I'm looking for an implementation of CRC32 in C or C++ that is explicitly licensed as being no cost or public domain. The implementation here seems nice, but the only thing it says about the license is "source code", which isn't good enough. I'd prefer non LGPL so I don't have to fool around with a DLL (my app is closed source). I saw the adler32 implementation in zlib, but I'm checking small chunks of data, which adler is not good for.
The SNIPPETS C Source Code Archive has a CRC32 implementation that is freely usable:
/* Copyright (C) 1986 Gary S. Brown. You may use this program, or
code or tables extracted from it, as desired without restriction.*/
(Unfortunately, c.snippets.org seems to have died. Fortunately, the Wayback Machine has it archived.)
In order to be able to compile the code, you'll need to add typedefs for BYTE as an unsigned 8-bit integer and DWORD as an unsigned 32-bit integer, along with the header files crc.h & sniptype.h.
The only critical item in the header is this macro (which could just as easily go in CRC_32.c itself:
#define UPDC32(octet, crc) (crc_32_tab[((crc) ^ (octet)) & 0xff] ^ ((crc) >> 8))
The crc code in zlib (http://zlib.net/) is among the fastest there is, and has a very liberal open source license.
And you should not use adler-32 except for special applications where speed is more important than error detection performance.
The mhash library works pretty good for me. It's fast enough, supports multiple types of hashing (crc32, MD5, SHA-1, HAVAL, RIPEMD128, RIPEMD160, TIGER, GOST, etc.). To get CRC32 of a string you would do something like this:
MHASH td = mhash_init(MHASH_CRC32);
if (td == MHASH_FAILED) return -1; // handle failure
mhash(td, s, strlen(s));
unsigned int digest = 0; // crc32 will be stored here
mhash_deinit(td, &digest);
// do endian swap here if desired
I am the author of the source code at the specified link. While the intention of the source code license is not clear (it will be later today), the code is in fact open and free for use in your free or commercial applications with no strings attached.
using zlib.h (http://refspecs.linuxbase.org/LSB_3.0.0/LSB-Core-generic/LSB-Core-generic/zlib-crc32-1.html):
#include <zlib.h>
unsigned long crc = crc32(0L, Z_NULL, 0);
crc = crc32(crc, (const unsigned char*)data_address, data_len);
The SNIPPETS C Source Code Archive has a CRC32 implementation that is freely usable:
/* Copyright (C) 1986 Gary S. Brown. You may use this program, or
code or tables extracted from it, as desired without restriction.*/
(Unfortunately, c.snippets.org seems to have died. Fortunately, the Wayback Machine has it archived.)
In order to be able to compile the code, you'll need to add typedefs for BYTE as an unsigned 8-bit integer and DWORD as an unsigned 32-bit integer, along with the header files crc.h & sniptype.h.
The only critical item in the header is this macro (which could just as easily go in CRC_32.c itself:
#define UPDC32(octet, crc) (crc_32_tab[((crc) ^ (octet)) & 0xff] ^ ((crc) >> 8))
I am the author of the source code at the specified link. While the intention of the source code license is not clear (it will be later today), the code is in fact open and free for use in your free or commercial applications with no strings attached.
rurban's fork of SMHasher (the original SMHasher seems abandoned) has hardware CRC32 support. The changes were added before the initial commit, but try comparing the new CMakeLists.txt and the old one (which doesn't mention SSE at all).
The best option is probably Intel's zlib fork with PCLMULQDQ support described in this paper. This library also has the SSE 4.2 optimizations.
If you don't need portability and you're on Linux, you can use the kernel's implementation (which is hardware accelerated if available): https://stackoverflow.com/a/11156040/309483
The SNIPPETS C Source Code Archive has a CRC32 implementation that is freely usable:
/* Copyright (C) 1986 Gary S. Brown. You may use this program, or
code or tables extracted from it, as desired without restriction.*/
(Unfortunately, c.snippets.org seems to have died. Fortunately, the Wayback Machine has it archived.)
