Metrics for Logistic regression 2

analyze86.0ms (1.4%)

Algorithm

1×

Search

Probability	Valid	Unknown	Precondition	Infinite	Domain	Can't	Iter
0%	0%	99.9%	0.1%	0%	0%	0%	0
0%	0%	99.9%	0.1%	0%	0%	0%	1
50%	50%	50%	0.1%	0%	0%	0%	2
50%	50%	50%	0.1%	0%	0%	0%	3
75%	74.9%	25%	0.1%	0%	0%	0%	4
75%	74.9%	25%	0.1%	0%	0%	0%	5
75%	74.9%	25%	0.1%	0%	0%	0%	6
75%	74.9%	25%	0.1%	0%	0%	0%	7
75%	74.9%	25%	0.1%	0%	0%	0%	8
75%	74.9%	25%	0.1%	0%	0%	0%	9
75%	74.9%	25%	0.1%	0%	0%	0%	10
75%	74.9%	25%	0.1%	0%	0%	0%	11
75%	74.9%	25%	0.1%	0%	0%	0%	12

Compiler

Compiled 12 to 9 computations (25% saved)

Precisions

Click to see histograms. Total time spent on operations: 45.0ms

Operation ival-log, time spent: 17.0ms, 38.0% of total-time

Operation ival-add, time spent: 8.0ms, 18.0% of total-time

Operation ival-mult, time spent: 7.0ms, 15.0% of total-time

Operation ival-exp, time spent: 7.0ms, 15.0% of total-time

Operation ival-sub, time spent: 5.0ms, 11.0% of total-time

Operation const, time spent: 1.0ms, 2.0% of total-time

sample3.6s (57.8%)

Results

1.2s	7511×	0	valid
2.1s	2661×	6	exit
104.0ms	744×	0	infinite
0.0ms	1×	0	valid

Precisions

Click to see histograms. Total time spent on operations: 2.1s

Operation ival-sub, time spent: 495.0ms, 23.0% of total-time

Operation ival-add, time spent: 494.0ms, 23.0% of total-time

Operation ival-exp, time spent: 440.0ms, 21.0% of total-time

Operation ival-log, time spent: 411.0ms, 19.0% of total-time

Operation ival-mult, time spent: 225.0ms, 11.0% of total-time

Operation const, time spent: 56.0ms, 3.0% of total-time

Bogosity

preprocess1.6s (26.5%)

Algorithm

1×	egg-herbie

Rules

1540×	`times-frac`
1328×	`fma-def`
800×	`fma-neg`
552×	`cancel-sign-sub-inv`
470×	`swap-sqr`

FPErrors

Click to see full error table

Example	Example	Subexpression
-	-	`x`
-	-	`(*.f64 x y)`
-	-	`(log.f64 (+.f64 1 (exp.f64 x)))`
-	-	`(+.f64 1 (exp.f64 x))`
-	-	`(exp.f64 x)`
-	-	`(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))`
-	-	`y`
-	-	`1`

Iterations

Useful iterations: 2 (0.0ms)

Iter	Nodes	Cost
0	30	234
1	73	226
2	129	210
3	286	210
4	468	210
5	769	210
6	1322	210
7	2788	210
8	5719	210
9	7847	210

Stop Event

1×	node limit

Calls

Call 1

Inputs
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))
(-.f64 (log.f64 (+.f64 1 (exp.f64 (neg.f64 x)))) (*.f64 (neg.f64 x) y))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x (neg.f64 y)))
(neg.f64 (-.f64 (log.f64 (+.f64 1 (exp.f64 (neg.f64 x)))) (*.f64 (neg.f64 x) y)))
(neg.f64 (-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x (neg.f64 y))))
(-.f64 (log.f64 (+.f64 1 (exp.f64 y))) (*.f64 y x))

Outputs
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
(-.f64 (log.f64 (+.f64 1 (exp.f64 (neg.f64 x)))) (*.f64 (neg.f64 x) y))
(+.f64 (log1p.f64 (exp.f64 (neg.f64 x))) (*.f64 x y))
(+.f64 (*.f64 x y) (log1p.f64 (exp.f64 (neg.f64 x))))
(fma.f64 x y (log1p.f64 (exp.f64 (neg.f64 x))))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x (neg.f64 y)))
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x (neg.f64 y)))
(+.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))
(fma.f64 x y (log1p.f64 (exp.f64 x)))
(neg.f64 (-.f64 (log.f64 (+.f64 1 (exp.f64 (neg.f64 x)))) (*.f64 (neg.f64 x) y)))
(neg.f64 (+.f64 (log1p.f64 (exp.f64 (neg.f64 x))) (*.f64 x y)))
(neg.f64 (+.f64 (*.f64 x y) (log1p.f64 (exp.f64 (neg.f64 x)))))
(neg.f64 (fma.f64 x y (log1p.f64 (exp.f64 (neg.f64 x)))))
(neg.f64 (-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x (neg.f64 y))))
(neg.f64 (-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x (neg.f64 y))))
(neg.f64 (+.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y)))
(neg.f64 (fma.f64 x y (log1p.f64 (exp.f64 x))))
(-.f64 (log.f64 (+.f64 1 (exp.f64 y))) (*.f64 y x))
(-.f64 (log1p.f64 (exp.f64 y)) (*.f64 x y))
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 y)))

