Merge 14128cbdb4 into c5d8b8265b

Fix number of threads on the new Intel Core Ultra CPUs

src/backend/common/Hashrate.cpp

@@ -30,10 +30,10 @@
 #include "base/tools/Handle.h"
 
 
-inline static const char *format(double h, char *buf, size_t size)
+inline static const char *format(std::pair<bool, double> h, char *buf, size_t size)
 {
-    if (std::isnormal(h)) {
-        snprintf(buf, size, (h < 100.0) ? "%04.2f" : "%03.1f", h);
+    if (h.first) {
+        snprintf(buf, size, (h.second < 100.0) ? "%04.2f" : "%03.1f", h.second);
         return buf;
     }
 
@@ -80,15 +80,16 @@ double xmrig::Hashrate::average() const
 }
 
 
-const char *xmrig::Hashrate::format(double h, char *buf, size_t size)
+const char *xmrig::Hashrate::format(std::pair<bool, double> h, char *buf, size_t size)
 {
     return ::format(h, buf, size);
 }
 
 
-rapidjson::Value xmrig::Hashrate::normalize(double d)
+rapidjson::Value xmrig::Hashrate::normalize(std::pair<bool, double> d)
 {
-    return Json::normalize(d, false);
+    using namespace rapidjson;
+    return d.first ? Value(floor(d.second * 100.0) / 100.0) : Value(kNullType);
 }
 
 
@@ -122,11 +123,11 @@ rapidjson::Value xmrig::Hashrate::toJSON(size_t threadId, rapidjson::Document &d
 #endif
 
 
-double xmrig::Hashrate::hashrate(size_t index, size_t ms) const
+std::pair<bool, double> xmrig::Hashrate::hashrate(size_t index, size_t ms) const
 {
     assert(index < m_threads);
     if (index >= m_threads) {
-        return nan("");
+        return { false, 0.0 };
     }
 
     uint64_t earliestHashCount = 0;
@@ -157,17 +158,27 @@ double xmrig::Hashrate::hashrate(size_t index, size_t ms) const
     } while (idx != idx_start);
 
     if (!haveFullSet || earliestStamp == 0 || lastestStamp == 0) {
-        return nan("");
+        return { false, 0.0 };
     }
 
-    if (lastestStamp - earliestStamp == 0) {
-        return nan("");
+    if (lastestHashCnt == earliestHashCount) {
+        return { true, 0.0 };
+    }
+
+    if (lastestStamp == earliestStamp) {
+        return { false, 0.0 };
     }
 
     const auto hashes = static_cast<double>(lastestHashCnt - earliestHashCount);
-    const auto time   = static_cast<double>(lastestStamp - earliestStamp) / 1000.0;
+    const auto time   = static_cast<double>(lastestStamp - earliestStamp);
 
-    return hashes / time;
+    const auto hr = hashes * 1000.0 / time;
+
+    if (!std::isnormal(hr)) {
+        return { false, 0.0 };
+    }
+
+    return { true, hr };
 }
 
 

src/backend/common/Hashrate.h

@@ -47,16 +47,16 @@ public:
     Hashrate(size_t threads);
     ~Hashrate();
 
-    inline double calc(size_t ms) const                                     { const double data = hashrate(0U, ms); return std::isnormal(data) ? data : 0.0; }
-    inline double calc(size_t threadId, size_t ms) const                    { return hashrate(threadId + 1, ms); }
+    inline std::pair<bool, double> calc(size_t ms) const                    { return hashrate(0U, ms); }
+    inline std::pair<bool, double> calc(size_t threadId, size_t ms) const   { return hashrate(threadId + 1, ms); }
     inline size_t threads() const                                           { return m_threads > 0U ? m_threads - 1U : 0U; }
     inline void add(size_t threadId, uint64_t count, uint64_t timestamp)    { addData(threadId + 1U, count, timestamp); }
     inline void add(uint64_t count, uint64_t timestamp)                     { addData(0U, count, timestamp); }
 
     double average() const;
 
-    static const char *format(double h, char *buf, size_t size);
-    static rapidjson::Value normalize(double d);
+    static const char *format(std::pair<bool, double> h, char *buf, size_t size);
+    static rapidjson::Value normalize(std::pair<bool, double> d);
 
