Calculate implied edge weights from joint probabilities

[smadhuk]

Add in edge weight calculation when creating a user_graph with correlations enabled. Populating edges with their implied weight is a TODO.

-- Performance before this PR --

[.................*<<|....................] 530 us (vs 290 us) (2.8 Mdets/s) (9.5 Mlayers/s) (1.9 Mshots/s) Decode_surface_r5_d5_p1000
[................*<<<|....................]  23 ms (vs  10 ms) (920 kdets/s) ( 59 klayers/s) (5.3 kshots/s) Decode_surface_r11_d11_p100
[................*<<<|....................] 3.3 ms (vs 1.5 ms) (2.9 Mdets/s) (1.6 Mlayers/s) (150 kshots/s) Decode_surface_r11_d11_p1000
[..................*<|....................] 140 us (vs  83 us) (3.6 Mdets/s) ( 19 Mlayers/s) (1.7 Mshots/s) Decode_surface_r11_d11_p10000
[..................*<|....................]  57 us (vs  33 us) (3.3 Mdets/s) (190 Mlayers/s) ( 17 Mshots/s) Decode_surface_r11_d11_p100000
[.................*<<|....................]  16 ms (vs 7.5 ms) (630 kdets/s) ( 10 klayers/s) (490 shots/s) Decode_surface_r21_d21_p100
[................*<<<|....................]  20 ms (vs 7.8 ms) (490 kdets/s) (8.0 klayers/s) (380 shots/s) Decode_surface_r21_d21_p100_with_dijkstra
[................*<<<|....................]  22 ms (vs 8.2 ms) (460 kdets/s) (7.5 klayers/s) (360 shots/s) Decode_surface_r21_d21_p100_to_edges
[.................*<<|....................]  12 ms (vs 6.3 ms) (2.8 Mdets/s) (410 klayers/s) ( 19 kshots/s) Decode_surface_r21_d21_p1000
[...............*<<<<|....................]  23 ms (vs 7.7 ms) (1.5 Mdets/s) (220 klayers/s) ( 10 kshots/s) Decode_surface_r21_d21_p1000_with_dijkstra
[...............*<<<<|....................]  25 ms (vs 8.4 ms) (1.4 Mdets/s) (210 klayers/s) ( 10 kshots/s) Decode_surface_r21_d21_p1000_to_edges
[................*<<<|....................] 2.3 ms (vs 980 us) (3.3 Mdets/s) (4.6 Mlayers/s) (220 kshots/s) Decode_surface_r21_d21_p10000
[...............*<<<<|....................] 4.1 ms (vs 1.3 ms) (1.8 Mdets/s) (2.5 Mlayers/s) (120 kshots/s) Decode_surface_r21_d21_p10000_with_dijkstra
[...............*<<<<|....................] 4.5 ms (vs 1.4 ms) (1.6 Mdets/s) (2.3 Mlayers/s) (110 kshots/s) Decode_surface_r21_d21_p10000_to_edges
[.................*<<|....................] 180 us (vs  94 us) (4.0 Mdets/s) ( 59 Mlayers/s) (2.8 Mshots/s) Decode_surface_r21_d21_p100000
[...............*<<<<|....................] 360 us (vs 130 us) (1.9 Mdets/s) ( 29 Mlayers/s) (1.3 Mshots/s) Decode_surface_r21_d21_p100000_with_dijkstra
[...............*<<<<|....................] 400 us (vs 130 us) (1.7 Mdets/s) ( 26 Mlayers/s) (1.2 Mshots/s) Decode_surface_r21_d21_p100000_to_edges
[..................*<|....................]  59 ms (vs  35 ms) (160 loads/s) Load_dem_r11_d11_p100
[..................*<|....................] 430 ms (vs 280 ms) ( 23 loads/s) Load_dem_r21_d21_p100
[...................*|....................]  45 ms (vs  35 ms) (210 loads/s) Load_dem_r11_d11_p100_correlations
[...................*|....................] 370 ms (vs 280 ms) ( 26 loads/s) Load_dem_r21_d21_p100_correlations
[................*<<<|....................] 9.5 ms (vs 3.5 ms) (1.0 Gcalls/s) Varying32_get_distance_at_time
[..............*<<<<<|....................]  25 ms (vs 5.9 ms) (390 Mcalls/s) Varying64_get_distance_at_time
[....................*....................] 5.0 us (vs 4.9 us) (190 MEnqueueDequeues/s) bucket_queue_sort
[..................*<|....................] 140 us (vs  99 us) ( 67 MEnqueueDequeues/s) bucket_queue_stream

