fireice-uk
7fdaceee72
Add return value
2017-10-22 13:12:52 +01:00
fireice-uk
f13f70c28e
Implement pool-controlled nonce allocation
2017-10-22 13:12:52 +01:00
fireice-uk and GitHub
610f4f0fa7
Merge pull request #58 from psychocrypt/fix-cuda9SuffleWarning
...
fix CUDA 9 shuffle warning
2017-10-21 20:01:06 +01:00
fireice-uk and GitHub
da6ff1b9e9
Merge pull request #59 from psychocrypt/topic-addVoltaSupport
...
add NVIDIA Volta support
2017-10-21 19:46:19 +01:00
fireice-uk and GitHub
bd31e5b019
Merge pull request #60 from psychocrypt/fix-amdAssert
...
fix assert in AMD backend
2017-10-21 19:45:27 +01:00
psychocrypt
630bcffb9f
fix assert in AMD backend
...
fix bug announced in #52
If the miner is compiled in debug mode the assert condition results in a compile time error.
2017-10-20 21:25:08 +02:00
psychocrypt
d7209b21fd
add NVIDIA Volta support
...
add `sm_70` if CUDA 9+ is avalable
2017-10-20 21:17:08 +02:00
psychocrypt
18ba334391
fix CUDA 9 shuffle warning
...
use `__shffl_snyc` if CUDA 9+ is avalable
2017-10-20 21:12:50 +02:00
psychocrypt and GitHub
1b639b6eaf
Merge pull request #54 from fireice-uk/topic-fix
...
add order_fix for double hash
2017-10-18 20:14:58 +02:00
fireice-uk
67e135a317
add order_fix for double hash
2017-10-18 15:21:52 +01:00
fireice-uk and GitHub
98630910b1
Merge pull request #50 from fireice-uk/topic-hwloc
...
Make hwloc check capabilities and stop it from leaking mem
2017-10-18 12:59:59 +01:00
psychocrypt and GitHub
6701b0c202
Merge pull request #47 from fireice-uk/topic-re-add-uac-manifest
...
Re-add UAC manifest and WIN_UAC cmake option
2017-10-17 21:24:29 +02:00
fireice-uk and GitHub
c7c5824364
Merge pull request #43 from fireice-uk/fix-aff-alt-c
...
Affinity fix version C
2017-10-17 20:14:45 +01:00
fireice-uk
3ec7fa7c7a
make hwloc check capabilities and stop it from leaking mem
2017-10-17 20:04:56 +01:00
fireice-uk
4253e7a028
rm pin_thd_affinity
2017-10-17 19:44:18 +01:00
fireice-uk and GitHub
dcd176ada4
Merge pull request #44 from fireice-uk/fix-uninit-access
...
Make sure all singletons are set to null and make env a global ptr
2017-10-17 16:55:32 +01:00
fireice-uk
1fc87d737f
Re-add UAC manifest and WIN_UAC cmake option
2017-10-14 14:44:34 +01:00
psychocrypt and GitHub
ca7e27b784
Merge pull request #37 from fireice-uk/topic-remove-amd-gpu_num
...
remove the gpu_thread_num safety
2017-10-13 21:31:47 +02:00
psychocrypt and GitHub
cf5cb145a3
Merge pull request #39 from fireice-uk/topic-fix-warning
...
Keep consistency between class and struct
2017-10-13 21:29:42 +02:00
fireice-uk
76aa167847
doh! != instead of ==
2017-10-13 17:13:44 +01:00
fireice-uk
d59495110d
Make sure all singletons are set to null and make env a global ptr
2017-10-12 22:28:35 +01:00
fireice-uk
8bbdd45b9e
Affinity - fix C
2017-10-12 12:52:50 +01:00
fireice-uk
73482f3065
Affinity - fix B
2017-10-12 12:38:14 +01:00
fireice-uk
75b2c8ca7a
Keep consistency between class and struct
2017-10-12 02:02:50 +01:00
fireice-uk
632d0758d2
remove the gpu_thread_num safety
2017-10-12 00:23:20 +01:00
fireice-uk
329a6d9ad1
swap order
2017-10-12 00:02:43 +01:00
fireice-uk
da070d4259
print warning
2017-10-11 23:59:17 +01:00
fireice-uk
c897fc0451
cleanup and give a return from set_affinity
2017-10-11 23:53:31 +01:00
fireice-uk and GitHub
b40f937bf5
Merge pull request #35 from psychocrypt/fix-cuda7_5Compile
...
fix cuda 7.5 compile
2017-10-11 22:39:22 +01:00
fireice-uk and GitHub
fa5518309e
Merge pull request #28 from psychocrypt/fix-missingCUDAOpenCLCmakeOutput
...