 #   ifdef XMRIG_FEATURE_API
     rapidjson::Value toJSON(rapidjson::Document &doc) const;
@@ -64,7 +64,7 @@ public:
 #   endif
 
 private:
-    double hashrate(size_t index, size_t ms) const;
+    std::pair<bool, double> hashrate(size_t index, size_t ms) const;
     void addData(size_t index, uint64_t count, uint64_t timestamp);
 
     constexpr static size_t kBucketSize = 2 << 11;

src/backend/cpu/interfaces/ICpuInfo.h

@@ -29,6 +29,10 @@
 #ifdef XMRIG_FEATURE_HWLOC
 using hwloc_const_bitmap_t  = const struct hwloc_bitmap_s *;
 using hwloc_topology_t      = struct hwloc_topology *;
+#define MEMBIND_SUCCESS    1
+#define MEMBIND_FAIL_SUPP -1
+#define MEMBIND_FAIL_NODE -2
+#define MEMBIND_FAIL_BIND -3
 #endif
 
 
@@ -126,7 +130,7 @@ public:
     virtual uint32_t model() const                                                  = 0;
 
 #   ifdef XMRIG_FEATURE_HWLOC
-    virtual bool membind(hwloc_const_bitmap_t nodeset)                              = 0;
+    virtual int8_t membind(hwloc_const_bitmap_t nodeset)                            = 0;
     virtual const std::vector<uint32_t> &nodeset() const                            = 0;
     virtual hwloc_topology_t topology() const                                       = 0;
 #   endif

src/backend/cpu/platform/HwlocCpuInfo.cpp

@@ -23,6 +23,7 @@
 
 #include <algorithm>
 #include <cmath>
+#include <errno.h>
 #include <hwloc.h>
 
 
@@ -191,16 +192,25 @@ xmrig::HwlocCpuInfo::~HwlocCpuInfo()
 }
 
 
-bool xmrig::HwlocCpuInfo::membind(hwloc_const_bitmap_t nodeset)
+int8_t xmrig::HwlocCpuInfo::membind(hwloc_const_bitmap_t nodeset)
 {
     if (!hwloc_topology_get_support(m_topology)->membind->set_thisthread_membind) {
-        return false;
+        return MEMBIND_FAIL_SUPP;
     }
 
 #   if HWLOC_API_VERSION >= 0x20000
-    return hwloc_set_membind(m_topology, nodeset, HWLOC_MEMBIND_BIND, HWLOC_MEMBIND_THREAD | HWLOC_MEMBIND_BYNODESET) >= 0;
+    int rv = hwloc_set_membind(m_topology, nodeset, HWLOC_MEMBIND_BIND, HWLOC_MEMBIND_THREAD | HWLOC_MEMBIND_BYNODESET);
+    int error = errno;
+
+    if (rv < 0) {
+        LOG_WARN("hwloc_set_membind() error: \"%s\"\n", strerror(error));
+        return MEMBIND_FAIL_BIND;
+    }
+
+    return MEMBIND_SUCCESS;
 #   else
-    return hwloc_set_membind_nodeset(m_topology, nodeset, HWLOC_MEMBIND_BIND, HWLOC_MEMBIND_THREAD) >= 0;
+    return (hwloc_set_membind_nodeset(m_topology, nodeset, HWLOC_MEMBIND_BIND, HWLOC_MEMBIND_THREAD) >= 0)
+        ? MEMBIND_SUCCESS : MEMBIND_FAIL_BIND;
 #   endif
 }
 
@@ -342,7 +352,7 @@ void xmrig::HwlocCpuInfo::processTopLevelCache(hwloc_obj_t cache, const Algorith
     }
 
