cfv vs Jacksum

Will access .gz files as gzipped. Can create gzipped files.

-n will rename bad files based on --renameformat.

Cfv shows a progress bar when working so you know that the process is actually taking place.

Runs on all systems Python supports. It has been verified to work on Linux, FreeBSD, OpenBSD, NetBSD, Solaris, OS X, and Windows.

cfv offers 2 recursive modes: -r will make separate chksum files for each dir and -rr will create a single file.

Jacksum 3.6.0 supports 489 algorithms, including checksums, CRCs, XOFs, cryptographic, and non-cryptographic hash functions. Jacksum also supports the "Rocksoft (tm) Model CRC Algorithm" to customize CRCs.
See also here.

The integration of Jacksum (and Jacksum's cross platform GUI called HashGarten) are available for many file browsers on Linux, Windows, and macOS.

It finds all files that match (or don't match) the hash values in a precalculated hash set.
Example: jacksum -w log4j.hashes /

Hashes partitions and disks on Linux, Unix (e.g. macOS), and Microsoft Windows

It finds all algorithms to a checksum/CRC by using a brute force algorithm.

Can perform a file integrity check on a single file, on standard input channel or by passing data on the command line. Can performs a file integrity check against a pre-calculated hash set, and detect ok, failed, missing, and new files. Can even hash and verify NTFS Alternate Data Streams (ADS).

Supports multi-threading on multi-processor and multi-core computer systems, it calculates multiple hashes simultaneously, files are read only once, and the calculation load is distributed on the available cores. It also processes multiple files simultaneously, i.e. files are read in parallel.

Encodings for representing hash values are available: Hex (lower- and uppercase, optional: grouping bytes by a separator), Base16, Base32 (with and without padding), Base32hex (with and without padding), Base64 (with and without padding), Base64url (with and without padding), BubbleBabble, and z-base-32.

Comprehensive format options are available to get the output the user needs, e.g. create ed2k-links, magnet-links or create a compatible Solaris' pkgmap.

Both input and output can occur in 9 predefined standard formats (BSD-, BSD-reversed, GNU/Linux-, openssl-, openssl-reversed, and Solaris' digest style both tagged and untagged, SFV or FCIV) which works with any supported algorithm. GNU file name escaping is supported.

Allows the user to specify the character set for both input and output files, examples: UTF-8, UTF-16, UTF-16BE, UTF-16LE, UTF-32, UTF-32BE, UTF-32LE, GB18030 (Chinese standard), etc.

It can process file sizes up to 8 Exbibytes (= 8,000,000,000 Gibibytes), presupposed the operating system respectively the file system is large file aware, too.

It calculates hashes from NTFS Alternate Data Streams (ADS) on Microsoft Windows.

Windows, Linux, macOS, and any platform with an OpenJDK compatible Java Runtime Environment (JRE) or Java Development Kit (JDK) can run Jacksum, no recompilation required.

Recursive operations can be controlled by specifying a walking depth. Also the following of symbolic links on files and/or directories can be controlled, and it also detects file system cycles to avoids endless loops.

The entire SHA-2 famliy and SHA-3 algorithms are supported: SHA-[224,256,384,512], and SHA-512/[224,256] as in NIST FIPS 180-4, and SHA3-[224,256,384,512], SHAKE[128,256] as in NIST FIPS 202.

It gives you an added security against attacks on the hash function. Multiple cores are used if available.