1
0
mirror of https://github.com/xmrig/xmrig.git synced 2025-12-09 16:52:40 -05:00

Compare commits

..

7 Commits

Author SHA1 Message Date
Tony Butler
ad9c6ecb48 Merge 862280f28c into cd2fd9d7a6 2024-11-14 20:29:26 -07:00
XMRig
cd2fd9d7a6 Simplified getting PCI topology for the OpenCL backend. 2024-11-08 13:03:35 +07:00
XMRig
064cd3ef20 Fixed and simplified OpenCL GPU type detection. 2024-11-08 07:09:35 +07:00
XMRig
e8bbd134f9 v6.22.3-dev 2024-11-03 15:06:54 +07:00
XMRig
cf86a1e05c Merge branch 'master' into dev 2024-11-03 15:06:22 +07:00
Tony Butler
862280f28c How about this way 2023-07-12 02:06:53 -06:00
Tony Butler
814e1de2a6 CN: Consistency cleanup 2023-07-12 02:06:53 -06:00
10 changed files with 304 additions and 322 deletions

View File

@@ -1,6 +1,6 @@
/* XMRig
* Copyright (c) 2018-2021 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2021 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -19,10 +19,8 @@
#ifndef XMRIG_PCITOPOLOGY_H
#define XMRIG_PCITOPOLOGY_H
#include <cstdio>
#include "base/tools/String.h"
@@ -33,7 +31,14 @@ class PciTopology
{
public:
PciTopology() = default;
PciTopology(uint32_t bus, uint32_t device, uint32_t function) : m_valid(true), m_bus(bus), m_device(device), m_function(function) {}
template<typename T>
inline PciTopology(T bus, T device, T function)
: m_valid(true),
m_bus(static_cast<uint8_t>(bus)),
m_device(static_cast<uint8_t>(device)),
m_function(static_cast<uint8_t>(function))
{}
inline bool isEqual(const PciTopology &other) const { return m_valid == other.m_valid && toUint32() == other.toUint32(); }
inline bool isValid() const { return m_valid; }
@@ -70,4 +75,4 @@ private:
} // namespace xmrig
#endif /* XMRIG_PCITOPOLOGY_H */
#endif // XMRIG_PCITOPOLOGY_H

View File

@@ -1,6 +1,6 @@
/* XMRig
* Copyright (c) 2018-2021 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2021 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -227,7 +227,7 @@ public:
# endif
Log::print("|" CYAN_BOLD("%3zu") " |" CYAN_BOLD("%4u") " |" YELLOW(" %7s") " |" CYAN_BOLD("%10d") " |" CYAN_BOLD("%8d") " |"
CYAN_BOLD("%7d") " |" CYAN_BOLD("%3d") " |" CYAN_BOLD("%4d") " |" CYAN("%7zu") " | " GREEN("%s"),
CYAN_BOLD("%7d") " |" CYAN_BOLD("%3d") " |" CYAN_BOLD("%4d") " |" CYAN("%7zu") " | " GREEN_BOLD("%s"),
i,
data.thread.index(),
data.device.topology().toString().data(),

View File

@@ -5,8 +5,8 @@
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017-2018 XMR-Stak <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
* Copyright 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2020 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -22,7 +22,6 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "backend/cuda/wrappers/CudaDevice.h"
#include "3rdparty/rapidjson/document.h"
#include "backend/cuda/CudaThreads.h"
@@ -41,7 +40,7 @@
xmrig::CudaDevice::CudaDevice(uint32_t index, int32_t bfactor, int32_t bsleep) :
m_index(index)
{
auto ctx = CudaLib::alloc(index, bfactor, bsleep);
auto *ctx = CudaLib::alloc(index, bfactor, bsleep);
if (!CudaLib::deviceInfo(ctx, 0, 0, Algorithm::INVALID)) {
CudaLib::release(ctx);
@@ -50,7 +49,7 @@ xmrig::CudaDevice::CudaDevice(uint32_t index, int32_t bfactor, int32_t bsleep) :
m_ctx = ctx;
m_name = CudaLib::deviceName(ctx);
m_topology = PciTopology(CudaLib::deviceUint(ctx, CudaLib::DevicePciBusID), CudaLib::deviceUint(ctx, CudaLib::DevicePciDeviceID), 0);
m_topology = { CudaLib::deviceUint(ctx, CudaLib::DevicePciBusID), CudaLib::deviceUint(ctx, CudaLib::DevicePciDeviceID), 0U };
}

View File

@@ -5,13 +5,7 @@ if (BUILD_STATIC AND XMRIG_OS_UNIX AND WITH_OPENCL)
endif()
if (WITH_OPENCL)
add_definitions(/DXMRIG_FEATURE_OPENCL)
add_definitions(/DCL_USE_DEPRECATED_OPENCL_1_2_APIS)
if (XMRIG_OS_APPLE)
add_definitions(/DCL_TARGET_OPENCL_VERSION=120)
elseif (WITH_OPENCL_VERSION)
add_definitions(/DCL_TARGET_OPENCL_VERSION=${WITH_OPENCL_VERSION})
endif()
add_definitions(/DXMRIG_FEATURE_OPENCL /DCL_USE_DEPRECATED_OPENCL_1_2_APIS)
set(HEADERS_BACKEND_OPENCL
src/backend/opencl/cl/OclSource.h
@@ -71,6 +65,13 @@ if (WITH_OPENCL)
src/backend/opencl/wrappers/OclPlatform.cpp
)
if (XMRIG_OS_APPLE)
add_definitions(/DCL_TARGET_OPENCL_VERSION=120)
list(APPEND SOURCES_BACKEND_OPENCL src/backend/opencl/wrappers/OclDevice_mac.cpp)
elseif (WITH_OPENCL_VERSION)
add_definitions(/DCL_TARGET_OPENCL_VERSION=${WITH_OPENCL_VERSION})
endif()
if (WIN32)
list(APPEND SOURCES_BACKEND_OPENCL src/backend/opencl/OclCache_win.cpp)
else()

