publish from 07ff8974 (gopub)

README.md

@@ -112,9 +112,10 @@ You may get the latest installers or binaries in [Releases](https://github.com/i
 ## Documentation
 * Refer to the [XPU Manager Installation Guide](doc/Install_guide.md) and for how to install/uninstall XPU Manager.
 * Refer to the [XPU-SMI Installation Guide](doc/smi_install_guide.md) and for how to install/uninstall XPU-SMI.
-* Refer to the [XPU Manager CLI User Guide](doc/CLI_user_guide.md) to start using XPU Manager.
-* Refer to the [XPU-SMI CLI User Guide](doc/smi_user_guide.md) to start using XPU-SMI.
-* Refer to the [XPU Manager amcmcli User Guide](doc/amcmcli_user_guide.md) to start using XPU Manager amcmcli.
+* Refer to the [XPU Manager CLI User Guide](doc/CLI_user_guide.md) to start to use XPU Manager.
+* Refer to the [XPU-SMI CLI User Guide](doc/smi_user_guide.md) to start to use XPU-SMI.
+* Refer to the [XPU Manager Windows CLI User Guide](doc/xpum_win_user_guide.md) to start to use XPU Manager Windows CLI.
+* Refer to the [XPU Manager amcmcli User Guide](doc/amcmcli_user_guide.md) to start to use XPU Manager amcmcli.
 * Refer to [DockerHub](https://hub.docker.com/r/intel/xpumanager) for a Docker container image that can be used as a Prometheus exporter in a Kubernetes environment.
 * Refer to [Building XPU Manager Installer](BUILDING.md) to build XPU Manager installer packages. 
 

VERSION

@@ -1 +1 @@
-1.2.2
+1.2.3

build.sh

@@ -36,14 +36,4 @@ cmake .. $ccache_opts $prefix_default $@
 make -j4
 
 echo "---------Create installation package-----------"
-cpack   
-
-if [ -f ~/password.sys_dcm ]; then
-    PackageName=$(cat package_file_name)
-    CSUser="ccr\\sys_dcm"
-    CSPwd=$(cat ~/password.sys_dcm)
-    echo "SignFile:${PackageName}" 
-    pushd "${WORK_DIR}"/install/tools/signfile
-    ./SignFile -vv -u "${CSUser}" -p "${CSPwd}" "${WORK_DIR}"/build/${PackageName}
-    popd
-fi
+cpack   

builder/requirements.txt

@@ -53,5 +53,5 @@ sphinxcontrib-qthelp==1.0.3
 sphinxcontrib-serializinghtml==1.1.5
 typing-extensions==4.1.1
 urllib3==1.26.9
-Werkzeug==2.0.3
+Werkzeug==2.2.3
 zipp==3.6.0

cli/CMakeLists.txt

@@ -66,6 +66,7 @@ include_directories(/usr/local/include)
 include_directories(${CMAKE_CURRENT_LIST_DIR}/src)
 include_directories(${CMAKE_CURRENT_LIST_DIR}/../core/include)
 include_directories(${CMAKE_CURRENT_LIST_DIR}/../core/src/api)
+include_directories(${CMAKE_CURRENT_LIST_DIR}/../core/src/amc)
 include_directories(${CMAKE_CURRENT_LIST_DIR}/../third_party/CLI11/include)
 include_directories(${CMAKE_CURRENT_LIST_DIR}/../third_party/json/include)
 if(EXISTS ${CMAKE_CURRENT_LIST_DIR}/test)

cli/src/cli_wrapper.cpp

@@ -111,8 +111,10 @@ int CLIWrapper::printResult(std::ostream &out) {
                     setenv("XPUM_METRICS", "0,4-38", 1);
                 else if (dump_comlet->dumpEUMetrics())
                     setenv("XPUM_METRICS", "0-31,36-38", 1);
-                else
+                else if(dump_comlet->dumpRASMetrics())
                     setenv("XPUM_METRICS", "0,4-31,36-38", 1);
+                else
+                    setenv("XPUM_METRICS", "0,4-19,29-31,36-38", 1);
             }
             if (comlet->getCommand().compare("dump") == 0 && std::dynamic_pointer_cast<ComletDump>(comlet)->dumpIdlePowerOnly()) {
                 this->coreStub = std::make_shared<LibCoreStub>(false);

cli/src/comlet_diagnostic.cpp

@@ -12,6 +12,7 @@
 #include "core_stub.h"
 #include "utility.h"
 #include "exit_code.h"
+#include <unordered_map>
 
 namespace xpum::cli {
 
@@ -56,26 +57,51 @@ static CharTableConfig ComletConfigDiagnosticDevice(R"({
     }]
 })"_json);
 
