New features in Version 19.1 of mprime
--------------------------------------

1)  A new undocumented feature, Stage1GCD, lets you speed up P-1 tests
    on large exponents - see undoc.txt for details.
2)  K6 CPUs now default to the Pentium code path instead of the Pentium
    Pro code path.  This is faster on the K6.
3)  A bug that caused negative elapsed time on 486 and Cyrix CPUs has been
    fixed.
4)  The rolling average formula was adjusted to work better on machines
    that are not on 24 hours a day.
5)  You can now put EntropiaIP=207.104.25.155 in primenet.ini to work
    around error 2250 problems when using the statically linked mprime
    on older Linux versions.

New features in Version 19.0 of mprime
--------------------------------------

1)  Faster - in some cases as much as 10% faster!.  The FFTs were recoded
    for improved memory and TLB efficiency.  Furthermore, optimizations
    specific to the Pentium Pro and later processors were added.
2)  New FFT lengths.  The program can now test exponents as large as
    79.3 million.  Also, smaller FFT lengths are supported for use in
    ECM and P-1 factoring.
3)  More conservative FFT breakpoints.  This could actually result in
    some exponents being slower to test in this version.  However, the
    chance of a fatal rounding error has been reduced.
4)  P-1 factoring has been added.  Although it is not very practical for
    large exponents because of a slow GCD routine, it can be used to
    find new factors of exponents below a million or so.
5)  ECM can now run on large exponents.  Once again, the slow GCD routine
    might make this impractical for large exponents.
6)  ECM and P-1 factoring now support save files.  Very handy on lengthy runs.
    The count of completed ECM curves is maintained in the save file instead
    of the worktodo.ini file.
7)  ECM and P-1 factoring lets you specify the amount of memory to use.
    More memory can improve execution speed slightly.
8)  A bug in guessing the CPU speed on initial install has been fixed.
9)  Error checking has been improved slightly.
10) Factoring is now "layered".  That is, prime95 now factors to 2^50, then
    2^51, 2^52, and so forth up to the appropriate limit.  The factoring
    output lines have been changed to show percent complete in the current
    "layer".
11) A bug in running two or more self or torture tests in the same directory
    has been fixed
12) Trial factoring above 2^64 is now supported.
13) More trial factoring is now done to take into account the cost of
    double-checking.
14) By default, the percent complete value is now displayed to 2 decimal
    places.  You can change this by setting PercentPrecision in prime.ini
    to a value between 0 and 6.
15) An option to get only 10,000,000 digit numbers to run primality tests
    on has been added to the Test/Primenet dialog box.  See
    http://www.mersenne.org/prize.htm for rules on claiming the EFF award
    for finding a 10,000,000 digit prime.
16) The User Information dialog box allows you to request newsletters and
    form a team user ID where the team members cannot alter the team name.
17) A bug in the reporting of error counts in the results.txt file has
    been fixed.
18) The server can now broadcast important messages to the mprime client.
19) Mprime now tries to prevent two copies of mprime from running
    simultaneously.  Although not foolproof, it should greatly reduce
    the chances of corrupting the intermediate files.  Note, multi-processor
    systems are not affected due to using the -An switch or running in
    different directories.
20) The -c option allows you to run mprime, have it communicate with the
    PrimeNet server, and then exit.
21) A computer ID will be generated for computers that do not have one.  If
    you want to assign your own computer ID, please try to make each
    computer ID using the same user ID unique.

New features in Version 18.1 of mprime
--------------------------------------

1)  Bug fixed in Lucas-Lehmer testing of exponents above 4194304.
2)  Added code so that server can distinguish between a v17 and v18 client.

New features in Version 17.1 of mprime
--------------------------------------

1)  You can now fine tune your control of the program by adding Time=
    lines to your prime.ini file - see the readme.txt file.
2)  ECM factoring for 2^N+1 is now available.
3)  By default, ECM factoring stops if a factor is found for exponents
    above 5825 and continues (if the cofactor is composite) for exponents
    below 5825.  You can override this behavior by setting ContinueECM=0
    or 1 in prime.ini.

New features in Version 17.0 of mprime
--------------------------------------

1)  Mprime can now be used to double-check its own results.
2)  Machines less powerful than a P133 will default to double-checking
    assignments.
3)  The total number of errors that occur during the Lucas-Lehmer test
    are now output in results.txt.

New features in Version 16.4 of mprime
--------------------------------------

1)  Lucas-Lehmer testing output now shows percent complete.
2)  ECM factoring is available for small exponents.  See
    http://www.mersenne.org/ecm.htm for a list of exponents that
    need factoring.
3)  A bug that prevented 486 & Cyrix machines from factoring above 2^62
    has been fixed.
4)  The LockUserInfo feature no longer sends user info every time prime95
    starts up.
5)  Test/Status now shows expected completion dates.

