FuzzBench: 2024-09-03-dgfuzz report

experiment summary

We show two different aggregate (cross-benchmark) rankings of fuzzers. The first is based on the average of per-benchmarks scores, where the score represents the percentage of the highest reached median code-coverage on a given benchmark (higher value is better). The second ranking shows the average rank of fuzzers, after we rank them on each benchmark according to their median reached code-covereges (lower value is better).

By avg. score

	average normalized score
fuzzer
dgfuzz_e0d055	99.18
aflplusplus	93.44
eclipser	86.82
mopt	70.87
aflsmart	69.36
fairfuzz	69.31
aflfast	67.03
centipede	65.76
libafl	41.67
honggfuzz	13.69
libfuzzer	13.15
afl	12.97

By avg. rank

	average rank
fuzzer
dgfuzz_e0d055	1.43
aflplusplus	2.00
eclipser	4.57
mopt	5.29
aflsmart	5.43
fairfuzz	6.43
libafl	6.71
aflfast	7.00
centipede	7.71
honggfuzz	8.43
libfuzzer	8.71
afl	9.00

Critical difference diagram

The diagram visualizes the average rank of fuzzers (second ranking above) while showing the significance of the differences as well. What is considered a "critical difference" (CD) is based on the Friedman/Nemenyi post-hoc test. See more in the documentation.

Note: If a fuzzer does not support all benchmarks, its ranking as shown in this diagram can be lower than it should be. So please check the list of supported benchmarks for the fuzzer(s) of your interest. The list could be specified in the fuzzer's README.md like this.

Median relative code-coverages on each benchmark

Note: The relative coverage summary table shows the median relative performance of each fuzzer to the experiment maximum. Thus the highest relative performance may not be 100%.
trial_relative_coverage = trial_coverage / experiment_max_coverage

	dgfuzz_e0d055	libafl	honggfuzz	libfuzzer	aflplusplus	afl	eclipser	centipede	mopt	aflsmart	fairfuzz	aflfast
FuzzerMedian	95.00	88.00	91.00	88.00	89.00	87.00	75.00	75.50	71.00	71.00	71.00	71.00
FuzzerMean	93.29	91.33	91.00	88.00	87.71	87.00	82.00	71.17	67.14	65.86	65.71	63.71
harfbuzz_hb-shape-fuzzer	98.00	nan	nan	nan	98.00	nan	97.00	nan	97.00	97.00	84.00	96.00
lcms_cms_transform_fuzzer	95.00	nan	nan	nan	89.00	nan	75.00	35.00	50.00	40.00	54.00	28.00
libpcap_fuzz_both	84.00	nan	nan	nan	88.00	nan	71.00	81.00	0.00	0.00	0.00	1.00
mbedtls_fuzz_dtlsclient	95.00	88.00	nan	nan	70.00	nan	68.00	66.00	68.00	68.00	71.00	66.00
openthread_ot-ip6-send-fuzzer	87.00	88.00	nan	nan	75.00	nan	74.00	70.00	71.00	71.00	68.00	71.00
stb_stbi_read_fuzzer	95.00	nan	91.00	88.00	95.00	87.00	91.00	85.00	86.00	87.00	86.00	86.00
vorbis_decode_fuzzer	99.00	98.00	nan	nan	99.00	nan	98.00	90.00	98.00	98.00	97.00	98.00

Fuzzers are sorted by "FuzzerMean" (average median relative coverage), highest on the left.
Green background = highest relative median coverage.
Blue gradient background = greater than 95% relative median coverage.

harfbuzz_hb-shape-fuzzer summary

Ranking by median reached code coverage

Mean code coverage growth over time

* The error bands show the 95% confidence interval around the mean code coverage.

The following fuzzers do not have enough samples: aflplusplus, fairfuzz, dgfuzz_e0d055.

Sample statistics and statistical significance (code coverage)

Code coverage sample statistics

		count	mean	std	min	25%	median	75%	max
fuzzer	time
dgfuzz_e0d055	82800	7.0	10863.428571	276.618183	10327.0	10783.50	10959.0	11060.00	11071.0
aflplusplus	82800	14.0	10909.214286	40.828306	10824.0	10876.75	10912.5	10939.00	10967.0
mopt	82800	20.0	10778.950000	52.249880	10651.0	10757.50	10789.5	10796.25	10863.0
eclipser	82800	20.0	10769.400000	51.530676	10623.0	10744.25	10777.0	10789.50	10868.0
aflsmart	82800	20.0	10763.700000	42.521945	10652.0	10753.00	10764.5	10791.00	10839.0
aflfast	82800	20.0	10677.750000	59.487172	10566.0	10633.00	10675.5	10714.50	10791.0
fairfuzz	82800	14.0	9509.571429	347.646530	9071.0	9259.50	9399.0	9715.75	10155.0

Vargha-Delaney A12 measure

The table summarizes the A12 values from the pairwise Vargha-Delaney A measure of effect size. Green cells indicate the probability the fuzzer in the row will outperform the fuzzer in the column.

Mann-Whitney U test

The table summarizes the p values of pairwise Mann-Whitney U tests. Green cells indicate that the reached coverage distribution of a given fuzzer pair is significantly different.

Unique code coverage plots

Ranking by unique code branches covered
Each bar shows the total number of code branches found by a given fuzzer. The colored area shows the number of unique code branches (i.e., branches that were not covered by any other fuzzers).

Pairwise unique code coverage
Each cell represents the number of code branches covered by the fuzzer of the column but not by the fuzzer of the row

Code coverage reports for each fuzzer on this benchmark

afl aflfast aflplusplus aflsmart centipede dgfuzz_e0d055 eclipser fairfuzz honggfuzz libafl libfuzzer mopt