loop.
sr5sieve --delete K deletes the sequence K*5^n+/-1 from sr5data.txt.

Version 0.3.1: (Posted 6 July 2006)
Compute -c/k (mod p) once per base b sequence instead of once per base b^Q
 subsequence. This makes it efficient to use a larger Q, and so the default
 choice of maximum Q should now be close to optimal.
Fixed a dangling pointer left when every term of a sequence was eliminated.
Fixed wrong estimate of completion date in sr5sieve.

Version 0.3.0: (Posted 5 July 2006)
Split sequences in base b into subsequences in base b^Q. Q can be limited to
 a divisor of X with --limit-base X. I don't know how to choose the optimal
 Q, so experimentation is needed. The wrong choice for Q may result in a
 large performance loss.

Version 0.2.4: (Posted 5 July 2006)
sr5sieve now reports percentage of total range done, and estimated
 completion date.
sr5work.txt format changed: <pmin>,<pmax> is in billions, no 'e' notation.

Version 0.2.3:
Fixed a bug introduced in version 0.2.0 causing a segfault in test_bit()
 when a duplicate factor is found near the bottom of the sieve range.
When --duplicates and --check are specified together, doublecheck duplicate
 factors as well.
Relaxed the restriction on b in k*b^n+c having common factors with k or c.
 I can't remember why I imposed this condition in the first place, but it
 doesn't seem to apply anymore.
Changed behaviour of --checkpoint: Don't resume from checkpoint.txt unless
 it contains a value p such that pmin < p < pmax, where pmin and pmax are
 calulated from defaults or from the input file or given on the command line
 as usual.
Removed the --twin feature, since for k*b^n+/-1 forms it was faster to sieve
 (k^2)*(b^2)^n-1 (thanks R.Gerbicz for pointing this out), and for other
 forms it was no better than sieving the individual sequences and removing
 the non-twin terms from the resulting sieve files (now a TODO for srfile).

Version 0.2.2: (Posted 1 July 2006)
sr5sieve is a version of srsieve specialised for the base 5 projects. It
 accepts no command line arguments, but invoking 'sr5sieve' is equivalent to
 'srsieve --check --checkpoint --factors --report 120 --save 5 sr5data.txt'
 except that --pmin and --pmax are read as a comma-separated pair from a
 file sr5work.txt. When sieving is complete the first line of sr5work.txt is
 deleted and sieving starts again with the next line, until empty.

Version 0.2.1:
'srfile --pfactor' will print each term in the sieve as a Prime95 P-1
 worktodo.ini entry 'Pfactor=k,b,n,c,bit_level,0'.
Reduce base modulo p before calling modular arithmetic functions in
 small_sieve(). Luckily only mulmod/powmod were affected, as the current
 versions can tolerate values greater than p anyway.

Version 0.2.0: (Posted 29 June 2006)
Compressed bitmaps by representing n_i as q*m_i+r where q=min_{nmin<n_i<nmax}
 gcd(n_{i+1}-n_i).
srfile --congruence FILE prints congruence information for sequences in FILE.

Version 0.1.24: (Posted 26 June 2006)
--twin sieves for twin primes.

Version 0.1.23:
If the order of b is found in the baby steps then skip the giant steps.

Version 0.1.22:
Added support for NewPGen type 19 and 20 sieve files (b^n+k, b^n-k) when
 using --newpgen or --prp. LLR version 3.7.0 can read these formats, but PRP
 version 24.14 can't (it incorrectly tests k*b^n+/-1 instead of b^n+/-k).
NOTE NEW FILE NAMES: A NewPGen type T sieve for K*B^n+/-1 or B^n+/-K is now
 written to file t<T>_b<B>_k<K>.npg with --newpgen or to file t<T>_b<B>.prp
 with --prp. T is one of 16,17,19,20.
If k=1 then the sequence k*b^n+c can be written as b^n+c.
Increased the small sieve limit proportional to the sieve range (nmax-nmin).
 This should greatly improve the sieve speed for very small p whan starting
 a large new sieve.

Version 0.1.21: (Posted 22 June 2006)
Added PPC assembler versions of mulmod and powmod supplied by Mark Rodenkirch.
Re-arranged the Makefile so that extra CFLAGS can be passed on the make
 command line, e.g. 'make CFLAGS=-O3' just adds -O3 to the other flags
 instead of replacing them.

Version 0.1.20: (Posted 20 June 2006)
Set prime sieve limits at runtime so all primes up to p_max can be generated.
If --pmax is not specified, then default to p_min + 4 trillion.
Use bitmaps instead of byte arrays in prime sieve.
Made the reporting period independent of the sieve block size.
Command line arguments can be given in exponential notation, 1e6=1000000 etc.
--hashtable X now limits the maximum hashtable density to X (default 0.8).
 (Use lower densities to increase hashtable size).
Actual hashtable density and size is reported when --verbose is used.