New features in Version 16.3 of mprime
--------------------------------------

1)  Mprime can now test exponents up to 20.5 million.
2)  Mprime can now find factors up to 2^64.
3)  The program now reports to the PrimeNet server new completion dates
    every 28 days (actually user adjustable from 1 to 60 days).  This will
    improve the reclaiming of exponents that are not being actively tested.
4)  The program now reports the current iteration number to the server
    for more informative status reports.
5)  Intermediate files are now less likely to fail.  If an error occurs
    during the writing of an intermediate file, the program continues on
    trying to write the intermediate file every 10 minutes.
6)  A bug in the rolling average code has been fixed.  Mprime now modifies
    the rolling average when factoring too.
7)  The prime.log file is now limited in size to 256K bytes.
8)  The user information dialog box now has a checkbox to tell the server
    not to send email.
9)  Setting LockUserInfo=1 in prime.ini will make the user information
    dialog box inaccessible.  This setting is also sent to the server
    so that no member of a team can accidentally transfer the credit
    for work previously performed by the team.

New features in Version 15.4 of mprime
--------------------------------------

1)  Mprime can now contact a central server to get exponents to test,
    report results, and change your email address.  No database is required!
    The old manual reservation system is still supported.
2)  The command line interface is replaced by a series of menus and
    questions to control your preferences.  These menus are similar
    to the menus in the Windows executable.  All preferences are
    now stored in local INI files that are compatible with the Windows
    program.  This should make mprime easier to use, especially for those
    users running both operating systems.
3)  The Pentium factoring code (up to 60 bit factors) is about 4.5% faster.
4)  The CPU menu now has an entry for the AMD K6 and an entry
    for how many hours per day you expect to be running the program.
5)  Mprime will no longer let you select Pentium Pro or Pentium CPU type
    unless your CPU supports the process-specific instructions.
6)  You can now change how often prime95 saves intermediate files to disk.
7)  You can now select one save file or two.
8)  -A switch added.  This allows you to run multiple copies of prime95
    from the same directory.  This is useful on multi-processor machines.
9)  Mprime now changes its working directory to the same directory that
    mprime is located.  This makes it easier to launch mprime from startup
    scripts.  If you must run mprime from a different directory than the
    rest of the files, then use the new -W switch.
10) The time estimate in the Status menu choice now takes into account
    the time you've invested in the current Lucas-Lehmer test.  It also
    takes into account how many hours per day you are running the program.
    It also adjusts its time estimate if your computer seems to be running
    faster or slower than expected.
11) Timestamps in the results.txt file.  
12) The results file is now named results.txt, not results.  

New features in Version 14.4 of mprime
--------------------------------------

1)  The Pentium Pro factoring code is nearly twice as
    fast compared to the previous version.

New features in Version 14.2 of mprime
--------------------------------------

1)  Corrected floating point round-off bugs in the factoring code that
    affected testing factors larger than 2^59.
2)  The Pentium factoring code is 5-10% faster.
3)  When an error is detected the program now waits 5 minutes before 
    restarting from the last save file.
4)  The detection of corrupt save files has improved.

New features in Version 14.1 of mprime
--------------------------------------

1)  The Pentium factoring code is 5% faster.

New features in Version 14.0 of mprime
--------------------------------------

1)  The program can now check exponents up to 5.26 million!  To test
    these big exponents, you will need a new database and great patience.
2)  The per-iteration error checking code has been improved.  It will
    detect more errors and automatically restart from the previous save
    file.
3)  The -c option now has a number for Cyrix machines.  There is no
    Cyrix-specific code.  This change does provide better estimates in
    status option.
4)  There is a machine torture test option which runs a continuous Self Test.
    This is great for "burning in" a new machine.
5)  The time estimates in status option have been changed slightly.
    Hopefully, the new estimates are a little more accurate.
6)  The self test option now takes nine hours.

New features in Version 13.3.2 of mprime
----------------------------------------

1) It's smaller.  Mprime is no longer linked
   with the static C runtime library.  The staticly
   linked version is still available for a.out Linux systems.
2) Improved error checking should detect more
   hardware problems with fewer false alarms.

New features in Version 13.1.2 of mprime
----------------------------------------

1) Faster!  Support for non-power-of-2 FFT lengths
   has been added.
2) The -I switch has been added.  This switch prints
   a status line to the results file every N iterations.
   For example, -I50000 will print a line to the results
   file after Lucas-Lehmer iteration 50000, 100000, etc.
3) The program now creates two backup files in case one
   is corrupted during a system crash.

New features in Version 1.3 of mprime
-------------------------------------

1) A minor bug in the status report code is fixed.
2) The -L and -P switches have been implemented, but
   not tested.  The -L switch takes a floating point
   number as an argument.  If the load as reported in
   /proc/loadavg rises above this threshold, then the
   program pauses for as many seconds as specified by
   the -P switch.  This allows you to free up mprime's
   memory for other tasks to use during high load
   conditions.

New features in Version 1.2 of mprime
-------------------------------------

1)  A minor bug in the Pentium factoring code has been fixed.

New features in Version 1.1 of mprime
-------------------------------------

1)  Factoring is now 6% faster due to a 16 pass
    process rather than a 2 pass process.

New features in Version 1.0 of mprime
-------------------------------------

1)  Initial release - all features are new!