View File

@@ -1,6 +1,7 @@
/* XMRig
* Copyright (c) 2018-2021 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2021 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2021 Spudz76 <https://github.com/Spudz76>
* Copyright (c) 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -17,6 +18,7 @@
*/
#include "backend/opencl/wrappers/OclDevice.h"
#include "3rdparty/fmt/core.h"
#include "3rdparty/rapidjson/document.h"
#include "backend/opencl/OclGenerator.h"
#include "backend/opencl/OclThreads.h"
@@ -30,19 +32,21 @@
#include <algorithm>
// NOLINTNEXTLINE(modernize-use-using)
typedef union
{
struct { cl_uint type; cl_uint data[5]; } raw;
struct { cl_uint type; cl_char unused[17]; cl_char bus; cl_char device; cl_char function; } pcie;
} topology_amd;
#include <map>
namespace xmrig {
struct topology_amd {
cl_uint type;
cl_char unused[17];
cl_char bus;
cl_char device;
cl_char function;
};
#ifdef XMRIG_ALGO_RANDOMX
extern bool ocl_generic_rx_generator(const OclDevice &device, const Algorithm &algorithm, OclThreads &threads);
#endif
@@ -81,9 +85,11 @@ static OclVendor getPlatformVendorId(const String &vendor, const String &extensi
return OCL_VENDOR_INTEL;
}
# ifdef XMRIG_OS_APPLE
if (extensions.contains("cl_APPLE_") || vendor.contains("Apple")) {
return OCL_VENDOR_APPLE;
}
# endif
return OCL_VENDOR_UNKNOWN;
}
@@ -103,117 +109,16 @@ static OclVendor getVendorId(const String &vendor)
return OCL_VENDOR_INTEL;
}
# ifdef XMRIG_OS_APPLE
if (vendor.contains("Apple")) {
return OCL_VENDOR_APPLE;
}
# endif
return OCL_VENDOR_UNKNOWN;
}
static OclDevice::Type getType(const String &name, const OclVendor platformVendorId)
{
if (platformVendorId == OCL_VENDOR_APPLE) {
// Apple Platform: uses product names, not gfx# or codenames
if (name.contains("AMD Radeon")) {
if (name.contains(" 450 ") ||
name.contains(" 455 ") ||
name.contains(" 460 ")) {
return OclDevice::Baffin;
}
if (name.contains(" 555 ") || name.contains(" 555X ") ||
name.contains(" 560 ") || name.contains(" 560X ") ||
name.contains(" 570 ") || name.contains(" 570X ") ||
name.contains(" 575 ") || name.contains(" 575X ")) {
return OclDevice::Polaris;
}
if (name.contains(" 580 ") || name.contains(" 580X ")) {
return OclDevice::Ellesmere;
}
if (name.contains(" Vega ")) {
if (name.contains(" 48 ") ||
name.contains(" 56 ") ||
name.contains(" 64 ") ||
name.contains(" 64X ")) {
return OclDevice::Vega_10;
}
if (name.contains(" 16 ") ||
name.contains(" 20 ") ||
name.contains(" II ")) {
return OclDevice::Vega_20;
}
}
if (name.contains(" 5700 ") || name.contains(" W5700X ")) {
return OclDevice::Navi_10;
}
if (name.contains(" 5600 ") || name.contains(" 5600M ")) {
return OclDevice::Navi_12;
}
if (name.contains(" 5300 ") || name.contains(" 5300M ") ||
name.contains(" 5500 ") || name.contains(" 5500M ")) {
return OclDevice::Navi_14;
}
if (name.contains(" W6800 ") || name.contains(" W6900X ")) {
return OclDevice::Navi_21;
}
}
}
if (name == "gfx900" || name == "gfx901") {
return OclDevice::Vega_10;
}
if (name == "gfx902" || name == "gfx903") {
return OclDevice::Raven;
}
if (name == "gfx906" || name == "gfx907") {
return OclDevice::Vega_20;
}
if (name == "gfx1010") {
return OclDevice::Navi_10;
}
if (name == "gfx1011") {
return OclDevice::Navi_12;
}
if (name == "gfx1012") {
return OclDevice::Navi_14;
}
if (name == "gfx1030") {
return OclDevice::Navi_21;
}
if (name == "gfx804") {
return OclDevice::Lexa;
}
if (name == "Baffin") {
return OclDevice::Baffin;
}
if (name.contains("Ellesmere")) {
return OclDevice::Ellesmere;
}
if (name == "gfx803" || name.contains("polaris")) {
return OclDevice::Polaris;
}
return OclDevice::Unknown;
}
} // namespace xmrig
@@ -231,21 +136,21 @@ xmrig::OclDevice::OclDevice(uint32_t index, cl_device_id id, cl_platform_id plat
{
m_vendorId = getVendorId(m_vendor);
m_platformVendorId = getPlatformVendorId(m_platformVendor, m_extensions);
m_type = getType(m_name, m_platformVendorId);
m_type = getType(m_name);
if (m_extensions.contains("cl_amd_device_attribute_query")) {
topology_amd topology;
if (OclLib::getDeviceInfo(id, CL_DEVICE_TOPOLOGY_AMD, sizeof(topology), &topology, nullptr) == CL_SUCCESS && topology.raw.type == CL_DEVICE_TOPOLOGY_TYPE_PCIE_AMD) {
m_topology = PciTopology(static_cast<uint32_t>(topology.pcie.bus), static_cast<uint32_t>(topology.pcie.device), static_cast<uint32_t>(topology.pcie.function));
topology_amd topology{};
if (OclLib::getDeviceInfo(id, CL_DEVICE_TOPOLOGY_AMD, sizeof(topology), &topology) == CL_SUCCESS && topology.type == CL_DEVICE_TOPOLOGY_TYPE_PCIE_AMD) {
m_topology = { topology.bus, topology.device, topology.function };
}
m_board = OclLib::getString(id, CL_DEVICE_BOARD_NAME_AMD);
}
else if (m_extensions.contains("cl_nv_device_attribute_query")) {
cl_uint bus = 0;
if (OclLib::getDeviceInfo(id, CL_DEVICE_PCI_BUS_ID_NV, sizeof (bus), &bus, nullptr) == CL_SUCCESS) {
if (OclLib::getDeviceInfo(id, CL_DEVICE_PCI_BUS_ID_NV, sizeof(bus), &bus) == CL_SUCCESS) {
cl_uint slot = OclLib::getUint(id, CL_DEVICE_PCI_SLOT_ID_NV);
m_topology = PciTopology(bus, (slot >> 3) & 0xff, slot & 7);
m_topology = { bus, (slot >> 3) & 0xff, slot & 7 };
}
}
}
@@ -253,17 +158,11 @@ xmrig::OclDevice::OclDevice(uint32_t index, cl_device_id id, cl_platform_id plat
xmrig::String xmrig::OclDevice::printableName() const
{
const size_t size = m_board.size() + m_name.size() + 64;
char *buf = new char[size]();
if (m_board.isNull()) {
snprintf(buf, size, GREEN_BOLD("%s"), m_name.data());
}
else {
snprintf(buf, size, GREEN_BOLD("%s") " (" CYAN_BOLD("%s") ")", m_board.data(), m_name.data());
return fmt::format(GREEN_BOLD("{}"), m_name).c_str();
}
return buf;
return fmt::format(GREEN_BOLD("{}") " (" CYAN_BOLD("{}") ")", m_board, m_name).c_str();
}
@@ -311,3 +210,35 @@ void xmrig::OclDevice::toJSON(rapidjson::Value &out, rapidjson::Document &doc) c
# endif
}
#endif
#ifndef XMRIG_OS_APPLE
xmrig::OclDevice::Type xmrig::OclDevice::getType(const String &name)
{
static std::map<const char *, OclDevice::Type> types = {
{ "gfx900", Vega_10 },
{ "gfx901", Vega_10 },
{ "gfx902", Raven },
{ "gfx903", Raven },
{ "gfx906", Vega_20 },
{ "gfx907", Vega_20 },
{ "gfx1010", Navi_10 },
{ "gfx1011", Navi_12 },
{ "gfx1012", Navi_14 },
{ "gfx1030", Navi_21 },
{ "gfx804", Lexa },
{ "Baffin", Baffin },
{ "Ellesmere", Ellesmere },
{ "gfx803", Polaris },
{ "polaris", Polaris },
};
for (auto &kv : types) {
if (name.contains(kv.first)) {
return kv.second;
}
}
return OclDevice::Unknown;
}
#endif

