Jens Steube
a0eae9050c
OpenCL Runtime: Workaround JiT compiler deadlock on NVIDIA driver >= 465.89
2021-04-11 13:35:40 +02:00
jsteube
083d6e1fb2
Refresh dynloader management
2019-01-14 10:11:23 +01:00
Jens Steube
6f8b90a11a
Add support for latest OpenCL headers from Khronos
2018-10-14 12:39:11 +02:00
Rosen Penev
84cbad2ba8
Only Apple has a different cl.h location. Simplify a bit
2017-03-04 18:28:50 -08:00
Rosen Penev
96783ed7a3
Fix compilation for native Cygwin
2016-11-29 14:10:39 -08:00
jsteube
e63bc4d328
Query PCI-Express bus/device/function and store for later use
2016-11-05 20:23:05 +01:00
jsteube
27bec8be13
Get rid of logging.c and most exit() calls replaced with return() for true library usage
2016-10-09 22:41:55 +02:00
jsteube
a02f6f5ad6
Give all those anonymous structures a name, makes it easier to load the project into VS
2016-10-01 11:54:00 +02:00
jsteube
3daf0af480
Added docs/credits.txt
...
Added docs/team.txt
2016-09-11 22:20:15 +02:00
jsteube
9eb47153d4
Replace all #ifdef with #if defined (...) for convention
2016-09-07 22:29:57 +02:00
jsteube
25726e4f4e
Remove some (hopefully) redundant macros and compiler options
2016-09-07 20:30:14 +02:00
jsteube
8dd57b715b
Create real opencl.c to better distinguish between opencl helper functions and library handling functions
2016-09-07 15:13:50 +02:00
jsteube
3ed4bfd237
Move tuningdb specific functions into their own source file
2016-09-06 19:44:27 +02:00
jsteube
e4bf86dcbe
Replace #pragma once with #ifndef #define #endif logic because #pragma once does not work cleanly on OSX
2016-09-06 18:44:05 +02:00
jsteube
1902e6c2eb
Create real cpu_crc32.c and rp_cpu.c
2016-09-06 15:28:56 +02:00
jsteube
4b79ccd416
Preparation for easier integration of:
...
* https://github.com/hashcat/hashcat/pull/485
* https://github.com/hashcat/hashcat/pull/486
2016-09-05 21:47:26 +02:00
Jens Steube
6ba0eb8b39
Get rid of exit() calls in OpenCL wrapper library with the goal to have a better control which error can be ignored under special circumstances
...
Note, not all return codes from some of the subfunctions are uset yet, it would be too much changes at once
This is a preparation for https://github.com/hashcat/hashcat/issues/416
2016-07-10 13:23:06 +02:00
jsteube
49f693a1c9
Replace LINUX macro with compiler predefined macro __linux__
2016-07-08 23:23:16 +02:00
jsteube
96f2ade4bb
Replace DARWIN macro with __APPLE__
...
As discussed on https://github.com/hashcat/hashcat/pull/409
2016-07-08 23:14:57 +02:00
Nikolai Lifanov
a2663fb93a
include the right cl.h on FreeBSD
2016-07-05 15:12:05 -04:00
Unix-Ninja
2188a01ae9
Refactor OSX preprocessor constants to read 'DARWIN'
2016-06-27 08:13:50 -04:00
jsteube
829289b109
Fix calling conventions for external libraries (mingw doesn't handle it alone); Fix first print of PROMPT after autotune;
2016-06-04 00:54:28 +02:00
Gabriele 'matrix' Gristina
2a1d0d21ff
Get rid of hc_clEnqueueNDRangeKernel double call
2016-02-12 17:11:23 +01:00
jsteube
5c01349ba6
Update event handling to workaround event handling error in nvidia opencl runtime
2016-02-11 09:54:50 +01:00
Jens Steube
9fc360e39d
Added the execution time of the running kernel to the status display
2016-02-09 20:01:50 +01:00
Gabriele 'matrix' Gristina
74be8e2e25
skip device if buildProgram() fail
2016-02-07 21:20:10 +01:00
Gabriele 'matrix' Gristina
89e5ed2574
handling possible clEnqueueFillBuffer not found in opencl lib
2016-02-04 21:06:19 +01:00
Gabriele 'matrix' Gristina
d120ede655
Update ADL Library Loader
2016-01-31 23:48:38 +01:00
Gabriele 'matrix' Gristina
f9598f691d
Fixed CL_INVALID_WORK_GROUP_SIZE error with Apple CPU
2016-01-30 22:28:41 +01:00
Gabriele 'matrix' Gristina
d147d89f5e
cleanup unnecessary clGetProgramBuildInfo code
2016-01-30 20:05:39 +01:00
Gabriele 'matrix' Gristina
5bae9de3a3
Implemented OpenCL library loader
2016-01-26 21:40:49 +01:00
Gabriele 'matrix' Gristina
5da79e4411
Added support for build without ADL/NVML/NVAPI (issue #120 )
...
Added support for build OSX native binaries (issue #63 )
2016-01-20 20:55:09 +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
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
philsmd
ef548d3088
fixes issue #5 : formatting problem with tabs vs spaces
2015-12-06 11:29:29 +01:00
Jens Steube
5065474b4e
Initial commit
2015-12-04 15:47:52 +01:00
