Generate a 1000M test file, the file content is all 0 (because it is read from /dev/zero, /dev/zero is the source of 0). But this is for actually writing to the hard disk. The speed of file generation depends on the read and write speed of the hard disk. If you want to generate very large files, the speed is very slow.

[root@tx]# dd if=/dev/zero of=testfile.txt bs=1M count=1000
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB, 1000 MiB) copied, 1.04506 s, 1.0 GB/s #Memory operation speed

In a certain scenario, we just want the file system to think that there is a very large file here, but we don't actually write it to the hard disk. then you can use seek

1、count=0 means reading and writing 0 times, specifying the size of the generated file to be 0M

[root@tx]# dd if=/dev/zero of=testfile2.txt bs=1M count=0 seek=100000
0+0 records in
0+0 records out
0 bytes copied, 0.000169502 s, 0.0 kB/s

2、count=50 means read and write 50 times, and specify the size of the generated file to be 50M

[root@tx]# dd if=/dev/zero of=testfile3.txt bs=1M count=50 seek=100000
50+0 records in
50+0 records out
52428800 bytes (52 MB, 50 MiB) copied, 0.0457615 s, 1.1 GB/s

The displayed size of the file created at this time in the file system is 100000MB, but it does not actually occupy a block, so the creation speed is equivalent to the memory speed.

[root@tx]# ls -l
-rw-r--r-- 1 root root 104857600000 Nov 7 13:20 testfile2.txt
-rw-r--r-- 1 root root 104910028800 Nov 7 13:21 testfile3.txt
-rw-r--r-- 1 root root 1048576000 Nov 7 13:20 testfile.txt

# command parsing

if (input file) represents the input file/directory, if not specified, the input will be read from stdin by default of (output file) represents the output file/directory. If of is not specified, stdout will be used as the default output by default bs represents the block size read in each time count: the number of blocks read bs*count=file size /dev/zero is a special character device provided by Linux. Its characteristic is that the file can be read forever, and the result of each read is binary 0 seek: Skips a section of the specified size in the output file without actually writing