Tree: 19d7bc1232

gh-pages

master

stdsimd

0.1.0

0.2.0

0.2.1

0.2.10

0.2.11

0.2.12

0.2.13

0.2.14

0.2.15

0.2.16

0.2.17

0.2.18

0.2.2

0.2.3

0.2.4

0.2.5

0.2.6

0.2.7

0.2.8

0.2.9

Branches Tags

${ item.name }

Create tag ${ searchTerm }

Create branch ${ searchTerm }

from '19d7bc1232'

${ noResults }

Commit Graph

18 Commits (19d7bc1232b1a2990702889516ba6e4a33623330)

 

All Branches   

Search

Author	SHA1	Message	Date
Cesar Eduardo Barros	19d7bc1232	When SIMD is enabled, use the SSE2 code for x86 When enabled, makes BLAKE2b ~20% faster and BLAKE2s ~60% faster on 32-bit x86.	9 years ago
Cesar Eduardo Barros	aaccc40e40	Bump version to 0.2.4	9 years ago
Cesar Eduardo Barros	c6c237116f	Make it possible to use advanced BLAKE2 features	9 years ago
Cesar Eduardo Barros	238fa7b09f	Simplify finalize code No need to write back self.h when self is about to be dropped.	9 years ago
Cesar Eduardo Barros	e6d70af3b7	Treat IV as constant vector	9 years ago
Cesar Eduardo Barros	1fdfa6310a	Avoid abstraction penalty in the SIMD code It seems that the compiler is not smart enough to avoid the abstraction penalty in the 32-bit case. Passing the vector directly instead of encapsulating it inside a struct made the SIMD Blake2s code now very slightly *faster* than the non-SIMD Blake2s code.	9 years ago
Cesar Eduardo Barros	63d8a42ebe	Attempt to vectorize the code Go all the way and manually vectorize the code, using a generic 4-element vector type, and add both a fallback implementation and an attempt at a SSE2 implementation using LLVM intrinsics. The experiment was not successful; the SIMD code for some reason is much slower. Interestingly, however, the fallback code was much faster for Blake2s, and only slightly slower for Blake2b. Now both implementations are around 80% of the speed of the SIMD-optimized reference libb2 code.	9 years ago
Cesar Eduardo Barros	7390881973	Bump version to 0.2.2	9 years ago
Cesar Eduardo Barros	5612c4455c	Try to help the autovectorizer Change the order of the round operations to make more parallelism visible to the compiler's autovectorizer. The results were mixed; on my 64-bit machine, it made Blake2b 14% faster, but Blake2s became 3% slower.	9 years ago
Cesar Eduardo Barros	a971812a6f	Add bench feature	9 years ago
Cesar Eduardo Barros	90e83febaf	Remove redundant mask	9 years ago
Cesar Eduardo Barros	d6ec2ff38f	Derive Debug for state	9 years ago
Cesar Eduardo Barros	f727ab6a78	Add .travis.yml	9 years ago
Cesar Eduardo Barros	3fbc2dec25	Bump version to 0.2.0	9 years ago
Cesar Eduardo Barros	ce889c001a	Implement Write	9 years ago
Cesar Eduardo Barros	a809ebeeaa	Fix PartialEq	9 years ago
Cesar Eduardo Barros	8046be2088	"a maximum of 5 keywords per crate are allowed"	9 years ago
Cesar Eduardo Barros	93d1460cc7	Initial commit	9 years ago