- 2-10% speed-up on most modern GPUs
- AESNI bug on stored files really fixed
- CUDA version is obsolete because OpenCL speed is the same or better
- a little faster on newest NVIDIA GPUs (Maxwell, Pascal).
AMD Polaris support.
- 64-bit version, 32-bit is now obsolete. Up to 5-20% speed-up on
different archive types.
- NVIDIA Maxwell support.
- Intel GPU suppurt temporarily removed due to a lot of errors
in Intel OpenCL driver.
- WinXP support fixed. But future WinXP support is not
- Haswell AVX2 support, 2 times faster, CPU rate up to 4.50!
- Linking CUDA 5.5 surprisingly resulted in 30%-50% speed-up
on Fermi and Kepler cards
- Intel HD grapichs OpenCL support
- AMD XOP code enhanced
- -n option to specify GPU device number
- serious bug (password not found) fixed
- correct recognition of AMD APUs
- non-ASCII passwords are now working
- fixed OpenCL AMD GCN bug with -d2 option
- Kepler tuning
- GPU rate and grid size are calculated more
- due to a lot of OpenCL compiler errors, the self
test is included. (OpenCL version only)
- serious bug (false negative) fixed
- up to 3% brute-force attack speed-up on CPU,
up to 10% dictionary attack speed-up
- GPU support using CUDA and OpenCL, including
Kepler and AMD GCN GPUs
- more accurate CPU rate and SHA-256 clocks. Turn
Turbo Boost off when to get correct results!
- AMD FX (Bulldozer) code added
- First version released.