 #   ifdef XMRIG_ALGO_RANDOMX
-    if ((algorithm.family() == Algorithm::RANDOM_X) && L3_exclusive && (PUs > cores.size()) && (PUs < cores.size() * 2)) {
+    if ((vendor() == VENDOR_INTEL) && (algorithm.family() == Algorithm::RANDOM_X) && L3_exclusive && (PUs < cores.size() * 2)) {
         // Use all L3+L2 on latest Intel CPUs with P-cores, E-cores and exclusive L3 cache
         cacheHashes = (L3 + L2) / scratchpad;
     }

src/backend/cpu/platform/HwlocCpuInfo.h

@@ -38,7 +38,7 @@ public:
     ~HwlocCpuInfo() override;
 
 protected:
-    bool membind(hwloc_const_bitmap_t nodeset) override;
+    int8_t membind(hwloc_const_bitmap_t nodeset) override;
     CpuThreads threads(const Algorithm &algorithm, uint32_t limit) const override;
 
     inline const char *backend() const override                     { return m_backend; }

src/backend/cuda/CudaBackend.cpp

@@ -372,15 +372,20 @@ void xmrig::CudaBackend::printHashrate(bool details)
 
     char num[16 * 3] = { 0 };
 
-    const double hashrate_short  = hashrate()->calc(Hashrate::ShortInterval);
-    const double hashrate_medium = hashrate()->calc(Hashrate::MediumInterval);
-    const double hashrate_large  = hashrate()->calc(Hashrate::LargeInterval);
+    auto hashrate_short  = hashrate()->calc(Hashrate::ShortInterval);
+    auto hashrate_medium = hashrate()->calc(Hashrate::MediumInterval);
+    auto hashrate_large  = hashrate()->calc(Hashrate::LargeInterval);
 
     double scale = 1.0;
     const char* h = " H/s";
 
-    if ((hashrate_short >= 1e6) || (hashrate_medium >= 1e6) || (hashrate_large >= 1e6)) {
+    if ((hashrate_short.second >= 1e6) || (hashrate_medium.second >= 1e6) || (hashrate_large.second >= 1e6)) {
         scale = 1e-6;
+
+        hashrate_short.second  *= scale;
+        hashrate_medium.second *= scale;
+        hashrate_large.second  *= scale;
+
         h = "MH/s";
     }
 
@@ -388,12 +393,20 @@ void xmrig::CudaBackend::printHashrate(bool details)
 
     size_t i = 0;
     for (const auto& data : d_ptr->threads) {
-         Log::print("| %8zu | %8" PRId64 " | %8s | %8s | %8s |" CYAN_BOLD(" #%u") YELLOW(" %s") GREEN(" %s"),
+        auto h0 = hashrate()->calc(i, Hashrate::ShortInterval);
+        auto h1 = hashrate()->calc(i, Hashrate::MediumInterval);
+        auto h2 = hashrate()->calc(i, Hashrate::LargeInterval);
+
+        h0.second *= scale;
+        h1.second *= scale;
+        h2.second *= scale;
+
+        Log::print("| %8zu | %8" PRId64 " | %8s | %8s | %8s |" CYAN_BOLD(" #%u") YELLOW(" %s") GREEN(" %s"),
                     i,
                     data.thread.affinity(),
-                    Hashrate::format(hashrate()->calc(i, Hashrate::ShortInterval)  * scale, num,          sizeof num / 3),
-                    Hashrate::format(hashrate()->calc(i, Hashrate::MediumInterval) * scale, num + 16,      sizeof num / 3),
-                    Hashrate::format(hashrate()->calc(i, Hashrate::LargeInterval)  * scale, num + 16 * 2, sizeof num / 3),
+                    Hashrate::format(h0, num,          sizeof num / 3),
+                    Hashrate::format(h1, num + 16,     sizeof num / 3),
+                    Hashrate::format(h2, num + 16 * 2, sizeof num / 3),
                     data.device.index(),
                     data.device.topology().toString().data(),
                     data.device.name().data()
@@ -403,9 +416,9 @@ void xmrig::CudaBackend::printHashrate(bool details)
     }
 
     Log::print(WHITE_BOLD_S "|        - |        - | %8s | %8s | %8s |",
-               Hashrate::format(hashrate_short  * scale, num,          sizeof num / 3),
-               Hashrate::format(hashrate_medium * scale, num + 16,     sizeof num / 3),
-               Hashrate::format(hashrate_large  * scale, num + 16 * 2, sizeof num / 3)
+               Hashrate::format(hashrate_short , num,          sizeof num / 3),
+               Hashrate::format(hashrate_medium, num + 16,     sizeof num / 3),
+               Hashrate::format(hashrate_large , num + 16 * 2, sizeof num / 3)
                );
 }
 

src/backend/opencl/OclBackend.cpp

@@ -352,15 +352,20 @@ void xmrig::OclBackend::printHashrate(bool details)
 
     char num[16 * 3] = { 0 };
 
