From 594a5b4d5b515af2b4f66cf940c10e103ceee40a Mon Sep 17 00:00:00 2001
From: psychocrypt <psychocryptHPC@gmail.com>
Date: Mon, 8 Oct 2018 21:43:25 +0200
Subject: [PATCH] CUDA: add compatibility mode
Add compatibility mode for CUDA to avoid invalid shares.
---
xmrstak/backend/nvidia/autoAdjust.hpp | 1 +
xmrstak/backend/nvidia/config.tpl | 3 +
xmrstak/backend/nvidia/jconf.cpp | 14 ++-
xmrstak/backend/nvidia/jconf.hpp | 1 +
xmrstak/backend/nvidia/minethd.cpp | 1 +
.../backend/nvidia/nvcc_code/cryptonight.hpp | 1 +
xmrstak/backend/nvidia/nvcc_code/cuda_core.cu | 109 ++++++++++++++----
7 files changed, 105 insertions(+), 25 deletions(-)
diff --git a/xmrstak/backend/nvidia/autoAdjust.hpp b/xmrstak/backend/nvidia/autoAdjust.hpp
index 1246809..6354f60 100644
--- a/xmrstak/backend/nvidia/autoAdjust.hpp
+++ b/xmrstak/backend/nvidia/autoAdjust.hpp
@@ -96,6 +96,7 @@ private:
" \"threads\" : " + std::to_string(ctx.device_threads) + ", \"blocks\" : " + std::to_string(ctx.device_blocks) + ",\n" +
" \"bfactor\" : " + std::to_string(ctx.device_bfactor) + ", \"bsleep\" : " + std::to_string(ctx.device_bsleep) + ",\n" +
" \"affine_to_cpu\" : false, \"sync_mode\" : 3,\n" +
+ " \"comp_mode\" : true,\n" +
" },\n";
}
}
diff --git a/xmrstak/backend/nvidia/config.tpl b/xmrstak/backend/nvidia/config.tpl
index 144da80..e2a76d9 100644
--- a/xmrstak/backend/nvidia/config.tpl
+++ b/xmrstak/backend/nvidia/config.tpl
@@ -16,6 +16,9 @@ R"===(// generated by XMRSTAK_VERSION
* 1 = cudaDeviceScheduleSpin - create a high load on one cpu thread per gpu
* 2 = cudaDeviceScheduleYield
* 3 = cudaDeviceScheduleBlockingSync (default)
+ * comp_mode - Compatibility if true it will use 64bit memory loads and if false it will use
+ * 128bit memory loads (can produce invalid results)
+ * (this option has only a meaning for cryptonight_v8 and monero)
*
* On the first run the miner will look at your system and suggest a basic configuration that will work,
* you can try to tweak it from there to get the best performance.
diff --git a/xmrstak/backend/nvidia/jconf.cpp b/xmrstak/backend/nvidia/jconf.cpp
index c9d4f19..b1059f3 100644
--- a/xmrstak/backend/nvidia/jconf.cpp
+++ b/xmrstak/backend/nvidia/jconf.cpp
@@ -123,7 +123,7 @@ bool jconf::GetGPUThreadConfig(size_t id, thd_cfg &cfg)
if(!oThdConf.IsObject())
return false;
- const Value *gid, *blocks, *threads, *bfactor, *bsleep, *aff, *syncMode;
+ const Value *gid, *blocks, *threads, *bfactor, *bsleep, *aff, *syncMode, *compMode;
gid = GetObjectMember(oThdConf, "index");
blocks = GetObjectMember(oThdConf, "blocks");
threads = GetObjectMember(oThdConf, "threads");
@@ -131,9 +131,11 @@ bool jconf::GetGPUThreadConfig(size_t id, thd_cfg &cfg)
bsleep = GetObjectMember(oThdConf, "bsleep");
aff = GetObjectMember(oThdConf, "affine_to_cpu");
syncMode = GetObjectMember(oThdConf, "sync_mode");
+ compMode = GetObjectMember(oThdConf, "comp_mode");
if(gid == nullptr || blocks == nullptr || threads == nullptr ||
- bfactor == nullptr || bsleep == nullptr || aff == nullptr || syncMode == nullptr)
+ bfactor == nullptr || bsleep == nullptr || aff == nullptr || syncMode == nullptr ||
+ compMode == nullptr)
{
return false;
}
@@ -161,13 +163,19 @@ bool jconf::GetGPUThreadConfig(size_t id, thd_cfg &cfg)
printer::inst()->print_msg(L0, "Error NVIDIA: sync_mode out of range or no number. ( range: 0 <= sync_mode < 4.)");
return false;
}
+
+ if(!compMode->IsBool())
+ return false;
+
+
cfg.id = gid->GetInt();
cfg.blocks = blocks->GetInt();
cfg.threads = threads->GetInt();
cfg.bfactor = bfactor->GetInt();
cfg.bsleep = bsleep->GetInt();
cfg.syncMode = syncMode->GetInt();
-
+ cfg.compMode = compMode->GetBool();
+
if(aff->IsNumber())
cfg.cpu_aff = aff->GetInt();
else