-- Performance after this PR --

[.................*<<|....................] 530 us (vs 290 us) (2.8 Mdets/s) (9.5 Mlayers/s) (1.9 Mshots/s) Decode_surface_r5_d5_p1000
[................*<<<|....................]  24 ms (vs  10 ms) (900 kdets/s) ( 58 klayers/s) (5.2 kshots/s) Decode_surface_r11_d11_p100
[................*<<<|....................] 3.5 ms (vs 1.5 ms) (2.8 Mdets/s) (1.5 Mlayers/s) (140 kshots/s) Decode_surface_r11_d11_p1000
[..................*<|....................] 140 us (vs  83 us) (3.6 Mdets/s) ( 19 Mlayers/s) (1.8 Mshots/s) Decode_surface_r11_d11_p10000
[..................*<|....................]  56 us (vs  33 us) (3.3 Mdets/s) (190 Mlayers/s) ( 18 Mshots/s) Decode_surface_r11_d11_p100000
[.................*<<|....................]  16 ms (vs 7.5 ms) (620 kdets/s) ( 10 klayers/s) (480 shots/s) Decode_surface_r21_d21_p100
[................*<<<|....................]  21 ms (vs 7.8 ms) (460 kdets/s) (7.6 klayers/s) (360 shots/s) Decode_surface_r21_d21_p100_with_dijkstra
[................*<<<|....................]  22 ms (vs 8.2 ms) (450 kdets/s) (7.4 klayers/s) (350 shots/s) Decode_surface_r21_d21_p100_to_edges
[.................*<<|....................]  13 ms (vs 6.3 ms) (2.8 Mdets/s) (410 klayers/s) ( 19 kshots/s) Decode_surface_r21_d21_p1000
[...............*<<<<|....................]  26 ms (vs 7.7 ms) (1.3 Mdets/s) (200 klayers/s) (9.7 kshots/s) Decode_surface_r21_d21_p1000_with_dijkstra
[................*<<<|....................]  23 ms (vs 8.4 ms) (1.5 Mdets/s) (230 klayers/s) ( 11 kshots/s) Decode_surface_r21_d21_p1000_to_edges
[................*<<<|....................] 2.2 ms (vs 980 us) (3.4 Mdets/s) (4.7 Mlayers/s) (220 kshots/s) Decode_surface_r21_d21_p10000
[...............*<<<<|....................] 4.1 ms (vs 1.3 ms) (1.8 Mdets/s) (2.5 Mlayers/s) (120 kshots/s) Decode_surface_r21_d21_p10000_with_dijkstra
[...............*<<<<|....................] 4.1 ms (vs 1.4 ms) (1.8 Mdets/s) (2.5 Mlayers/s) (120 kshots/s) Decode_surface_r21_d21_p10000_to_edges
[.................*<<|....................] 170 us (vs  94 us) (4.1 Mdets/s) ( 61 Mlayers/s) (2.9 Mshots/s) Decode_surface_r21_d21_p100000
[................*<<<|....................] 320 us (vs 130 us) (2.2 Mdets/s) ( 33 Mlayers/s) (1.5 Mshots/s) Decode_surface_r21_d21_p100000_with_dijkstra
[................*<<<|....................] 330 us (vs 130 us) (2.1 Mdets/s) ( 32 Mlayers/s) (1.5 Mshots/s) Decode_surface_r21_d21_p100000_to_edges
[..................*<|....................]  53 ms (vs  35 ms) (180 loads/s) Load_dem_r11_d11_p100
[..................*<|....................] 400 ms (vs 280 ms) ( 24 loads/s) Load_dem_r21_d21_p100
[....................|*...................]  25 ms (vs  35 ms) (390 loads/s) Load_dem_r11_d11_p100_correlations
[....................|>*..................] 180 ms (vs 280 ms) ( 54 loads/s) Load_dem_r21_d21_p100_correlations
[................*<<<|....................] 9.6 ms (vs 3.5 ms) (1.0 Gcalls/s) Varying32_get_distance_at_time
[..............*<<<<<|....................]  25 ms (vs 5.9 ms) (390 Mcalls/s) Varying64_get_distance_at_time
[....................*....................] 4.8 us (vs 4.9 us) (200 MEnqueueDequeues/s) bucket_queue_sort
[..................*<|....................] 150 us (vs  99 us) ( 63 MEnqueueDequeues/s) bucket_queue_stream

[oscarhiggott]

Mostly looks good, however the weights should be kept as floats while they are in the UserGraph and only converted to ints after they are rescaled during the conversion to a MatchingGraph

[oscarhiggott]

The weights should just be stored as floats, not ints, at this stage. The weights need to be rescaled before being turned into integers, but this is only done when the UserGraph is converted to the MatchingGraph

[smadhuk]

Fixed (this code was actually removed) by changes from an upstream PR

[oscarhiggott]

This is something we don't want to happen (should rescale before converting to int)

[smadhuk]

Fixed (this code was actually removed) by changes from an upstream PR

[oscarhiggott]

It should be:

double new_weight;

instead of

weight_int new_weight;

[smadhuk]

Fixed by changes from an upstream PR