View File

@@ -1,6 +1,6 @@
/* XMRig
* Copyright (c) 2018-2021 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2021 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -86,6 +86,8 @@ public:
# endif
private:
static OclDevice::Type getType(const String &name);
cl_device_id m_id = nullptr;
cl_platform_id m_platform = nullptr;
const String m_platformVendor;

View File

@@ -0,0 +1,77 @@
/* XMRig
* Copyright (c) 2021 Spudz76 <https://github.com/Spudz76>
* Copyright (c) 2018-2024 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2024 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "backend/opencl/wrappers/OclDevice.h"
xmrig::OclDevice::Type xmrig::OclDevice::getType(const String &name)
{
// Apple Platform: uses product names, not gfx# or codenames
if (name.contains("AMD Radeon")) {
if (name.contains(" 450 ") ||
name.contains(" 455 ") ||
name.contains(" 460 ")) {
return Baffin;
}
if (name.contains(" 555 ") || name.contains(" 555X ") ||
name.contains(" 560 ") || name.contains(" 560X ") ||
name.contains(" 570 ") || name.contains(" 570X ") ||
name.contains(" 575 ") || name.contains(" 575X ")) {
return Polaris;
}
if (name.contains(" 580 ") || name.contains(" 580X ")) {
return Ellesmere;
}
if (name.contains(" Vega ")) {
if (name.contains(" 48 ") ||
name.contains(" 56 ") ||
name.contains(" 64 ") ||
name.contains(" 64X ")) {
return Vega_10;
}
if (name.contains(" 16 ") ||
name.contains(" 20 ") ||
name.contains(" II ")) {
return Vega_20;
}
}
if (name.contains(" 5700 ") || name.contains(" W5700X ")) {
return Navi_10;
}
if (name.contains(" 5600 ") || name.contains(" 5600M ")) {
return Navi_12;
}
if (name.contains(" 5300 ") || name.contains(" 5300M ") ||
name.contains(" 5500 ") || name.contains(" 5500M ")) {
return Navi_14;
}
if (name.contains(" W6800 ") || name.contains(" W6900X ")) {
return Navi_21;
}
}
return OclDevice::Unknown;
}

View File

@@ -37,14 +37,36 @@ class CnAlgo
public:
constexpr CnAlgo() {};
constexpr inline Algorithm::Id base() const { static_assert(Algorithm::isCN(ALGO), "invalid CRYPTONIGHT algorithm"); return Algorithm::base(ALGO); }
constexpr inline bool isHeavy() const { return Algorithm::family(ALGO) == Algorithm::CN_HEAVY; }
constexpr inline bool isR() const { return ALGO == Algorithm::CN_R; }
constexpr inline size_t memory() const { static_assert(Algorithm::isCN(ALGO), "invalid CRYPTONIGHT algorithm"); return Algorithm::l3(ALGO); }
constexpr inline uint32_t iterations() const { static_assert(Algorithm::isCN(ALGO), "invalid CRYPTONIGHT algorithm"); return CN_ITER; }
# define ASSERT_CN static_assert(Algorithm::isCN(ALGO), "invalid CRYPTONIGHT algorithm")
constexpr inline Algorithm::Id base() const { ASSERT_CN; return Algorithm::base(ALGO); }
constexpr inline size_t memory() const { ASSERT_CN; return Algorithm::l3(ALGO); }
constexpr inline uint32_t iterations() const { ASSERT_CN; return CN_ITER; }
constexpr inline uint32_t mask() const { return static_cast<uint32_t>(((memory() - 1) / 16) * 16); }
constexpr inline uint32_t half_mem() const { return mask() < memory() / 2; }
constexpr inline bool isBase1() const { ASSERT_CN; return Algorithm::base(ALGO) == Algorithm::CN_1; }
constexpr inline bool isBase2() const { ASSERT_CN; return Algorithm::base(ALGO) == Algorithm::CN_2; }
constexpr inline bool is2() const { return ALGO == Algorithm::CN_2; }
constexpr inline bool isR() const { return ALGO == Algorithm::CN_R; }
constexpr inline bool isHalf() const { return ALGO == Algorithm::CN_HALF; }
constexpr inline bool isRTO() const { return ALGO == Algorithm::CN_RTO; }
constexpr inline bool isRWZ() const { return ALGO == Algorithm::CN_RWZ; }
constexpr inline bool isZLS() const { return ALGO == Algorithm::CN_ZLS; }
constexpr inline bool isDouble() const { return ALGO == Algorithm::CN_DOUBLE; }
constexpr inline bool isCCX() const { return ALGO == Algorithm::CN_CCX; }
constexpr inline bool isHeavy() const { ASSERT_CN; return Algorithm::family(ALGO) == Algorithm::CN_HEAVY; }
constexpr inline bool isHeavyTube() const { return ALGO == Algorithm::CN_HEAVY_TUBE; }
constexpr inline bool isHeavyXHV() const { return ALGO == Algorithm::CN_HEAVY_XHV; }
constexpr inline bool isPico0() const { return ALGO == Algorithm::CN_PICO_0; }
constexpr inline bool isPicoTLO() const { return ALGO == Algorithm::CN_PICO_TLO; }
constexpr inline bool isUPX2() const { return ALGO == Algorithm::CN_UPX2; }
constexpr inline bool isGR0() const { return ALGO == Algorithm::CN_GR_0; }
constexpr inline bool isGR1() const { return ALGO == Algorithm::CN_GR_1; }
constexpr inline bool isGR2() const { return ALGO == Algorithm::CN_GR_2; }
constexpr inline bool isGR3() const { return ALGO == Algorithm::CN_GR_3; }
constexpr inline bool isGR4() const { return ALGO == Algorithm::CN_GR_4; }
constexpr inline bool isGR5() const { return ALGO == Algorithm::CN_GR_5; }
inline static uint32_t iterations(Algorithm::Id algo)
{
switch (algo) {