Version 0.1.19: (Posted 17 June 2006)
Leave small prime terms in the sieve.
--no-base prevents base-specific mulmods being used.

Version 0.1.18: (Posted 14 June 2006)
Much improved hashtable code. When SHORT_HASHTABLE=1 the number of baby
 steps is limited to 2^15-1, which allows the memory used in both the main
 table and overflow list to be halved. References to the overflow list are
 avoided unless there is a collision. All this helps to keep the main table
 in cache longer, resulting in speedups on my P4 of about 60% for larger
 sieves and 5-10% for small sieves. For very large sieves a warning is given
 that it may be best to recompile with SHORT_HASHTABLE=0.
Hashtable size is reported when the --verbose switch is specified.
Reduced prime sieve block sizes as large blocks don't seem to gain anything.

Version 0.1.17:
Delete the checkpoint file if the sieving job is complete.
Fixed a potential bug in the assembler version of mulmod64() present since
 it was inlined in version 0.0.15. GCC could have re-ordered mulmod64() to
 put it before mod64_init() or after mod64_fini(). I don't think this ever
 actually happened in any released version of srsieve, as it would probably
 have been caught by 'make check'.

Version 0.1.16:
Added base specific mulmods for bases 2 and 3. There are probably gains to
 be had for other small bases too, I will add them if requested.

Version 0.1.15: (Posted 7 June 2006)
Added base five specific functions mul5mod32(a,p) and mul5mod64(a,p). They
 are employed in the baby steps only. The 32/64 bit base 5 sieve is approx.
 30%/10% faster than the general sieve on my P3, and 40%/30% faster on my P4.

Version 0.1.14: (Posted 6 June 2006)
Implemented pre2_mulmod64(a,b,p) which uses a pre-computed value of b/p held
 on the FPU stack. This is exploited in bsgs() where b/p is constant through
 the baby steps and (with a different b) the giant steps, resulting in a
 small speedup for the 64 bit sieve. (About 7% on my P3/450 for 20 sequences
 over 50K < n < 150K, perhaps a little more for larger sieves).

Version 0.1.13: (Posted 3 June 2006)
Added srfile --delete option to delete a sequence from the sieve.
Only report the total number of terms removed with srfile --known-factors
 unless --verbose is also specified.

Version 0.1.12: (Posted 2 June 2006)
Added --abcd switch to write a multiple header ABCD format sieve file.
Added --checkpoint switch which causes a one line checkpoint.txt file
 contining the current sieve prime to be written, and causes srsieve to
 resume sieving from checkpoint.txt if it exists. If both --checkpoint and
 --factors are given then no sieve file will be written unless explicitly
 requested. This will be useful when the input sieve is very large.

Version 0.1.11: (Posted 1 June 2006)
Allow blank lines and comments in input files. Lines with all whitespace
 characters or whose first non-whitespace character is '#' are ignored.
Report the line number where the error occured in input error reports.

Version 0.1.10: (Posted 30 May 2006)
Converted printf/scanf formats to use macros in <inttypes.h> for
 compatibility with mingw32.

Version 0.1.9:
Avoid using alarms, poll every time p increases by 1 million instead.
Misc other changes to compile cleanly with the Debian mingw32 cross compiler.

Version 0.1.8:
Added config.h option HAVE_ALARMS and code for systems that don't have them.

Version 0.1.7: (Posted 29 May 2006)
Avoid excessive memory use when writing sorted --prp or --pfgw files.
Rewrite of the generic C mulmod64() based on simpler code by Phil Carmody. I
 expect it to be better suited to non-x86 architectures.
Define limits for a,b,p along with the definition of mulmod64(a,b,p).

Version 0.1.6: (Posted 27 May 2006)
Minimised MAX_PRIMES_IN_TABLE in sieve.c to avoid unnecessary memory use,
 and fixed an array bounds bug and a bug that caused some small primes to be
 sieved twice.
Made USE_PRIME_SIEVE the default, define USE_PP_SIEVE to use the old
 probable prime code instead.

Version 0.1.5: (Posted 26 May 2006)
Reverted to original default sieve file format, as pfgw can't parse multiple
 header ABC files, only ABCD files, and these are not nice for hand editing.
Don't report each individual NewPGen file written unless --verbose is used.
Removed gcd functions from arithmetic.h as they are not critical.

Version 0.1.4: (Posted 25 May 2006)
Makefile option USE_PRIME_SIEVE uses Mark Rodenkirch's sieve code in sieve.c
 instead of the probable prime code, now in pp_sieve.c.

