jsteube
7807eb6f2d
Fix -m 7800 for NV
2015-12-30 00:29:39 +01:00
jsteube
77df413886
Add allocatable memory location per GPU
2015-12-28 11:29:07 +01:00
jsteube
a78173af7e
prepare scrypt tmto value for manual tuning
2015-12-28 09:20:18 +01:00
philsmd
286bde422b
limit the salt length of -m 22 = Juniper Netscreen/SSG (ScreenOS) to 10
2015-12-27 09:48:52 +01:00
jsteube
3dc8d526db
Update timeout-patch url to wiki
2015-12-24 10:32:22 +01:00
jsteube
f1cdf15540
Added option --gpu-platform to select a single OpenCL platform in case multiple OpenCL platforms are present
2015-12-23 18:02:01 +01:00
jsteube
c17bf5e865
- Fixed PHY memory handling for scrypt based algorithms
...
- Bring back kernel exec timeout checking for NV
2015-12-23 15:51:55 +01:00
jsteube
19c77034a6
Updated algorithm used to automatically select an ideal --scrypt-tmto value for NV
2015-12-23 11:24:41 +01:00
philsmd
0d905317a6
fix OpenCL memory problem (exhaustion): -4 error should be gone
2015-12-22 19:03:35 +01:00
jsteube
7be2c2fd8d
Fixed some speeds:
...
1800
3200
7100
7200
7400
8200
9300
9000
9600
11300
11600
2015-12-21 14:29:04 +01:00
jsteube
e2412c9b10
We do not depend on CUDA any longer, so get rid of it :)
2015-12-21 12:20:08 +01:00
jsteube
378258d789
Fix caching system for use with AMD and NV, drop BINARY_KERNEL define
2015-12-21 12:01:38 +01:00
jsteube
c8f7b7b5d3
Workaround missing clEnqueueFillBuffer() support in NVidia's OpenCL runtime
2015-12-21 10:17:54 +01:00
jsteube
1d4bece384
Revert "Fix bsdicrypt for NV"
...
This reverts commit 284990e4d3
2015-12-20 22:08:59 +01:00
jsteube
284990e4d3
Fix bsdicrypt for NV
2015-12-20 22:07:44 +01:00
jsteube
52c416ba72
Fixed empty line problem
2015-12-16 17:01:14 +01:00
jsteube
80f86c52a1
Cleanup makefile
...
Remove unused CUDA library defs
2015-12-15 19:18:45 +01:00
jsteube
4a0c47ca08
It's safe now to use AMD's OpenCL includes, tested on NV
2015-12-15 18:58:02 +01:00
philsmd
cff683eac1
change in library loading: libnvidia-ml now does load dynamically only on NVidia systems
2015-12-15 20:34:07 +01:00
jsteube
2283d5c843
Fix more append_* functions in kernels
2015-12-15 16:50:21 +01:00
philsmd
3ab7a23cd5
fixed hardware monitor for amd/nv (one problem that is still there since last commit: dynamic loading of libnvidia-ml)
2015-12-15 18:41:11 +01:00
jsteube
0bf4e3c34a
- Dropped all vector code since new GPU's are all scalar, makes the code much easier
...
- Some performance on low-end GPU may drop because of that, but only for a few hash-modes
- Dropped scalar code (aka warp) since we do not have any vector datatypes anymore
- Renamed C++ overloading functions memcat32_9 -> memcat_c32_w4x4_a3x4
- Still need to fix kernels to new function names, needs to be done manually
- Temperature Management needs to be rewritten partially because of conflicting datatypes names
- Added code to create different codepaths for NV on AMD in runtime in host (see data.vendor_id)
- Added code to create different codepaths for NV on AMD in runtime in kernels (see IS_NV and IS_AMD)
- First tests working for -m 0, for example
- Great performance increases in general for NV so far
- Tested amp_* and markov_* kernel
- Migrated special NV optimizations for rule processor
2015-12-15 12:04:22 +01:00
Jens Steube
586441fa25
Revert "Fixed a bug where oclHashcat rejected to load a rule which calls 15 functions although it is supported"
...
This reverts commit f230ed73dc
2015-12-13 18:46:21 +01:00
Jens Steube
f230ed73dc
Fixed a bug where oclHashcat rejected to load a rule which calls 15 functions although it is supported
2015-12-13 14:54:39 +01:00
Jens Steube
1537390fae
Prepare reordering or files to help integration into linux distributions as discussed in https://github.com/hashcat/oclHashcat/issues/20
...
TODOS:
- Let oclHashcat actually use the new paths
- Find a better way for native compilation
- Replace /bin/cp with /usr/bin/install where it has to copy files recursive
2015-12-13 12:21:36 +01:00
Jens Steube
c460b554d0
- Fixed a bug related to --restore if the User
...
- Updated readme, we actually need catalyst 15.7 (for the binary package) because of OpenCL 2.0
2015-12-11 19:27:07 +01:00
philsmd
c15f0d6146
issue #16 : use just "deps" instead of "hashcat-deps" as directory name
2015-12-09 15:55:06 +01:00
philsmd
29547c4f9b
issue #14 : move /opt/hashcat-deps/ to [git_clone_dir]/hascat-deps/
2015-12-09 13:33:21 +01:00
philsmd
feeb44882b
allow both enabling and disabling of restore point update (implemented with issue #10 )
2015-12-07 23:51:51 +01:00
jsteube
968265fffb
- Prepared for JIT use of hash-mode 1500, 8900 and 9300, works already on OpenCL (AMD)
...
- Changed PROMPT
2015-12-07 21:37:12 +01:00
philsmd
946b4a37d0
implements the enhancement mentioned in issue #10 (cancel on next checkpoint)
2015-12-07 18:31:45 +01:00
philsmd
ffcf763b13
typo fixed in Makefile
2015-12-06 18:48:20 +01:00
philsmd
3f79ac5964
Merge remote-tracking branch 'upstream/master'
2015-12-06 16:12:25 +01:00
Jens Steube
ab66fa1772
Revert "Remove NPROCS from Makefile, make is able to automatically detect the optimal number of parallel threads"
...
Problem here is; This process requires alot of memory. If host memory is too small it will segfault somewhere inside AMD's OpenCL runtime library. Therefore it's better to stick to NPROCS as part of a workaround.
This reverts commit e5adccbf38
2015-12-06 13:09:17 +01:00
philsmd
ef548d3088
fixes issue #5 : formatting problem with tabs vs spaces
2015-12-06 11:29:29 +01:00
Jens Steube
e5adccbf38
Remove NPROCS from Makefile, make is able to automatically detect the optimal number of parallel threads
2015-12-05 14:30:09 +01:00
Jens Steube
490ff03fe1
It's important to release a fixed version of v2.00 because of the issue #1 even if AMD is not affected
2015-12-05 13:42:02 +01:00
philsmd
9684d8793a
fixes issue #1 : cuMemsetD8() 1 error (nvidia only, problem did not affect mask attacks)
2015-12-05 08:48:00 +01:00
Jens Steube
5065474b4e
Initial commit
2015-12-04 15:47:52 +01:00