+static CharTableConfig ComletConfigSpecificDiagnosticDevice(R"({
+    "showTitleRow": false,
+    "columns": [{
+        "title": "none"
+    }, {
+        "title": "none"
+    }],
+    "rows": [{
+        "instance": "",
+        "cells": [
+            { "rowTitle": "Device ID" },
+            "device_id"
+        ]
+    }, {
+        "instance": "component_list[]",
+        "cells": [
+            { "value": "component_type" }, [
+            { "label": "Result", "value": "result" },
+            { "label": "Message", "value": "message" },
+            { "value": "process_list[]", "subrow": true, "subs": [
+                { "label": "  PID", "value": "process_id" },
+                { "label": "Command", "value": "process_name" }
+            ]},
+            { "value": "media_codec_list[]", "subrow": true, "subs": [
+                { "label": "", "value": "fps" }
+            ]}
+        ]]
+    }]
+})"_json);
+
 static CharTableConfig ComletConfigDiagnosticPreCheck(R"({
     "showTitleRow": true,
     "columns": [{
         "title": "Component",
         "size": 16
     }, {
-        "title": "Status"
+        "title": "Details"
     }],
     "rows": [{
-        "instance": "",
-        "cells": [[
-            { "rowTitle": "GPU" },
-            { "rowTitle": "Driver" },
-            { "rowTitle": "GPU Status" },
-            { "rowTitle": "CPU Status" }
-        ], [
-            { "value": "gpu_basic_info" },
-            { "value": "gpu_driver_info" },
-            { "value": "gpu_status_info" },
-            { "value": "cpu_status_info" }
+        "instance": "component_list[]",
+        "cells": [
+            { "value": "type" }, [
+            { "value": "error_details[]", "subrow": true, "subs": [
+                { "value": "field_value" }
+            ]}
         ]]
     }]
 })"_json);
@@ -109,6 +135,16 @@ void ComletDiagnostic::setupOptions() {
     auto stressFlag = addFlag("-s,--stress", this->opts->stress, "Stress the GPU(s) for the specified time");
     auto stressTimeOpt = addOption("--stresstime", this->opts->stressTime, "Stress time (in minutes)");
     auto preCheckOpt = addFlag("--precheck", this->opts->preCheck, "Do the precheck on the GPU and GPU driver");
+    auto onlyGPUOpt = addFlag("--gpu", this->opts->onlyGPU, "Show the GPU status only");
+
+    auto singleTestId = addOption("--singletest", this->opts->singleTestId,
+              "Selectively run a particular test\n\
+      1. Computation\n\
+      2. Memory Error\n\
+      3. Memory Bandwidth\n\
+      4. Media Codec\n\
+      5. PCIe Bandwidth\n\
+      6. Power");
 
     preCheckOpt->excludes(deviceIdOpt);
     preCheckOpt->excludes(level);
@@ -117,18 +153,22 @@ void ComletDiagnostic::setupOptions() {
     level->excludes(preCheckOpt);
     level->excludes(stressFlag);
     level->excludes(stressTimeOpt);
+    level->excludes(singleTestId);
+    singleTestId->excludes(level);
+
     deviceIdOpt->excludes(preCheckOpt);
     if (stressFlag == nullptr) {
         deviceIdOpt->needs(level);
     }
     stressTimeOpt->needs(stressFlag);
 
