Skip to content

Latest commit

 

History

History
129 lines (117 loc) · 5.12 KB

QUERIES.md

File metadata and controls

129 lines (117 loc) · 5.12 KB

##Structure ####settings Contains the settings the tests were run with.

  • ID: An ID just for the database
  • N: The size of the number that is going to be factorized. If N is approximately 10^14, then N holds the number 14
  • c: A parameter of the program
  • dim: A parameter of the program. Sometimes called n in the program.
  • s_max: A parameter of the program also called "the pruning level".
  • reduce_ratio: A parameter that is most of the time fixed to 0.8 and not widely used.
  • start_reduction: A parameter that is also named a in the master thesis.
  • bkz_strong: The block size of the strong BKZ reduction
  • bkz_slight: The block size of the slight BKZ reduction
  • seed: The seed the program was started with
  • cf: An indicator if continued fractions (1), centered continued fraction (2) or none of them (0) are used.
  • scaling_type: the scaling type that is used. This is an integer between 0 and 5.
  • file: The file where you can see the results.

###statistics Some statistics about the tests:

  • settings_id Foreign key to the settings
  • runtime_total
  • runtime_per_uniq_eqn
  • rounds: The number of rounds in total
  • rounds_wo_dist_red: The number of rounds without a reduction of the distance
  • rounds_w_eqns: The number of rounds where at least one equation was found
  • uniq_eqns: The number of uniq equations
  • duplicate_eqns: The number of equations found at least twice
  • total_checked_stages_rwe: Total checked stages in rounds where equations were found (rwe - rounds with equations)
  • min_checked_stages_rwe
  • max_checked_stages_rwe
  • avg_checked_stages_rwe
  • total_checked_stages_rwoe: Total checked stages in rounds where no equations were found (rwoe - rounds without equations)
  • min_checked_stages_rwoe
  • max_checked_stages_rwoe
  • avg_checked_stages_rwoe
  • min_time_newenum_total: Min time NewEnum took (any round)
  • max_time_newenum_total
  • avg_time_newenum_total
  • min_time_newenum_we: Min time NewEnum took in rounds where equations were found
  • max_time_newenum_we
  • avg_time_newenum_we
  • min_time_slight_bkz: Min time the slight BKZ reduction took
  • max_time_slight_bkz
  • avg_time_slight_bkz
  • max_distance_reduction: The max distance reduction. This is a relative number (relative to the GNF of the reduced lattice basis)
  • min_distance_reduction
  • avg_distance_reduction
  • eqns_newenum_only: The number of equations that were found by NewEnum. (Not in the continued fraction extension)
  • min_v_value_neo: Minimal value of v used in equations from NewEnum only (neo)
  • max_v_value_neo
  • avg_v_value_neo
  • median_v_value_neo
  • eqns_cf_only: The number of equations that were found by the continued fractions extension only.
  • min_v_value_cfo Minimal value of v used in equations from continued fraction only (cfo)
  • max_v_value_cfo
  • avg_v_value_cfo
  • median_v_value_cfo

###rounds

  • ID: An ID used as foreign key in the equations table
  • settings_id: Foreign key to the settings table
  • num_primes: the number of scaled primes
  • primes: The scaled primes it self. This is formatted as a JSON string
  • round: The number of the round in the corresponding test

###equations The equations $u\equiv \pm |u-vN| mod N$ that were found during the tests.

  • rounds_id: Foreign key to the rounds table
  • equation_id
  • u: u from the equation. Stored as a JSON string containing primes and the exponent.
  • u_min_prime: smallest prime in u
  • u_max_prime: biggest prime in u
  • u_median_prime
  • u_num_primes: The number of different prime factors in u
  • u_max_e: The highest exponent of a prime in u
  • v: v from the equation. Stored as a string, since the number can get very big.
  • v_smooth: Indicator if v is a smooth number (1) or not (0)
  • rs: |u-vN| from the equation. ALso stored as a JSON string
  • rs_sign: The sign of u-vN
  • rs_min_prime: smallest prime in |u-vN|
  • rs_max_prime biggest prime in |u-vN|
  • rs_median_prime
  • rs_num_primes: The number of different prime factors in |u-vN|
  • rs_max_e: The highest exponent of a prime in |u-vN|
  • cf: An indicator that shows if the equation was found via continued fractions (1) or from NewEnum (0)
  • dist: The relative distance the equation was found at.
  • level: The level the equation was found at.

##Queries

  • The number of tests and the time they took in hours grouped by N.
SELECT
   N,
   COUNT(*),
   SUM(runtime_total) / 3600
 FROM statistics
   JOIN settings ON statistics.settings_id = settings.ID
   WHERE file NOT LIKE 'special\\\\%'
 GROUP BY N ASC WITH ROLLUP;
  • The number of distinct equations found grouped by N
SELECT COUNT(DISTINCT u, rs)
FROM equations
  JOIN rounds ON equations.rounds_id = rounds.ID
  JOIN settings ON rounds.settings_id = settings.ID
WHERE file NOT LIKE 'special\\\\%'
GROUP BY N ASC WITH ROLLUP;
  • The number of equations found in the c for N = 20 test series and the number of tests grouped by c
SELECT
  c,
  COUNT(*),
  COUNT(DISTINCT settings_id)
FROM settings
  JOIN rounds ON settings.ID = rounds.settings_id
  JOIN equations ON rounds.ID = equations.rounds_id
WHERE file LIKE '%c\\\\20%' AND equations.cf = 0
GROUP BY c;