diff --git a/xmrstak/backend/nvidia/jconf.hpp b/xmrstak/backend/nvidia/jconf.hpp
index b4ebaa0..5ee1f81 100644
--- a/xmrstak/backend/nvidia/jconf.hpp
+++ b/xmrstak/backend/nvidia/jconf.hpp
@@ -29,6 +29,7 @@ public:
bool bNoPrefetch;
int32_t cpu_aff;
int syncMode;
+ bool compMode;
long long iCpuAff;
};
diff --git a/xmrstak/backend/nvidia/minethd.cpp b/xmrstak/backend/nvidia/minethd.cpp
index 0153eed..135f26e 100644
--- a/xmrstak/backend/nvidia/minethd.cpp
+++ b/xmrstak/backend/nvidia/minethd.cpp
@@ -78,6 +78,7 @@ minethd::minethd(miner_work& pWork, size_t iNo, const jconf::thd_cfg& cfg)
ctx.device_bfactor = (int)cfg.bfactor;
ctx.device_bsleep = (int)cfg.bsleep;
ctx.syncMode = cfg.syncMode;
+ ctx.compMode = cfg.compMode;
this->affinity = cfg.cpu_aff;
std::future<void> numa_guard = numa_promise.get_future();
diff --git a/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp b/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp
index d588641..8167395 100644
--- a/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp
+++ b/xmrstak/backend/nvidia/nvcc_code/cryptonight.hpp
@@ -16,6 +16,7 @@ typedef struct {
int device_bfactor;
int device_bsleep;
int syncMode;
+ bool compMode;
uint32_t *d_input;
uint32_t inputlen;
diff --git a/xmrstak/backend/nvidia/nvcc_code/cuda_core.cu b/xmrstak/backend/nvidia/nvcc_code/cuda_core.cu
index 8e69c20..1c9c9df 100644
--- a/xmrstak/backend/nvidia/nvcc_code/cuda_core.cu
+++ b/xmrstak/backend/nvidia/nvcc_code/cuda_core.cu
@@ -5,6 +5,7 @@
#include <string.h>
#include <cuda.h>
#include <cuda_runtime.h>
+#include <bitset>
#include "xmrstak/jconf.hpp"
#include "xmrstak/backend/nvidia/nvcc_code/cuda_fast_int_math_v2.hpp"
@@ -254,8 +255,12 @@ struct u64 : public uint2
}
};
-
-template<size_t ITERATIONS, uint32_t MEMORY, uint32_t MASK, xmrstak_algo ALGO>
+/** cryptonight with two threads per hash
+ *
+ * @tparam COMP_MODE if true than 64bit memory transfers per thread will be used to store/load data within shared memory
+ * else 128bit operations will be used
+ */
+template<size_t ITERATIONS, uint32_t MEMORY, uint32_t MASK, xmrstak_algo ALGO, bool COMP_MODE>
#ifdef XMR_STAK_THREADS
__launch_bounds__( XMR_STAK_THREADS * 2 )
#endif
@@ -329,7 +334,16 @@ __global__ void cryptonight_core_gpu_phase2_double( int threads, int bfactor, in
{
ptr0 = (uint64_t *)&l0[idx0 & MASK & 0x1FFFC0];
- ((ulong4*)myChunks)[sub] = ((ulong4*)ptr0)[sub];
+ if(COMP_MODE)
+ {
+ #pragma unroll 4
+ for(int x = 0; x < 8; x += 2)
+ {
+ myChunks[x + sub] = ptr0[ x + sub ];
+ }
+ }
+ else
+ ((ulong4*)myChunks)[sub] = ((ulong4*)ptr0)[sub];
uint32_t idx1 = (idx0 & 0x30) >> 3;
@@ -358,13 +372,31 @@ __global__ void cryptonight_core_gpu_phase2_double( int threads, int bfactor, in
}
myChunks[ idx1 + sub ] = cx_aes ^ bx0;
- ((ulong4*)ptr0)[sub] = ((ulong4*)myChunks)[sub];
+ if(COMP_MODE)
+ {
+ #pragma unroll 4
+ for(int x = 0; x < 8; x += 2)
+ {
+ ptr0[ x + sub ] = myChunks[x + sub];
+ }
+ }
+ else
+ ((ulong4*)ptr0)[sub] = ((ulong4*)myChunks)[sub];
idx0 = shuffle<2>(sPtr, sub, cx_aes.x, 0);
idx1 = (idx0 & 0x30) >> 3;
ptr0 = (uint64_t *)&l0[idx0 & MASK & 0x1FFFC0];
- ((ulong4*)myChunks)[sub] = ((ulong4*)ptr0)[sub];
+ if(COMP_MODE)
+ {
+ #pragma unroll 4
+ for(int x = 0; x < 8; x += 2)
+ {
+ myChunks[x + sub] = ptr0[ x + sub ];
+ }
+ }
+ else
+ ((ulong4*)myChunks)[sub] = ((ulong4*)ptr0)[sub];
if(ALGO != cryptonight_monero_v8)
bx0 = cx_aes;
@@ -420,7 +452,16 @@ __global__ void cryptonight_core_gpu_phase2_double( int threads, int bfactor, in
bx0 = cx_aes;
}
myChunks[ idx1 + sub ] = ax0;
- ((ulong4*)ptr0)[sub] = ((ulong4*)myChunks)[sub];
+ if(COMP_MODE)
+ {
+ #pragma unroll 4
+ for(int x = 0; x < 8; x += 2)
+ {
+ ptr0[ x + sub ] = myChunks[x + sub];
+ }
+ }
+ else
