| Updated the description of -jumbo. |
|
More...
|
about 13 years ago
|
| Documented support for DES-based tripcodes. |
|
More...
|
about 13 years ago
|
| Documented OpenMP parallelization of bitslice DES. |
|
More...
|
about 13 years ago
|
| Documented OpenMP parallelization of bitslice DES. Made edits to other recent change log entries. |
|
More...
|
about 13 years ago
|
| Disable DES_BS_ASM when OpenMP is requested and SIMD intrinsics are available. |
|
More...
|
about 13 years ago
|
| In DES_bs_cmp_all(), process only the required number of thread slots. |
|
More...
|
about 13 years ago
|
| Moved DES_bs_set_salt() from DES_bs.c to DES_bs_b.c and into parallel regions. |
|
More...
|
about 13 years ago
|
| Dropped the union with "double dummy;" approach to aligning buffer[] elements. |
|
More...
|
about 13 years ago
|
| Allocate "buffer" dynamically. |
|
More...
|
about 13 years ago
|
| OpenMP parallelization of bitslice DES. |
|
More...
|
about 13 years ago
|
| Only set FMT_OMP if BF_mt indicates so. Although FMT_OMP is conditional on _OPENMP by itself, the check of BF_mt makes sense for the special case when BF_ASM is set, which inhibits OpenMP parallelization of the Blowfish code. |
|
More...
|
about 13 years ago
|
| Added keys_count parameter to DES_bs_crypt*() (to be made use of when max_keys_per_crypt exceeds DES_BS_DEPTH, but the actual keys_count might be lower than max). |
|
More...
|
about 13 years ago
|
| Corrected an error in one of the definitions of vshl1(). |
|
More...
|
about 13 years ago
|
| Include -funroll-loops in OPT_INLINE. |
|
More...
|
about 13 years ago
|
| 1.7.8.3 |
|
More...
|
about 13 years ago
|
| Merged DES_bs_finalize_keys*() into DES_bs_crypt*(), moved some code around in DES_bs_b.c. |
|
More...
|
about 13 years ago
|
| Introduced vshl1(). |
|
More...
|
about 13 years ago
|
| Defined vshl() and vshr() for many vector types, excluding AltiVec for now. |
|
More...
|
about 13 years ago
|
| Allow DES_bs_finalize_keys* not to be vectorized even when DES_bs_crypt* are. This is desirable to better support platforms for which we do not (yet) have fast enough vshl() and vshr() macros. |
|
More...
|
about 13 years ago
|
| Moved DES_bs_finalize_keys*() from DES_bs.c to DES_bs_b.c, converted to use of v*() macros. |
|
More...
|
about 13 years ago
|
| Do not #undef DES_BS_VECTOR - instead, #define DES_BS_VECTOR_LOOPS to 0 if our v*() macros process full vectors. |
|
More...
|
about 13 years ago
|
| Added #if 0'ed section with non-vectorized bitslice DES settings (for testing). |
|
More...
|
about 13 years ago
|
| Added DES_bs_all.masks and mask* macros in preparation for moving DES_bs_finalize_keys*() from DES_bs.c to DES_bs_b.c (and to use of vectors). |
|
More...
|
about 13 years ago
|
| Removed the #warning about use of SSE2 intrinsics with recent gcc. |
|
More...
|
about 13 years ago
|
| Restricted the scope of the "zero" and "ones" variables inside *crypt*(). |
|
More...
|
about 13 years ago
|
| Define "zero" in the same way as we define "ones". Use both in the *crypt*() functions. |
|
More...
|
about 13 years ago
|
| Simplified the definitions of vnot(). |
|
More...
|
about 13 years ago
|
| Properly distinguish the two 256-bit SSE2-based virtual vector types. |
|
More...
|
about 13 years ago
|
| Restored support for weird vector widths that are not powers of two. (This was broken with recent commits.) |
|
More...
|
about 13 years ago
|
| Corrected the iteration count for DES_bs_finalize_keys_LM_loop (7, not 8). |
|
More...
|
about 13 years ago
|
