File checksum using Adler32 algorithm.
This project aims to compare the performance, pros and cons of reading a file as a stream or with mapped memory.
Checksum reads a file to calculate its Adler32 checksum. We can select the reading method:
Open the file as a stream and read it into memory.
Map the file into memory.
autoreconf -vi
./configure [ OPTIONS ]
make
Feature
Description
--disable-assertTurn off assertions.
--enable-sanitizeEnable AddressSanitizer.
--enable-coverageEnable coverage test.
Option
Description
Range
Default
BLOCKSIZEStream reader block size.
Positive number.
65536
checksum [ -h ] [ -m ] [ file ]
If no file specified or file "-" is given, checksum reads stdin .
Option
Description
-cCompare two files.
-hShow help.
-mUse memory mapping method.
-rUse CRC-32 hash algorithm.
-sUse SHA-1 hash algorithm.
Method
500 MB file
1500 MB file
File streaming
6.817 s
Memory error.
Memory mapping
2.830 s
8.367 s
Method
500 MB file
1500 MB file
File streaming
2.366 s
7.483 s
Memory mapping
2.966 s
8.640 s
Method
500 MB file
1500 MB file
File streaming
0.682 s
2.076 s
Memory mapping
1.032 s
3.502 s
Analyzing file comparison times.
Best case: different files
Method
500 MB file
1500 MB file
File streaming
0.001 s
0.001 s
Memory mapping
1.073 s
3.436 s
Worst case: similar files
Method
500 MB file
1500 MB file
File streaming
0.482 s
1.845 s
Memory mapping
1.568 s
5.639 s
Comparing different hash algorithms with a 1500 MB file.
Algorithm
File streaming
Memory mapping
Adler-32
2.390 s
4.390 s
CRC-32
5.036 s
6.827 s
SHA-1
2.939 s
5.033 s