Compiler

Compiled 69 to 38 computations (44.9% saved)

Precisions

Click to see histograms. Total time spent on operations: 1.0s

Operation ival-add, time spent: 406.0ms, 39.0% of total-time

Operation ival-exp, time spent: 311.0ms, 30.0% of total-time

Operation ival-log, time spent: 204.0ms, 20.0% of total-time

Operation ival-sub, time spent: 60.0ms, 6.0% of total-time

Operation ival-mult, time spent: 35.0ms, 3.0% of total-time

Operation const, time spent: 16.0ms, 2.0% of total-time

eval0.0ms (0%)

Compiler: Compiled 16 to 8 computations (50% saved)

prune1.0ms (0%)

Alt Table

Click to see full alt table

Status	Accuracy	Program
▶	97.3%	(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
▶	97.3%	(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))

Compiler

Compiled 18 to 12 computations (33.3% saved)

localize388.0ms (6.3%)

Compiler: Compiled 36 to 9 computations (75% saved)
Precisions: Click to see histograms. Total time spent on operations: 282.0ms
Operation ival-exp, time spent: 126.0ms, 45.0% of total-time
Operation ival-log1p, time spent: 76.0ms, 27.0% of total-time
Operation ival-sub, time spent: 35.0ms, 12.0% of total-time
Operation ival-fma, time spent: 21.0ms, 7.0% of total-time
Operation ival-mult, time spent: 13.0ms, 5.0% of total-time
Operation ival-neg, time spent: 8.0ms, 3.0% of total-time
Operation const, time spent: 2.0ms, 1.0% of total-time

eval0.0ms (0%)

Compiler: Compiled 2 to 2 computations (0% saved)

prune2.0ms (0%)

Pruning

2 alts after pruning (0 fresh and 2 done)

	Kept	Total
New	0	0
Fresh	0	0
Picked	2	2
Done	0	0
Total	2	2

Accuracy

97.3%

Counts

2 → 2

Alt Table

Click to see full alt table

Status	Accuracy	Program
✓	97.3%	(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
✓	97.3%	(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))

Compiler

Compiled 43 to 24 computations (44.2% saved)

regimes10.0ms (0.2%)

Counts

3 → 1

Calls

Call 1

Inputs
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))

Outputs
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))

Calls

3 calls:

3.0ms	y
3.0ms	x
3.0ms	(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))

Results

Accuracy	Segments	Branch
97.3%	1	`x`
97.3%	1	`y`
97.3%	1	`(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))`

Compiler

Compiled 17 to 12 computations (29.4% saved)

regimes14.0ms (0.2%)

Accuracy

Total -0.0b remaining (-0.2%)

Threshold costs -0b (-0.2%)

Counts

2 → 1

Calls

Call 1

Inputs
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))
(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))

Outputs
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))

Calls

3 calls:

8.0ms	y
3.0ms	x
2.0ms	(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))

Results

Accuracy	Segments	Branch
97.3%	1	`x`
97.3%	1	`y`
97.3%	1	`(-.f64 (log.f64 (+.f64 1 (exp.f64 x))) (*.f64 x y))`

Compiler

Compiled 17 to 12 computations (29.4% saved)

simplify3.0ms (0.1%)

Algorithm

1×	egg-herbie

Rules

8×	`sub-neg`
8×	`+-commutative`
8×	`*-commutative`
6×	`neg-mul-1`
6×	`neg-sub0`

Iterations

Useful iterations: 0 (0.0ms)

Iter	Nodes	Cost
0	15	60
1	25	60
2	36	60
3	41	60
4	43	60

Stop Event

1×	done
1×	saturated

Calls

Call 1

Inputs
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))

Outputs
(fma.f64 x (neg.f64 y) (log1p.f64 (exp.f64 x)))
(-.f64 (log1p.f64 (exp.f64 x)) (*.f64 x y))

Compiler

Compiled 18 to 12 computations (33.3% saved)

soundness417.0ms (6.8%)

Rules

1540×	`times-frac`
1540×	`times-frac`
1328×	`fma-def`
1328×	`fma-def`
800×	`fma-neg`

Iterations

Useful iterations: 2 (0.0ms)

Iter	Nodes	Cost
0	30	234
1	73	226
2	129	210
3	286	210
4	468	210
5	769	210
6	1322	210
7	2788	210
8	5719	210
9	7847	210
0	30	234
1	73	226
2	129	210
3	286	210
4	468	210
5	769	210
6	1322	210
7	2788	210
8	5719	210
9	7847	210

Stop Event

1×	node limit
1×	node limit

Compiler

Compiled 85 to 45 computations (47.1% saved)

end0.0ms (0%)

preprocess48.0ms (0.8%)

Compiler: Compiled 110 to 72 computations (34.5% saved)

analyze86.0ms (1.4%)

sample3.6s (57.8%)

preprocess1.6s (26.5%)

eval0.0ms (0%)

prune1.0ms (0%)

localize388.0ms (6.3%)

eval0.0ms (0%)

prune2.0ms (0%)

regimes10.0ms (0.2%)

regimes14.0ms (0.2%)

simplify3.0ms (0.1%)

soundness417.0ms (6.8%)

end0.0ms (0%)

preprocess48.0ms (0.8%)

Profiling