View File

@@ -603,7 +603,7 @@ static inline void cryptonight_monero_tweak(uint64_t *mem_out, const uint8_t *l,
constexpr CnAlgo<ALGO> props;
if (props.base() == Algorithm::CN_2) {
VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx, (((ALGO == Algorithm::CN_RWZ) || (ALGO == Algorithm::CN_UPX2)) ? 1 : 0));
VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1, cx, ((props.isRWZ() || props.isUPX2()) ? 1 : 0));
_mm_store_si128(reinterpret_cast<__m128i *>(mem_out), _mm_xor_si128(bx0, cx));
} else {
__m128i tmp = _mm_xor_si128(bx0, cx);
@@ -665,15 +665,8 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
constexpr CnAlgo<ALGO> props;
constexpr size_t MASK = props.mask();
constexpr Algorithm::Id BASE = props.base();
# ifdef XMRIG_ALGO_CN_HEAVY
constexpr bool IS_CN_HEAVY_TUBE = ALGO == Algorithm::CN_HEAVY_TUBE;
# else
constexpr bool IS_CN_HEAVY_TUBE = false;
# endif
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32);
return;
}
@@ -694,10 +687,7 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
V4_Instruction code[256];
const int code_size = v4_random_math_init<ALGO>(code, height);
if (ALGO == Algorithm::CN_R) {
v4_soft_aes_compile_code(code, code_size, reinterpret_cast<void*>(ctx[0]->generated_code), Assembly::NONE);
}
v4_soft_aes_compile_code(code, code_size, reinterpret_cast<void*>(ctx[0]->generated_code), Assembly::NONE);
ctx[0]->generated_code_data = { ALGO, height };
}
@@ -718,26 +708,26 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
__m128i bx1 = _mm_set_epi64x(static_cast<int64_t>(h0[9] ^ h0[11]), static_cast<int64_t>(h0[8] ^ h0[10]));
__m128 conc_var;
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
conc_var = _mm_setzero_ps();
RESTORE_ROUNDING_MODE();
}
for (size_t i = 0; i < props.iterations(); i++) {
__m128i cx;
if (IS_CN_HEAVY_TUBE || !SOFT_AES) {
if (props.isHeavyTube() || !SOFT_AES) {
cx = _mm_load_si128(reinterpret_cast<const __m128i *>(&l0[interleaved_index<interleave>(idx0 & MASK)]));
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
cryptonight_conceal_tweak(cx, conc_var);
}
}
const __m128i ax0 = _mm_set_epi64x(static_cast<int64_t>(ah0), static_cast<int64_t>(al0));
if (IS_CN_HEAVY_TUBE) {
if (props.isHeavyTube()) {
cx = aes_round_tweak_div(cx, ax0);
}
else if (SOFT_AES) {
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
cx = _mm_load_si128(reinterpret_cast<const __m128i*>(&l0[interleaved_index<interleave>(idx0 & MASK)]));
cryptonight_conceal_tweak(cx, conc_var);
cx = soft_aesenc(&cx, ax0, reinterpret_cast<const uint32_t*>(saes_table));
@@ -750,7 +740,7 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
cx = _mm_aesenc_si128(cx, ax0);
}
if (BASE == Algorithm::CN_1 || BASE == Algorithm::CN_2) {
if (props.isBase1() || props.isBase2()) {
cryptonight_monero_tweak<ALGO>(reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)]), l0, idx0 & MASK, ax0, bx0, bx1, cx);
} else {
_mm_store_si128(reinterpret_cast<__m128i *>(&l0[interleaved_index<interleave>(idx0 & MASK)]), _mm_xor_si128(bx0, cx));
@@ -762,13 +752,11 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
cl = (reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)]))[0];
ch = (reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)]))[1];
if (BASE == Algorithm::CN_2) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT4_RANDOM_MATH(0, al0, ah0, cl, bx0, bx1);
if (ALGO == Algorithm::CN_R) {
al0 ^= r0[2] | (static_cast<uint64_t>(r0[3]) << 32);
ah0 ^= r0[0] | (static_cast<uint64_t>(r0[1]) << 32);
}
al0 ^= r0[2] | (static_cast<uint64_t>(r0[3]) << 32);
ah0 ^= r0[0] | (static_cast<uint64_t>(r0[1]) << 32);
} else {
VARIANT2_INTEGER_MATH(0, cl, cx);
}
@@ -776,11 +764,11 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
lo = __umul128(idx0, cl, &hi);
if (BASE == Algorithm::CN_2) {
if (ALGO == Algorithm::CN_R) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx0, bx1, cx, 0);
} else {
VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo, (((ALGO == Algorithm::CN_RWZ) || (ALGO == Algorithm::CN_UPX2)) ? 1 : 0));
VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo, ((props.isRWZ() || props.isUPX2()) ? 1 : 0));
}
}
@@ -789,9 +777,9 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)])[0] = al0;
if (IS_CN_HEAVY_TUBE || ALGO == Algorithm::CN_RTO) {
if (props.isHeavyTube() || props.isRTO()) {
reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)])[1] = ah0 ^ tweak1_2_0 ^ al0;
} else if (BASE == Algorithm::CN_1) {
} else if (props.isBase1()) {
reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)])[1] = ah0 ^ tweak1_2_0;
} else {
reinterpret_cast<uint64_t*>(&l0[interleaved_index<interleave>(idx0 & MASK)])[1] = ah0;
@@ -819,7 +807,7 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
((int64_t*)&l0[interleaved_index<interleave>(idx0 & MASK)])[0] = n ^ q;
if (ALGO == Algorithm::CN_HEAVY_XHV) {
if (props.isHeavyXHV()) {
d = ~d;
}
@@ -827,7 +815,7 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si
}
# endif
if (BASE == Algorithm::CN_2) {
if (props.isBase2()) {
bx1 = bx0;
}
@@ -960,7 +948,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