Version 0.1.3:
Allow c in k*b^n+c to be any number relatively prime to b and k, not just
 +1,-1.
Added --pfgw option to write the sieve in a format (abc, sorted by n)
 suitable for input to pfgw, as the --newpgen and --prp formats do not allow
 k*b^n+c with c other than +1,-1. Updated srfile to match.
Write a multiple-header abc format sieve file by default. This format is
 almost identical to the old srsieve format but can be read by pfgw and maybe
 other programs. Old format can still be read, but will not be written.

Version 0.1.2:
Put inline assembler functions in files asm-<architecture>-<compiler>.h.

Version 0.1.1: (Posted 22 May 2006)
Fixed a potential bug in the GCC assembler version of mulmod32 (it could
 only strike if explicitly invoked as mulmod32(a,b,b) which didn't happen).
Assembler versions of submod32/64().

Version 0.1.0:
Translated mulmod62() from assembler to C, GMP no longer required.
Removed sqrmod(a,p) as GCC generates the same code from inline mulmod(a,a,p).

Version 0.0.15: (Posted 18 May 2006)
Inlined the mulmod62() assembler for GCC, this speeds up the 64 bit sieve by
 about 5% on my P3/450.
Run 'make check' to perform two simple tests of the srsieve binary.

Version 0.0.14:
Changed "%ju" back to "%llu" in printf()/scanf().
--pmin was being ignored when starting a new sieve.

Version 0.0.13: (Posted 15 May 2006)
Allow command line --pmin to override pmin from input files.
Prefix underscores to global labels when compiling for Cygwin (define
 NEED_UNDERSCORE in config.h for other machines that need this) Thanks to
 `Greenbank' for fixing this.
Alternative millisec_clock() function for machines lacking getrusage().

Version 0.0.12: (Posted 13 May 2006)
Report the rate at which factors are being found (seconds per factor).
--stop-rate X stops sieving when the rate reaches X seconds per factor.
Added a new program srfile for converting between file formats and removing
 factors, other functions will be added later. Not well tested yet.

Version 0.0.11:
Fixed an incorrect assertation (j == count) in write_prp_file().
Replaced "%llu" with "%ju" which should work with 32 and 64 bit systems.
Included a .zip verison of the source archive as well as .tar.gz
Added MD5SUM.txt and README.txt for the yahoo directory.

Version 0.0.10: (Posted 10 May 2006)
If every term of a sequence is eliminated then the sequence is removed from
 the sieve. If there are no more sequences left then the program ends. These
 events are logged in srsieve.log.

Version 0.0.9: (Posted 9 May 2006)
The previous bug fix didn't work in all cases, fixed now.

Version 0.0.8: (Posted 8 May 2006)
Fixed a bug where composite p might have a common factor with k.
Removed --benchmark, it is too simplistic to be accurate.

Version 0.0.7:
Inlined the mulmod32() assembler for GCC, thanks to Anand Nair for his help
with this, it speeds up the 32 bit sieve by about 10% on my P3/450.
Removed powmod32.S as the C version is just as fast with mulmod32() inline.
Inlined common cases for insert() and lookup().
--hashtable allows a larger hashtable to be used.

Version 0.0.6: (Posted 6 May 2006)
--prp writes the sieve as two files suitable as input to PRP (sorted by n).
Simplified giant steps for the common case of p large relative to (nmax-nmin).
Fully buffer the factors file when p is small.
--babystep adjusts the ratio of baby steps to giant steps.

Version 0.0.5: (Posted 4 May 2006)
Added i386 assembler versions of mulmod32(), powmod32() and mulmod64() (now
 mulmod62()) supplied by Mark Rodenkirch. On my P3/450 mulmod32/62 are twice
 and five times faster respectively compared to the C/GMP versions.
Added a --benchmark command line option to time some critical functions.
Added --duplicates and --order options to help with debugging and tuning.

Version 0.0.4: (Posted 2 May 2006)
Added a missing bitmap sentinel bit to prevent a segfault caused by an
 unbounded next_bit(). Thanks to Mark Rodenkirch for spotting this bug.
Added a compile option BSGS_HASH to select a hash function as the residue
 modulo a prime or (Mark Rodenkirch's suggestion) modulo a power of 2.
 The latter seems to be faster.
--report R makes a simple timing report (p/cpu.sec) every R seconds.
Block signals while writing save files.

Version 0.0.3: (Posted 1 May 2006).
Added a more compact save file format, in addition to the NewPGen
 one-candidate-sequence-per-file format.
Use next_prime() instead of getprime(), which takes too long to initialise
 at large pmin. next_prime() needs tuning (or rewriting).

Version 0.0.2:
GMP (mpn) arithmetic for primes above 2^32. This is just a temporary measure
 until proper 64 bit arithmetic is implemented.
Simple signal handling to write save files and screen progress reports.
Hashtable only stores indices into an existing array of data, not the data
 itself.

Version 0.0.1: (Posted 27 April 2006).
Baby steps giant steps algorithm seems to work.