-    const double hashrate_short  = hashrate()->calc(Hashrate::ShortInterval);
-    const double hashrate_medium = hashrate()->calc(Hashrate::MediumInterval);
-    const double hashrate_large  = hashrate()->calc(Hashrate::LargeInterval);
+    auto hashrate_short = hashrate()->calc(Hashrate::ShortInterval);
+    auto hashrate_medium = hashrate()->calc(Hashrate::MediumInterval);
+    auto hashrate_large = hashrate()->calc(Hashrate::LargeInterval);
 
     double scale = 1.0;
     const char* h = " H/s";
 
-    if ((hashrate_short >= 1e6) || (hashrate_medium >= 1e6) || (hashrate_large >= 1e6)) {
+    if ((hashrate_short.second >= 1e6) || (hashrate_medium.second >= 1e6) || (hashrate_large.second >= 1e6)) {
         scale = 1e-6;
+
+        hashrate_short.second  *= scale;
+        hashrate_medium.second *= scale;
+        hashrate_large.second  *= scale;
+
         h = "MH/s";
     }
 
@@ -368,12 +373,16 @@ void xmrig::OclBackend::printHashrate(bool details)
 
     size_t i = 0;
     for (const auto& data : d_ptr->threads) {
-         Log::print("| %8zu | %8" PRId64 " | %8s | %8s | %8s |" CYAN_BOLD(" #%u") YELLOW(" %s") " %s",
+        auto h0 = hashrate()->calc(i, Hashrate::ShortInterval);
+        auto h1 = hashrate()->calc(i, Hashrate::MediumInterval);
+        auto h2 = hashrate()->calc(i, Hashrate::LargeInterval);
+
+        Log::print("| %8zu | %8" PRId64 " | %8s | %8s | %8s |" CYAN_BOLD(" #%u") YELLOW(" %s") " %s",
                     i,
                     data.affinity,
-                    Hashrate::format(hashrate()->calc(i, Hashrate::ShortInterval)  * scale, num,          sizeof num / 3),
-                    Hashrate::format(hashrate()->calc(i, Hashrate::MediumInterval) * scale, num + 16,     sizeof num / 3),
-                    Hashrate::format(hashrate()->calc(i, Hashrate::LargeInterval)  * scale, num + 16 * 2, sizeof num / 3),
+                    Hashrate::format(h0, num, sizeof num / 3),
+                    Hashrate::format(h1, num + 16, sizeof num / 3),
+                    Hashrate::format(h2, num + 16 * 2, sizeof num / 3),
                     data.device.index(),
                     data.device.topology().toString().data(),
                     data.device.printableName().data()
@@ -383,9 +392,9 @@ void xmrig::OclBackend::printHashrate(bool details)
     }
 
     Log::print(WHITE_BOLD_S "|        - |        - | %8s | %8s | %8s |",
-               Hashrate::format(hashrate_short  * scale, num,          sizeof num / 3),
-               Hashrate::format(hashrate_medium * scale, num + 16,     sizeof num / 3),
-               Hashrate::format(hashrate_large  * scale, num + 16 * 2, sizeof num / 3)
+               Hashrate::format(hashrate_short , num,          sizeof num / 3),
+               Hashrate::format(hashrate_medium, num + 16,     sizeof num / 3),
+               Hashrate::format(hashrate_large , num + 16 * 2, sizeof num / 3)
                );
 }
 

src/core/Miner.cpp

@@ -173,7 +173,7 @@ public:
         Value total(kArrayType);
         Value threads(kArrayType);
 
-        double t[3] = { 0.0 };
+        std::pair<bool, double> t[3] = { { true, 0.0 }, { true, 0.0 }, { true, 0.0 } };
 
         for (IBackend *backend : backends) {
             const Hashrate *hr = backend->hashrate();
@@ -181,9 +181,13 @@ public:
                 continue;
             }
 