}
cn_explode_scratchpad<ALGO, false, 0>(ctx[0]);
if (ALGO == Algorithm::CN_2) {
if (props.is2()) {
if (ASM == Assembly::INTEL) {
cnv2_mainloop_ivybridge_asm(ctx);
}
@@ -971,7 +959,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
cnv2_mainloop_bulldozer_asm(ctx);
}
}
else if (ALGO == Algorithm::CN_HALF) {
else if (props.isHalf()) {
if (ASM == Assembly::INTEL) {
cn_half_mainloop_ivybridge_asm(ctx);
}
@@ -983,7 +971,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
}
}
# ifdef XMRIG_ALGO_CN_PICO
else if (ALGO == Algorithm::CN_PICO_0) {
else if (props.isPico0()) {
if (ASM == Assembly::INTEL) {
cn_trtl_mainloop_ivybridge_asm(ctx);
}
@@ -994,7 +982,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
cn_trtl_mainloop_bulldozer_asm(ctx);
}
}
else if (ALGO == Algorithm::CN_PICO_TLO) {
else if (props.isPicoTLO()) {
if (ASM == Assembly::INTEL) {
cn_tlo_mainloop_ivybridge_asm(ctx);
}
@@ -1006,10 +994,10 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
}
}
# endif
else if (ALGO == Algorithm::CN_RWZ) {
else if (props.isRWZ()) {
cnv2_rwz_mainloop_asm(ctx);
}
else if (ALGO == Algorithm::CN_ZLS) {
else if (props.isZLS()) {
if (ASM == Assembly::INTEL) {
cn_zls_mainloop_ivybridge_asm(ctx);
}
@@ -1020,7 +1008,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
cn_zls_mainloop_bulldozer_asm(ctx);
}
}
else if (ALGO == Algorithm::CN_DOUBLE) {
else if (props.isDouble()) {
if (ASM == Assembly::INTEL) {
cn_double_mainloop_ivybridge_asm(ctx);
}
@@ -1032,7 +1020,7 @@ inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_
}
}
# ifdef XMRIG_ALGO_CN_FEMTO
else if (ALGO == Algorithm::CN_UPX2) {
else if (props.isUPX2()) {
cn_upx2_mainloop_asm(ctx);
}
# endif
@@ -1078,22 +1066,22 @@ inline void cryptonight_double_hash_asm(const uint8_t *__restrict__ input, size_
cn_explode_scratchpad<ALGO, false, 0>(ctx[1]);
}
if (ALGO == Algorithm::CN_2) {
if (props.is2()) {
cnv2_double_mainloop_sandybridge_asm(ctx);
}
else if (ALGO == Algorithm::CN_HALF) {
else if (props.isHalf()){
cn_half_double_mainloop_sandybridge_asm(ctx);
}
# ifdef XMRIG_ALGO_CN_PICO
else if (ALGO == Algorithm::CN_PICO_0) {
else if (props.isPico0()) {
cn_trtl_double_mainloop_sandybridge_asm(ctx);
}
else if (ALGO == Algorithm::CN_PICO_TLO) {
else if (props.isPicoTLO()) {
cn_tlo_double_mainloop_sandybridge_asm(ctx);
}
# endif
# ifdef XMRIG_ALGO_CN_FEMTO
else if (ALGO == Algorithm::CN_UPX2) {
else if (props.isUPX2()) {
if (Cpu::info()->arch() == ICpuInfo::ARCH_ZEN3) {
cnv2_upx_double_mainloop_zen3_asm(ctx);
}
@@ -1102,13 +1090,13 @@ inline void cryptonight_double_hash_asm(const uint8_t *__restrict__ input, size_
}
}
# endif
else if (ALGO == Algorithm::CN_RWZ) {
else if (props.isRWZ()) {
cnv2_rwz_double_mainloop_asm(ctx);
}
else if (ALGO == Algorithm::CN_ZLS) {
else if (props.isZLS()) {
cn_zls_double_mainloop_sandybridge_asm(ctx);
}
else if (ALGO == Algorithm::CN_DOUBLE) {
else if (props.isDouble()) {
cn_double_double_mainloop_sandybridge_asm(ctx);
}
else if (props.isR()) {
@@ -1146,9 +1134,8 @@ template<Algorithm::Id ALGO>
static NOINLINE void cryptonight_single_hash_gr_sse41(const uint8_t* __restrict__ input, size_t size, uint8_t* __restrict__ output, cryptonight_ctx** __restrict__ ctx, uint64_t height)
{
constexpr CnAlgo<ALGO> props;
constexpr Algorithm::Id BASE = props.base();
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32);
return;
}
@@ -1163,12 +1150,12 @@ static NOINLINE void cryptonight_single_hash_gr_sse41(const uint8_t* __restrict_
VARIANT1_INIT(0);
ctx[0]->tweak1_2 = tweak1_2_0;
ctx[0]->tweak1_table = tweak1_table;
if (ALGO == Algorithm::CN_GR_0) cn_gr0_single_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_1) cn_gr1_single_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_2) cn_gr2_single_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_3) cn_gr3_single_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_4) cn_gr4_single_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_5) cn_gr5_single_mainloop_asm(ctx);
if (props.isGR0()) cn_gr0_single_mainloop_asm(ctx);
if (props.isGR1()) cn_gr1_single_mainloop_asm(ctx);
if (props.isGR2()) cn_gr2_single_mainloop_asm(ctx);
if (props.isGR3()) cn_gr3_single_mainloop_asm(ctx);
if (props.isGR4()) cn_gr4_single_mainloop_asm(ctx);
if (props.isGR5()) cn_gr5_single_mainloop_asm(ctx);
cn_implode_scratchpad<ALGO, false, 0>(ctx[0]);
keccakf(reinterpret_cast<uint64_t*>(ctx[0]->state), 24);
@@ -1180,9 +1167,8 @@ template<Algorithm::Id ALGO>
static NOINLINE void cryptonight_double_hash_gr_sse41(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx, uint64_t height)
{
constexpr CnAlgo<ALGO> props;
constexpr Algorithm::Id BASE = props.base();
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 64);
return;
}
@@ -1196,7 +1182,7 @@ static NOINLINE void cryptonight_double_hash_gr_sse41(const uint8_t *__restrict_
}
# ifdef XMRIG_VAES
if (!props.isHeavy() && cn_vaes_enabled) {
if (cn_vaes_enabled) {
cn_explode_scratchpad_vaes_double(ctx[0], ctx[1], props.memory(), props.half_mem());
}
else