+ ((ulong4*)ptr0)[sub] = ((ulong4*)myChunks)[sub];
ax0 ^= c;
idx0 = shuffle<2>(sPtr, sub, static_cast<uint32_t>(ax0), 0);
}
@@ -699,7 +740,7 @@ __global__ void cryptonight_core_gpu_phase3( int threads, int bfactor, int parti
MEMCPY8( d_ctx_state + thread * 50 + sub + 16, text, 2 );
}
-template<size_t ITERATIONS, uint32_t MASK, uint32_t MEMORY, xmrstak_algo ALGO>
+template<size_t ITERATIONS, uint32_t MASK, uint32_t MEMORY, xmrstak_algo ALGO, bool COMP_MODE>
void cryptonight_core_gpu_hash(nvid_ctx* ctx, uint32_t nonce)
{
dim3 grid( ctx->device_blocks );
@@ -741,7 +782,7 @@ void cryptonight_core_gpu_hash(nvid_ctx* ctx, uint32_t nonce)
CUDA_CHECK_MSG_KERNEL(
ctx->device_id,
"\n**suggestion: Try to increase the value of the attribute 'bfactor' or \nreduce 'threads' in the NVIDIA config file.**",
- cryptonight_core_gpu_phase2_double<ITERATIONS,MEMORY,MASK,ALGO><<<
+ cryptonight_core_gpu_phase2_double<ITERATIONS,MEMORY,MASK,ALGO, COMP_MODE><<<
grid,
block2,
sizeof(uint64_t) * block2.x * 8 +
@@ -807,26 +848,50 @@ void cryptonight_core_gpu_hash(nvid_ctx* ctx, uint32_t nonce)
}
}
-typedef void (*cuda_hash_fn)(nvid_ctx* ctx, uint32_t nonce);
-
void cryptonight_core_cpu_hash(nvid_ctx* ctx, xmrstak_algo miner_algo, uint32_t startNonce)
{
+ typedef void (*cuda_hash_fn)(nvid_ctx* ctx, uint32_t nonce);
+
if(miner_algo == invalid_algo) return;
static const cuda_hash_fn func_table[] = {
- cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_lite>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_heavy>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_aeon>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_ipbc>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_stellite>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_MASARI_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_masari>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_haven>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_bittube2>,
- cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero_v8>
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_lite, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_lite, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_heavy, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_heavy, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_aeon, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_aeon, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_ipbc, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_LITE_ITER, CRYPTONIGHT_LITE_MASK, CRYPTONIGHT_LITE_MEMORY/4, cryptonight_ipbc, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_stellite, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_stellite, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_MASARI_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_masari, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_MASARI_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_masari, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_haven, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_haven, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_bittube2, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_HEAVY_ITER, CRYPTONIGHT_HEAVY_MASK, CRYPTONIGHT_HEAVY_MEMORY/4, cryptonight_bittube2, true>,
+
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero_v8, false>,
+ cryptonight_core_gpu_hash<CRYPTONIGHT_ITER, CRYPTONIGHT_MASK, CRYPTONIGHT_MEMORY/4, cryptonight_monero_v8, true>
};
- cuda_hash_fn selected_function = func_table[ miner_algo - 1u ];
+ std::bitset<1> digit;
+ digit.set(0, ctx->compMode);
+
+ cuda_hash_fn selected_function = func_table[ ((miner_algo - 1u) << 1) | digit.to_ulong() ];
selected_function(ctx, startNonce);
}
--
GitLab