fix CMake find package output
2017-10-11 22:38:27 +01:00
psychocrypt
8d313484e4
fix cuda 7.5 compile
...
fix error: `identifier "__builtin_ia32_monitorx" is undefined`
2017-10-11 21:25:48 +02:00
fireice-uk and GitHub
01b4460b95
Merge pull request #24 from psychocrypt/fix-amdEarlyReturn
...
remove early returns
2017-10-11 16:36:46 +01:00
fireice-uk and GitHub
0c332962d9
Merge pull request #27 from psychocrypt/topic-improveAmdAutoAdjust
...
improve AMD auto suggestion
2017-10-11 16:35:07 +01:00
psychocrypt and GitHub
984ddf9ebb
Merge pull request #31 from fireice-uk/topic-fix-typo
...
Minor doc fixes
2017-10-10 21:31:17 +02:00
psychocrypt
e52d04fb23
fix CMake find package output
...
fix #3
- remove QUIET from find package for OpenCL and CUDA
- remove pascal architecture from default CUDA_ARCH for CUDA < 8.0
2017-10-10 21:26:15 +02:00
psychocrypt
e0b93c2216
improve AMD auto suggestion
...
- use always a multiple of the number of compute units
- add more device information to the config file
2017-10-10 21:12:34 +02:00
psychocrypt
2c2d05e57a
remove early returns
...
Early returns within a kernel can have unexpected behavior, this strongly depends on the opencl
runtime compiler. To avoid errors all early returns are removed.
There is no negative effect(e.g.performance) if the threads stay alive up to the end of the kernel.
One source for the early return is: http://al-key-opencl.blogspot.de/2014/09/be-careful-not-to-mix-early-return-and.html
This pull request also fix a race condition where all fill the shared memory.
Also a bug introduced with #16 is fixed, because of the early return not was possible that
the last block works with an wrong initilized shared memory (result should be a wrong hash if the result target size is valid).
2017-10-10 21:04:37 +02:00
fireice-uk
204dded5e0
Minor doc fixes
2017-10-10 18:52:06 +01:00
fireice-uk and GitHub
e79debcfcb
Merge pull request #23 from psychocrypt/fix-possibleIllegalMemoryAccessAMD
...
avoid possible illegal memory access
2017-10-10 18:01:16 +01:00
fireice-uk and GitHub
9af8d7f6ad
Merge pull request #29 from psychocrypt/fix-travisOSXBuild
...
fix travis OSX build
2017-10-10 01:40:22 +01:00
psychocrypt
9ad1b04ffa
fix broken travis macosx compile
...
add `brew update` to avoid that `brew tap homebrew/science` crash
2017-10-08 22:33:50 +02:00
fireice-uk and GitHub
adb2e196c9
Merge pull request #26 from psychocrypt/fix-hashStatWrongTimer
...
fix wrong timer for stats
2017-10-07 00:52:42 +01:00
psychocrypt
73f75f95b7
fix wrong timer for stats
...
fix #21 , hash values randomly not available.
The reason is that I changed the lables but forget to change all timer.
A GPU needs sometimes ore than 2.5s to finish one hash round, therefore
the results are only visible randomly.
2017-10-06 22:52:54 +02:00
psychocrypt
611378eb96
avoid possible illegal memory access
...
Is is possible that the number of results are greater than the result output array,
in this case invalid memory can be access within the device and on the host side.
2017-10-06 21:50:51 +02:00
fireice-uk and GitHub
83752eccc0
Merge pull request #16 from psychocrypt/fix-invalidKernelCall
...
fix invalid kernel call parameter
2017-10-06 15:22:29 +01:00
psychocrypt and GitHub
62217a2488
Merge pull request #19 from fireice-uk/topic-crash-fix
...
Port the crash fix from xmr-stak-cpu
2017-10-06 10:18:17 +02:00
psychocrypt
65d47dbb31
add assert to validate number of global threads
...
check that the number of global threads is a multiple of the workgroup size
2017-10-05 22:22:30 +02:00
psychocrypt and GitHub
e96a761d5a
Merge pull request #12 from fireice-uk/topic-large-pages-win
...
Fix large pages on Windows and make them more user-friendly
2017-10-05 22:15:32 +02:00
fireice-uk and GitHub
34a87c64a1
Merge pull request #18 from fireice-uk/topic-larger-print-buf
...
Increase the scratchpad size to make sure we don't malform JSON
2017-10-05 15:24:57 +01:00
fireice-uk
cb65ae3f78
Port the crash fix from xmr-stak-cpu
2017-10-05 15:08:00 +01:00
fireice-uk