@@ -1214,15 +1200,15 @@ static NOINLINE void cryptonight_double_hash_gr_sse41(const uint8_t *__restrict_
ctx[0]->tweak1_table = tweak1_table;
if (ALGO == Algorithm::CN_GR_0) cn_gr0_double_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_1) cn_gr1_double_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_2) cn_gr2_double_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_3) cn_gr3_double_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_4) cn_gr4_double_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_5) cn_gr5_double_mainloop_asm(ctx);
if (props.isGR0()) cn_gr0_double_mainloop_asm(ctx);
if (props.isGR1()) cn_gr1_double_mainloop_asm(ctx);
if (props.isGR2()) cn_gr2_double_mainloop_asm(ctx);
if (props.isGR3()) cn_gr3_double_mainloop_asm(ctx);
if (props.isGR4()) cn_gr4_double_mainloop_asm(ctx);
if (props.isGR5()) cn_gr5_double_mainloop_asm(ctx);
# ifdef XMRIG_VAES
if (!props.isHeavy() && cn_vaes_enabled) {
if (cn_vaes_enabled) {
cn_implode_scratchpad_vaes_double(ctx[0], ctx[1], props.memory(), props.half_mem());
}
else
@@ -1267,15 +1253,8 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
constexpr CnAlgo<ALGO> props;
constexpr size_t MASK = props.mask();
constexpr Algorithm::Id BASE = props.base();
# ifdef XMRIG_ALGO_CN_HEAVY
constexpr bool IS_CN_HEAVY_TUBE = ALGO == Algorithm::CN_HEAVY_TUBE;
# else
constexpr bool IS_CN_HEAVY_TUBE = false;
# endif
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 64);
return;
}
@@ -1323,7 +1302,7 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
__m128i bx11 = _mm_set_epi64x(h1[9] ^ h1[11], h1[8] ^ h1[10]);
__m128 conc_var0, conc_var1;
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
conc_var0 = _mm_setzero_ps();
conc_var1 = _mm_setzero_ps();
RESTORE_ROUNDING_MODE();
@@ -1334,10 +1313,10 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
for (size_t i = 0; i < props.iterations(); i++) {
__m128i cx0, cx1;
if (IS_CN_HEAVY_TUBE || !SOFT_AES) {
if (props.isHeavyTube() || !SOFT_AES) {
cx0 = _mm_load_si128(reinterpret_cast<const __m128i *>(&l0[idx0 & MASK]));
cx1 = _mm_load_si128(reinterpret_cast<const __m128i *>(&l1[idx1 & MASK]));
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
cryptonight_conceal_tweak(cx0, conc_var0);
cryptonight_conceal_tweak(cx1, conc_var1);
}
@@ -1345,12 +1324,12 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
const __m128i ax0 = _mm_set_epi64x(ah0, al0);
const __m128i ax1 = _mm_set_epi64x(ah1, al1);
if (IS_CN_HEAVY_TUBE) {
if (props.isHeavyTube()) {
cx0 = aes_round_tweak_div(cx0, ax0);
cx1 = aes_round_tweak_div(cx1, ax1);
}
else if (SOFT_AES) {
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
cx0 = _mm_load_si128(reinterpret_cast<const __m128i*>(&l0[idx0 & MASK]));
cx1 = _mm_load_si128(reinterpret_cast<const __m128i*>(&l1[idx1 & MASK]));
cryptonight_conceal_tweak(cx0, conc_var0);
@@ -1368,7 +1347,7 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
cx1 = _mm_aesenc_si128(cx1, ax1);
}
if (BASE == Algorithm::CN_1 || BASE == Algorithm::CN_2) {
if (props.isBase1() || props.isBase2()) {
cryptonight_monero_tweak<ALGO>((uint64_t*)&l0[idx0 & MASK], l0, idx0 & MASK, ax0, bx00, bx01, cx0);
cryptonight_monero_tweak<ALGO>((uint64_t*)&l1[idx1 & MASK], l1, idx1 & MASK, ax1, bx10, bx11, cx1);
} else {
@@ -1383,13 +1362,11 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
cl = ((uint64_t*) &l0[idx0 & MASK])[0];
ch = ((uint64_t*) &l0[idx0 & MASK])[1];
if (BASE == Algorithm::CN_2) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT4_RANDOM_MATH(0, al0, ah0, cl, bx00, bx01);
if (ALGO == Algorithm::CN_R) {
al0 ^= r0[2] | ((uint64_t)(r0[3]) << 32);
ah0 ^= r0[0] | ((uint64_t)(r0[1]) << 32);
}
al0 ^= r0[2] | ((uint64_t)(r0[3]) << 32);
ah0 ^= r0[0] | ((uint64_t)(r0[1]) << 32);
} else {
VARIANT2_INTEGER_MATH(0, cl, cx0);
}
@@ -1397,11 +1374,11 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
lo = __umul128(idx0, cl, &hi);
if (BASE == Algorithm::CN_2) {
if (ALGO == Algorithm::CN_R) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT2_SHUFFLE(l0, idx0 & MASK, ax0, bx00, bx01, cx0, 0);
} else {
VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo, (((ALGO == Algorithm::CN_RWZ) || (ALGO == Algorithm::CN_UPX2)) ? 1 : 0));
VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo, ((props.isRWZ() || props.isUPX2()) ? 1 : 0));
}
}
@@ -1410,9 +1387,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
((uint64_t*)&l0[idx0 & MASK])[0] = al0;
if (IS_CN_HEAVY_TUBE || ALGO == Algorithm::CN_RTO) {
if (props.isHeavyTube() || props.isRTO()) {
((uint64_t*) &l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0;
} else if (BASE == Algorithm::CN_1) {
} else if (props.isBase1()) {
((uint64_t*) &l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0;
} else {
((uint64_t*) &l0[idx0 & MASK])[1] = ah0;
@@ -1430,7 +1407,7 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
((int64_t*)&l0[idx0 & MASK])[0] = n ^ q;
if (ALGO == Algorithm::CN_HEAVY_XHV) {
if (props.isHeavyXHV()) {
d = ~d;
}
@@ -1441,13 +1418,11 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
cl = ((uint64_t*) &l1[idx1 & MASK])[0];