-            t[0] += hr->calc(Hashrate::ShortInterval);
-            t[1] += hr->calc(Hashrate::MediumInterval);
-            t[2] += hr->calc(Hashrate::LargeInterval);
+            const auto h0 = hr->calc(Hashrate::ShortInterval);
+            const auto h1 = hr->calc(Hashrate::MediumInterval);
+            const auto h2 = hr->calc(Hashrate::LargeInterval);
+
+            if (h0.first) { t[0].second += h0.second; } else { t[0].first = false; }
+            if (h1.first) { t[1].second += h1.second; } else { t[1].first = false; }
+            if (h2.first) { t[2].second += h2.second; } else { t[2].first = false; }
 
             if (version > 1) {
                 continue;
@@ -204,7 +208,7 @@ public:
         total.PushBack(Hashrate::normalize(t[2]),  allocator);
 
         hashrate.AddMember("total",   total, allocator);
-        hashrate.AddMember("highest", Hashrate::normalize(maxHashrate[algorithm]), allocator);
+        hashrate.AddMember("highest", Hashrate::normalize({ maxHashrate[algorithm] > 0.0, maxHashrate[algorithm] }), allocator);
 
         if (version == 1) {
             hashrate.AddMember("threads", threads, allocator);
@@ -283,7 +287,7 @@ public:
     void printHashrate(bool details)
     {
         char num[16 * 5] = { 0 };
-        double speed[3]  = { 0.0 };
+        std::pair<bool, double> speed[3] = { { true, 0.0 }, { true, 0.0 }, { true, 0.0 } };
         uint32_t count   = 0;
 
         double avg_hashrate = 0.0;
@@ -293,9 +297,13 @@ public:
             if (hashrate) {
                 ++count;
 
-                speed[0] += hashrate->calc(Hashrate::ShortInterval);
-                speed[1] += hashrate->calc(Hashrate::MediumInterval);
-                speed[2] += hashrate->calc(Hashrate::LargeInterval);
+                const auto h0 = hashrate->calc(Hashrate::ShortInterval);
+                const auto h1 = hashrate->calc(Hashrate::MediumInterval);
+                const auto h2 = hashrate->calc(Hashrate::LargeInterval);
+
+                if (h0.first) { speed[0].second += h0.second; } else { speed[0].first = false; }
+                if (h1.first) { speed[1].second += h1.second; } else { speed[1].first = false; }
+                if (h2.first) { speed[2].second += h2.second; } else { speed[2].first = false; }
 
                 avg_hashrate += hashrate->average();
             }
@@ -312,8 +320,13 @@ public:
         double scale  = 1.0;
         const char* h = "H/s";
 
-        if ((speed[0] >= 1e6) || (speed[1] >= 1e6) || (speed[2] >= 1e6) || (maxHashrate[algorithm] >= 1e6)) {
+        if ((speed[0].second >= 1e6) || (speed[1].second >= 1e6) || (speed[2].second >= 1e6) || (maxHashrate[algorithm] >= 1e6)) {
             scale = 1e-6;
+
+            speed[0].second *= scale;
+            speed[1].second *= scale;
+            speed[2].second *= scale;
+
             h = "MH/s";
         }
 
@@ -322,16 +335,16 @@ public:
 
 #       ifdef XMRIG_ALGO_GHOSTRIDER
         if (algorithm.family() == Algorithm::GHOSTRIDER) {
-            snprintf(avg_hashrate_buf, sizeof(avg_hashrate_buf), " avg " CYAN_BOLD("%s %s"), Hashrate::format(avg_hashrate * scale, num + 16 * 4, 16), h);
+            snprintf(avg_hashrate_buf, sizeof(avg_hashrate_buf), " avg " CYAN_BOLD("%s %s"), Hashrate::format({ true, avg_hashrate * scale }, num + 16 * 4, 16), h);
         }
 #       endif
 
         LOG_INFO("%s " WHITE_BOLD("speed") " 10s/60s/15m " CYAN_BOLD("%s") CYAN(" %s %s ") CYAN_BOLD("%s") " max " CYAN_BOLD("%s %s") "%s",
                  Tags::miner(),
-                 Hashrate::format(speed[0] * scale,                 num,          16),
-                 Hashrate::format(speed[1] * scale,                 num + 16,     16),
-                 Hashrate::format(speed[2] * scale,                 num + 16 * 2, 16), h,
-                 Hashrate::format(maxHashrate[algorithm] * scale,   num + 16 * 3, 16), h,
+                 Hashrate::format(speed[0],                 num,          16),
+                 Hashrate::format(speed[1],                 num + 16,     16),
+                 Hashrate::format(speed[2],                 num + 16 * 2, 16), h,
+                 Hashrate::format({ maxHashrate[algorithm] > 0.0, maxHashrate[algorithm] * scale },   num + 16 * 3, 16), h,
                  avg_hashrate_buf
                  );
 