In order to be able to compile the code, you'll need to add typedefs for BYTE as an unsigned 8-bit integer and DWORD as an unsigned 32-bit integer, along with the header files crc.h & sniptype.h.
The only critical item in the header is this macro (which could just as easily go in CRC_32.c itself:
#define UPDC32(octet, crc) (crc_32_tab[((crc) ^ (octet)) & 0xff] ^ ((crc) >> 8))
pycrc is a Python script that generates C CRC code, with options to select the CRC size, algorithm and model.
It's released under the MIT licence. Is that acceptable for your purposes?
The most simple and straightforward C/C++ implementation that I found is in a link at the bottom of this page:
Web Page: http://www.barrgroup.com/Embedded-Systems/How-To/CRC-Calculation-C-Code
Code Download Link: https://barrgroup.com/code/crc.zip
It is a simple standalone implementation with one .h and one .c file. There is support for CRC32, CRC16 and CRC_CCITT thru the use of a define. Also, the code lets the user change parameter settings like the CRC polynomial, initial/final XOR value, and reflection options if you so desire.
The license is not explicitly defined ala LGPL or similar. However the site does say that they are placing the code in the public domain for any use. The actual code files also say this.
Hope it helps!
rurban's fork of SMHasher (the original SMHasher seems abandoned) has hardware CRC32 support. The changes were added before the initial commit, but try comparing the new CMakeLists.txt and the old one (which doesn't mention SSE at all).
The best option is probably Intel's zlib fork with PCLMULQDQ support described in this paper. This library also has the SSE 4.2 optimizations.
If you don't need portability and you're on Linux, you can use the kernel's implementation (which is hardware accelerated if available): https://stackoverflow.com/a/11156040/309483
The mhash library works pretty good for me. It's fast enough, supports multiple types of hashing (crc32, MD5, SHA-1, HAVAL, RIPEMD128, RIPEMD160, TIGER, GOST, etc.). To get CRC32 of a string you would do something like this:
MHASH td = mhash_init(MHASH_CRC32);
if (td == MHASH_FAILED) return -1; // handle failure
mhash(td, s, strlen(s));
unsigned int digest = 0; // crc32 will be stored here
mhash_deinit(td, &digest);
// do endian swap here if desired
The most simple and straightforward C/C++ implementation that I found is in a link at the bottom of this page:
Web Page: http://www.barrgroup.com/Embedded-Systems/How-To/CRC-Calculation-C-Code
Code Download Link: https://barrgroup.com/code/crc.zip
It is a simple standalone implementation with one .h and one .c file. There is support for CRC32, CRC16 and CRC_CCITT thru the use of a define. Also, the code lets the user change parameter settings like the CRC polynomial, initial/final XOR value, and reflection options if you so desire.
The license is not explicitly defined ala LGPL or similar. However the site does say that they are placing the code in the public domain for any use. The actual code files also say this.
Hope it helps!
The crc code in zlib (http://zlib.net/) is among the fastest there is, and has a very liberal open source license.
And you should not use adler-32 except for special applications where speed is more important than error detection performance.
using zlib.h (http://refspecs.linuxbase.org/LSB_3.0.0/LSB-Core-generic/LSB-Core-generic/zlib-crc32-1.html):
#include <zlib.h>
unsigned long crc = crc32(0L, Z_NULL, 0);
crc = crc32(crc, (const unsigned char*)data_address, data_len);
I am the author of the source code at the specified link. While the intention of the source code license is not clear (it will be later today), the code is in fact open and free for use in your free or commercial applications with no strings attached.
pycrc is a Python script that generates C CRC code, with options to select the CRC size, algorithm and model.
It's released under the MIT licence. Is that acceptable for your purposes?
I am the author of the source code at the specified link. While the intention of the source code license is not clear (it will be later today), the code is in fact open and free for use in your free or commercial applications with no strings attached.
Source: Stackoverflow.com