ch = ((uint64_t*) &l1[idx1 & MASK])[1];
if (BASE == Algorithm::CN_2) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT4_RANDOM_MATH(1, al1, ah1, cl, bx10, bx11);
if (ALGO == Algorithm::CN_R) {
al1 ^= r1[2] | ((uint64_t)(r1[3]) << 32);
ah1 ^= r1[0] | ((uint64_t)(r1[1]) << 32);
}
al1 ^= r1[2] | ((uint64_t)(r1[3]) << 32);
ah1 ^= r1[0] | ((uint64_t)(r1[1]) << 32);
} else {
VARIANT2_INTEGER_MATH(1, cl, cx1);
}
@@ -1455,11 +1430,11 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
lo = __umul128(idx1, cl, &hi);
if (BASE == Algorithm::CN_2) {
if (ALGO == Algorithm::CN_R) {
if (props.isBase2()) {
if (props.isR()) {
VARIANT2_SHUFFLE(l1, idx1 & MASK, ax1, bx10, bx11, cx1, 0);
} else {
VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo, (((ALGO == Algorithm::CN_RWZ) || (ALGO == Algorithm::CN_UPX2)) ? 1 : 0));
VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo, ((props.isRWZ() || props.isUPX2()) ? 1 : 0));
}
}
@@ -1468,9 +1443,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
((uint64_t*)&l1[idx1 & MASK])[0] = al1;
if (IS_CN_HEAVY_TUBE || ALGO == Algorithm::CN_RTO) {
if (props.isHeavyTube() || props.isRTO()) {
((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1 ^ al1;
} else if (BASE == Algorithm::CN_1) {
} else if (props.isBase1()) {
((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1;
} else {
((uint64_t*)&l1[idx1 & MASK])[1] = ah1;
@@ -1488,7 +1463,7 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
((int64_t*)&l1[idx1 & MASK])[0] = n ^ q;
if (ALGO == Algorithm::CN_HEAVY_XHV) {
if (props.isHeavyXHV()) {
d = ~d;
}
@@ -1496,7 +1471,7 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si
}
# endif
if (BASE == Algorithm::CN_2) {
if (props.isBase2()) {
bx01 = bx00;
bx11 = bx10;
}
@@ -1529,9 +1504,8 @@ template<Algorithm::Id ALGO>
static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__ input, size_t size, uint8_t* __restrict__ output, cryptonight_ctx** __restrict__ ctx, uint64_t height)
{
constexpr CnAlgo<ALGO> props;
constexpr Algorithm::Id BASE = props.base();
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32 * 4);
return;
}
@@ -1549,7 +1523,7 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
}
# ifdef XMRIG_VAES
if (!props.isHeavy() && cn_vaes_enabled) {
if (cn_vaes_enabled) {
cn_explode_scratchpad_vaes_double(ctx[0], ctx[1], props.memory(), props.half_mem());
cn_explode_scratchpad_vaes_double(ctx[2], ctx[3], props.memory(), props.half_mem());
}
@@ -1569,15 +1543,15 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
ctx[0]->tweak1_table = tweak1_table;
if (ALGO == Algorithm::CN_GR_0) cn_gr0_quad_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_1) cn_gr1_quad_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_2) cn_gr2_quad_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_3) cn_gr3_quad_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_4) cn_gr4_quad_mainloop_asm(ctx);
if (ALGO == Algorithm::CN_GR_5) cn_gr5_quad_mainloop_asm(ctx);
if (props.isGR0()) cn_gr0_quad_mainloop_asm(ctx);
if (props.isGR1()) cn_gr1_quad_mainloop_asm(ctx);
if (props.isGR2()) cn_gr2_quad_mainloop_asm(ctx);
if (props.isGR3()) cn_gr3_quad_mainloop_asm(ctx);
if (props.isGR4()) cn_gr4_quad_mainloop_asm(ctx);
if (props.isGR5()) cn_gr5_quad_mainloop_asm(ctx);
# ifdef XMRIG_VAES
if (!props.isHeavy() && cn_vaes_enabled) {
if (cn_vaes_enabled) {
cn_implode_scratchpad_vaes_double(ctx[0], ctx[1], props.memory(), props.half_mem());
cn_implode_scratchpad_vaes_double(ctx[2], ctx[3], props.memory(), props.half_mem());
}
@@ -1606,14 +1580,14 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
#define CN_STEP1(a, b0, b1, c, l, ptr, idx, conc_var) \
ptr = reinterpret_cast<__m128i*>(&l[idx & MASK]); \
c = _mm_load_si128(ptr); \
if (ALGO == Algorithm::CN_CCX) { \
if (props.isCCX()) { \
cryptonight_conceal_tweak(c, conc_var); \
}
#define CN_STEP2(a, b0, b1, c, l, ptr, idx) \
if (IS_CN_HEAVY_TUBE) { \
if (props.isHeavyTube()) { \
c = aes_round_tweak_div(c, a); \
} \
else if (SOFT_AES) { \
@@ -1622,7 +1596,7 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
c = _mm_aesenc_si128(c, a); \
} \
\
if (BASE == Algorithm::CN_1 || BASE == Algorithm::CN_2) { \
if (props.isBase1() || props.isBase2()) { \
cryptonight_monero_tweak<ALGO>((uint64_t*)ptr, l, idx & MASK, a, b0, b1, c); \
} else { \
_mm_store_si128(ptr, _mm_xor_si128(b0, c)); \
@@ -1638,36 +1612,34 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
#define CN_STEP4(part, a, b0, b1, c, l, mc, ptr, idx) \
uint64_t al##part, ah##part; \
if (BASE == Algorithm::CN_2) { \
if (props.isBase2()) { \
if (props.isR()) { \
al##part = _mm_cvtsi128_si64(a); \
ah##part = _mm_cvtsi128_si64(_mm_srli_si128(a, 8)); \
VARIANT4_RANDOM_MATH(part, al##part, ah##part, cl##part, b0, b1); \
if (ALGO == Algorithm::CN_R) { \
al##part ^= r##part[2] | ((uint64_t)(r##part[3]) << 32); \
ah##part ^= r##part[0] | ((uint64_t)(r##part[1]) << 32); \
} \
al##part ^= r##part[2] | ((uint64_t)(r##part[3]) << 32); \
ah##part ^= r##part[0] | ((uint64_t)(r##part[1]) << 32); \
} else { \