@@ -646,7 +659,10 @@ void xmrig::Miner::onTimer(const Timer *)
         }
 
         if (backend->hashrate()) {
-            maxHashrate += backend->hashrate()->calc(Hashrate::ShortInterval);
+            const auto h = backend->hashrate()->calc(Hashrate::ShortInterval);
+            if (h.first) {
+                maxHashrate += h.second;
+            }
         }
     }
 

src/crypto/common/VirtualMemory_hwloc.cpp

@@ -34,8 +34,13 @@ uint32_t xmrig::VirtualMemory::bindToNUMANode(int64_t affinity)
 
     auto pu = hwloc_get_pu_obj_by_os_index(Cpu::info()->topology(), static_cast<unsigned>(affinity));
 
-    if (pu == nullptr || !Cpu::info()->membind(pu->nodeset)) {
-        LOG_WARN("CPU #%02" PRId64 " warning: \"can't bind memory\"", affinity);
+    if (pu == nullptr) {
+        LOG_WARN("CPU #%02" PRId64 " warning: \"can't bind memory: hwloc_get_pu_obj_by_os_index() failed\"", affinity);
+
+        return 0;
+    }
+    if (Cpu::info()->membind(pu->nodeset) < 0) {
+        LOG_WARN("CPU #%02" PRId64 " warning: \"can't bind memory: Cpu::info()->membind() failed\"", affinity);
 
         return 0;
     }

src/crypto/rx/RxNUMAStorage.cpp

@@ -42,20 +42,20 @@ constexpr size_t oneMiB = 1024 * 1024;
 static std::mutex mutex;
 
 
-static bool bindToNUMANode(uint32_t nodeId)
+static int8_t bindToNUMANode(uint32_t nodeId)
 {
     auto node = hwloc_get_numanode_obj_by_os_index(Cpu::info()->topology(), nodeId);
     if (!node) {
-        return false;
+        return MEMBIND_FAIL_NODE;
     }
 
-    if (Cpu::info()->membind(node->nodeset)) {
+    if (Cpu::info()->membind(node->nodeset) > 0) {
         Platform::setThreadAffinity(static_cast<uint64_t>(hwloc_bitmap_first(node->cpuset)));
 
-        return true;
+        return MEMBIND_SUCCESS;
     }
 
-    return false;
+    return MEMBIND_FAIL_BIND;
 }
 
 
@@ -210,10 +210,20 @@ private:
     static void allocate(RxNUMAStoragePrivate *d_ptr, uint32_t nodeId, bool hugePages, bool oneGbPages)
     {
         const uint64_t ts = Chrono::steadyMSecs();
+        const int8_t   br = bindToNUMANode(nodeId);
 
-        if (!bindToNUMANode(nodeId)) {
-            printSkipped(nodeId, "can't bind memory");
-
+        if (br < 0) {
+            switch (br) {
+            case MEMBIND_FAIL_SUPP:
+                printSkipped(nodeId, "can't bind memory: hwloc_topology_get_support() failed");
+                break;
+            case MEMBIND_FAIL_NODE:
+                printSkipped(nodeId, "can't bind memory: hwloc_get_numanode_obj_by_os_index() failed");
+                break;
+            case MEMBIND_FAIL_BIND:
+                printSkipped(nodeId, "can't bind memory: Cpu::info()->membind() failed");
+                break;
+            }
             return;
         }
 

src/version.h

@@ -22,7 +22,7 @@
 #define APP_ID        "xmrig"
 #define APP_NAME      "XMRig"
 #define APP_DESC      "XMRig miner"
-#define APP_VERSION   "6.22.1"
+#define APP_VERSION   "6.22.2-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  1
+#define APP_VER_PATCH  2
 
 #ifdef _MSC_VER
 #   if (_MSC_VER >= 1930)