From dc4e3793454a8ac7ae85704e7997d878c378b0aa Mon Sep 17 00:00:00 2001
From: psychocrypt <psychocryptHPC@gmail.com>
Date: Wed, 21 Feb 2018 22:23:00 +0100
Subject: [PATCH] fix broken memchunk feature
fix double definition of define `MEM_CHUNK`
---
xmrstak/backend/amd/amd_gpu/gpu.cpp | 2 +-
xmrstak/backend/amd/amd_gpu/opencl/cryptonight.cl | 2 +-
xmrstak/backend/amd/autoAdjust.hpp | 2 +-
xmrstak/backend/amd/config.tpl | 2 +-
4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/xmrstak/backend/amd/amd_gpu/gpu.cpp b/xmrstak/backend/amd/amd_gpu/gpu.cpp
index 8c4a40d..95d30f7 100644
--- a/xmrstak/backend/amd/amd_gpu/gpu.cpp
+++ b/xmrstak/backend/amd/amd_gpu/gpu.cpp
@@ -336,7 +336,7 @@ size_t InitOpenCLGpu(cl_context opencl_ctx, GpuContext* ctx, const char* source_
char options[256];
snprintf(options, sizeof(options),
- "-DITERATIONS=%d -DMASK=%d -DWORKSIZE=%llu -DSTRIDED_INDEX=%d -DMEM_CHUNK=%d -DCOMP_MODE=%d",
+ "-DITERATIONS=%d -DMASK=%d -DWORKSIZE=%llu -DSTRIDED_INDEX=%d -DMEM_CHUNK_EXPONENT=%d -DCOMP_MODE=%d",
hasIterations, threadMemMask, int_port(ctx->workSize), ctx->stridedIndex, int(1u<<ctx->memChunk), ctx->compMode ? 1 : 0);
ret = clBuildProgram(ctx->Program, 1, &ctx->DeviceID, options, NULL, NULL);
if(ret != CL_SUCCESS)
diff --git a/xmrstak/backend/amd/amd_gpu/opencl/cryptonight.cl b/xmrstak/backend/amd/amd_gpu/opencl/cryptonight.cl
index dbe8991..9383b04 100644
--- a/xmrstak/backend/amd/amd_gpu/opencl/cryptonight.cl
+++ b/xmrstak/backend/amd/amd_gpu/opencl/cryptonight.cl
@@ -416,7 +416,7 @@ void AESExpandKey256(uint *keybuf)
}
}
-#define MEM_CHUNK (1<<4)
+#define MEM_CHUNK (1<<MEM_CHUNK_EXPONENT)
#if(STRIDED_INDEX==0)
# define IDX(x) (x)
diff --git a/xmrstak/backend/amd/autoAdjust.hpp b/xmrstak/backend/amd/autoAdjust.hpp
index 8d60b94..8950105 100644
--- a/xmrstak/backend/amd/autoAdjust.hpp
+++ b/xmrstak/backend/amd/autoAdjust.hpp
@@ -143,7 +143,7 @@ private:
// set 8 threads per block (this is a good value for the most gpus)
conf += std::string(" { \"index\" : ") + std::to_string(ctx.deviceIdx) + ",\n" +
" \"intensity\" : " + std::to_string(intensity) + ", \"worksize\" : " + std::to_string(8) + ",\n" +
- " \"affine_to_cpu\" : false, \"strided_index\" : 1, \"mem_chunk\" : 4,\n"
+ " \"affine_to_cpu\" : false, \"strided_index\" : 1, \"mem_chunk\" : 2,\n"
" \"comp_mode\" : true\n" +
" },\n";
}
diff --git a/xmrstak/backend/amd/config.tpl b/xmrstak/backend/amd/config.tpl
index 84251c7..28855f0 100644
--- a/xmrstak/backend/amd/config.tpl
+++ b/xmrstak/backend/amd/config.tpl
@@ -19,7 +19,7 @@ R"===(
* in this case set the intensity to a multiple of the worksize or activate comp_mode.
* "gpu_threads_conf" :
* [
- * { "index" : 0, "intensity" : 1000, "worksize" : 8, "affine_to_cpu" : false, "strided_index" : true, "mem_chunk" : 4, "comp_mode" : true },
+ * { "index" : 0, "intensity" : 1000, "worksize" : 8, "affine_to_cpu" : false, "strided_index" : true, "mem_chunk" : 2, "comp_mode" : true },
* ],
* If you do not wish to mine with your AMD GPU(s) then use:
* "gpu_threads_conf" :
--
GitLab