VARIANT2_INTEGER_MATH(part, cl##part, c); \
} \
} \
lo = __umul128(idx, cl##part, &hi); \
if (BASE == Algorithm::CN_2) { \
if (ALGO == Algorithm::CN_R) { \
if (props.isBase2()) { \
if (props.isR()) { \
VARIANT2_SHUFFLE(l, idx & MASK, a, b0, b1, c, 0); \
} else { \
VARIANT2_SHUFFLE2(l, idx & MASK, a, b0, b1, hi, lo, (((ALGO == Algorithm::CN_RWZ) || (ALGO == Algorithm::CN_UPX2)) ? 1 : 0)); \
VARIANT2_SHUFFLE2(l, idx & MASK, a, b0, b1, hi, lo, ((props.isRWZ() || props.isUPX2()) ? 1 : 0)); \
} \
} \
if (ALGO == Algorithm::CN_R) { \
if (props.isR()) { \
a = _mm_set_epi64x(ah##part, al##part); \
} \
a = _mm_add_epi64(a, _mm_set_epi64x(lo, hi)); \
\
if (BASE == Algorithm::CN_1) { \
if (props.isBase1()) { \
_mm_store_si128(ptr, _mm_xor_si128(a, mc)); \
\
if (IS_CN_HEAVY_TUBE || ALGO == Algorithm::CN_RTO) { \
if (props.isHeavyTube() || props.isRTO()) { \
((uint64_t*)ptr)[1] ^= ((uint64_t*)ptr)[0]; \
} \
} else { \
@@ -1681,13 +1653,13 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
int32_t d = ((int32_t*)&l[idx & MASK])[2]; \
int64_t q = n / (d | 0x5); \
((int64_t*)&l[idx & MASK])[0] = n ^ q; \
if (IS_CN_HEAVY_XHV) { \
if (props.isHeavyXHV()) { \
d = ~d; \
} \
\
idx = d ^ q; \
} \
if (BASE == Algorithm::CN_2) { \
if (props.isBase2()) { \
b1 = b0; \
} \
b0 = c;
@@ -1697,11 +1669,11 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
__m128i mc##n; \
__m128i division_result_xmm_##n; \
__m128i sqrt_result_xmm_##n; \
if (BASE == Algorithm::CN_1) { \
if (props.isBase1()) { \
mc##n = _mm_set_epi64x(*reinterpret_cast<const uint64_t*>(input + n * size + 35) ^ \
*(reinterpret_cast<const uint64_t*>((ctx)->state) + 24), 0); \
} \
if (BASE == Algorithm::CN_2) { \
if (props.isBase2()) { \
division_result_xmm_##n = _mm_cvtsi64_si128(h##n[12]); \
sqrt_result_xmm_##n = _mm_cvtsi64_si128(h##n[13]); \
} \
@@ -1710,7 +1682,7 @@ static NOINLINE void cryptonight_quad_hash_gr_sse41(const uint8_t* __restrict__
__m128i bx##n##1 = _mm_set_epi64x(h##n[9] ^ h##n[11], h##n[8] ^ h##n[10]); \
__m128i cx##n = _mm_setzero_si128(); \
__m128 conc_var##n; \
if (ALGO == Algorithm::CN_CCX) { \
if (props.isCCX()) { \
conc_var##n = _mm_setzero_ps(); \
} \
VARIANT4_RANDOM_MATH_INIT(n);
@@ -1721,17 +1693,8 @@ inline void cryptonight_triple_hash(const uint8_t *__restrict__ input, size_t si
{
constexpr CnAlgo<ALGO> props;
constexpr size_t MASK = props.mask();
constexpr Algorithm::Id BASE = props.base();
# ifdef XMRIG_ALGO_CN_HEAVY
constexpr bool IS_CN_HEAVY_TUBE = ALGO == Algorithm::CN_HEAVY_TUBE;
constexpr bool IS_CN_HEAVY_XHV = ALGO == Algorithm::CN_HEAVY_XHV;
# else
constexpr bool IS_CN_HEAVY_TUBE = false;
constexpr bool IS_CN_HEAVY_XHV = false;
# endif
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32 * 3);
return;
}
@@ -1755,7 +1718,7 @@ inline void cryptonight_triple_hash(const uint8_t *__restrict__ input, size_t si
CONST_INIT(ctx[1], 1);
CONST_INIT(ctx[2], 2);
VARIANT2_SET_ROUNDING_MODE();
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
RESTORE_ROUNDING_MODE();
}
@@ -1819,17 +1782,8 @@ inline void cryptonight_quad_hash(const uint8_t *__restrict__ input, size_t size
constexpr CnAlgo<ALGO> props;
constexpr size_t MASK = props.mask();
constexpr Algorithm::Id BASE = props.base();
# ifdef XMRIG_ALGO_CN_HEAVY
constexpr bool IS_CN_HEAVY_TUBE = ALGO == Algorithm::CN_HEAVY_TUBE;
constexpr bool IS_CN_HEAVY_XHV = ALGO == Algorithm::CN_HEAVY_XHV;
# else
constexpr bool IS_CN_HEAVY_TUBE = false;
constexpr bool IS_CN_HEAVY_XHV = false;
# endif
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32 * 4);
return;
}
@@ -1869,7 +1823,7 @@ inline void cryptonight_quad_hash(const uint8_t *__restrict__ input, size_t size
CONST_INIT(ctx[2], 2);
CONST_INIT(ctx[3], 3);
VARIANT2_SET_ROUNDING_MODE();
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
RESTORE_ROUNDING_MODE();
}
@@ -1930,17 +1884,8 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz
{
constexpr CnAlgo<ALGO> props;
constexpr size_t MASK = props.mask();
constexpr Algorithm::Id BASE = props.base();
# ifdef XMRIG_ALGO_CN_HEAVY
constexpr bool IS_CN_HEAVY_TUBE = ALGO == Algorithm::CN_HEAVY_TUBE;
constexpr bool IS_CN_HEAVY_XHV = ALGO == Algorithm::CN_HEAVY_XHV;
# else
constexpr bool IS_CN_HEAVY_TUBE = false;
constexpr bool IS_CN_HEAVY_XHV = false;
# endif
if (BASE == Algorithm::CN_1 && size < 43) {
if (props.isBase1() && size < 43) {
memset(output, 0, 32 * 5);
return;
}
@@ -1970,7 +1915,7 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz
CONST_INIT(ctx[3], 3);
CONST_INIT(ctx[4], 4);
VARIANT2_SET_ROUNDING_MODE();
if (ALGO == Algorithm::CN_CCX) {
if (props.isCCX()) {
RESTORE_ROUNDING_MODE();
}

View File

@@ -22,7 +22,7 @@
#define APP_ID "xmrig"
#define APP_NAME "XMRig"
#define APP_DESC "XMRig miner"
#define APP_VERSION "6.22.2"
#define APP_VERSION "6.22.3-dev"
#define APP_DOMAIN "xmrig.com"
#define APP_SITE "www.xmrig.com"
#define APP_COPYRIGHT "Copyright (C) 2016-2024 xmrig.com"
@@ -30,7 +30,7 @@
#define APP_VER_MAJOR 6
#define APP_VER_MINOR 22
#define APP_VER_PATCH 2
#define APP_VER_PATCH 3
#ifdef _MSC_VER
# if (_MSC_VER >= 1930)