+    onlyGPUOpt->needs(preCheckOpt);
+
 #ifndef DAEMONLESS
     preCheckOpt->excludes(groupIdOpt);
     groupIdOpt->excludes(preCheckOpt);
     groupIdOpt->excludes(stressFlag);
     groupIdOpt->excludes(stressTimeOpt);
-    groupIdOpt->needs(level);
     stressFlag->needs(stressTimeOpt);
 #endif
 }
@@ -152,6 +192,13 @@ std::unique_ptr<nlohmann::json> ComletDiagnostic::run() {
         (*json)["errno"] = XPUM_CLI_ERROR_DIAGNOSTIC_INVALID_LEVEL;
         return json;
     }
+
+    if (this->opts->singleTestId != INT_MIN && (this->opts->singleTestId < 1 || this->opts->singleTestId > 6)) {
+        (*json)["error"] = "invalid single test";
+        (*json)["errno"] = XPUM_CLI_ERROR_DIAGNOSTIC_INVALID_SINGLE_TEST;
+        return json;
+    }
+
     if (this->opts->level >= 1 && this->opts->level <= 3) {
         if (this->opts->deviceIds[0] != "-1") {
             int targetId = -1;
@@ -163,19 +210,48 @@ std::unique_ptr<nlohmann::json> ComletDiagnostic::run() {
                     return convertResult;
                 }
             }
-            json = this->coreStub->runDiagnostics(targetId, this->opts->level, this->opts->rawComponentTypeStr);
+            json = this->coreStub->runDiagnostics(targetId, this->opts->level, -1, this->opts->rawJson);
             return json;
         } 
 #ifndef DAEMONLESS
         else if (this->opts->groupId > 0 && this->opts->groupId != UINT_MAX) {
-            json = this->coreStub->runDiagnosticsByGroup(this->opts->groupId, this->opts->level, this->opts->rawComponentTypeStr);
+            json = this->coreStub->runDiagnosticsByGroup(this->opts->groupId, this->opts->level, -1, this->opts->rawJson);
+            return json;
+        }
+#endif
+    }
+
+    std::unordered_map<int, int> testIdToType = {{1, XPUM_DIAG_PERFORMANCE_COMPUTATION}, 
+                                                 {2, XPUM_DIAG_MEMORY_ERROR}, 
+                                                 {3, XPUM_DIAG_PERFORMANCE_MEMORY_BANDWIDTH}, 
+                                                 {4, XPUM_DIAG_MEDIA_CODEC}, 
+                                                 {5, XPUM_DIAG_INTEGRATION_PCIE}, 
+                                                 {6, XPUM_DIAG_PERFORMANCE_POWER}};
+
+    if (this->opts->singleTestId >= 1 && this->opts->singleTestId <= 6) {
+        if (this->opts->deviceIds[0] != "-1") {
+            int targetId = -1;
+            if (isNumber(this->opts->deviceIds[0])) {
+                targetId = std::stoi(this->opts->deviceIds[0]);
+            } else {
+                auto convertResult = this->coreStub->getDeivceIdByBDF(this->opts->deviceIds[0].c_str(), &targetId);
+                if (convertResult->contains("error")) {
+                    return convertResult;
+                }
+            }
+            json = this->coreStub->runDiagnostics(targetId, -1, testIdToType[this->opts->singleTestId], this->opts->rawJson);
+            return json;
+        } 
+#ifndef DAEMONLESS
+        else if (this->opts->groupId > 0 && this->opts->groupId != UINT_MAX) {
+            json = this->coreStub->runDiagnosticsByGroup(this->opts->groupId, -1, testIdToType[this->opts->singleTestId], this->opts->rawJson);
             return json;
         }
 #endif
     }
 
     if (this->opts->preCheck) {
-        json = this->coreStub->getPreCheckInfo();
+        json = this->coreStub->getPreCheckInfo(this->opts->onlyGPU, this->opts->rawJson);
         return json;
     }
 
@@ -204,10 +280,13 @@ std::unique_ptr<nlohmann::json> ComletDiagnostic::run() {
     return json;
 }
 
-static void showDeviceDiagnostic(std::ostream &out, std::shared_ptr<nlohmann::json> json, bool precheck = false, const bool cont = false) {
-    if (precheck) {
+static void showDeviceDiagnostic(std::ostream &out, std::shared_ptr<nlohmann::json> json, ShowMode mode, const bool cont = false) {
+    if (mode == PRE_CHECK) {
         CharTable table(ComletConfigDiagnosticPreCheck, *json, cont);
         table.show(out);
+    } else if (mode == SINGLE_TEST) {
+        CharTable table(ComletConfigSpecificDiagnosticDevice, *json, cont);
+        table.show(out);
     } else {
         CharTable table(ComletConfigDiagnosticDevice, *json, cont);
         table.show(out);
@@ -232,9 +311,9 @@ static void showStreesedDevices(std::ostream &out, const std::vector<std::string
 }
 
 void ComletDiagnostic::getTableResult(std::ostream &out) {
-    this->opts->rawComponentTypeStr = false;
+    this->opts->rawJson = false;
     auto res = run();
-    this->opts->rawComponentTypeStr = true;
+    this->opts->rawJson = true;
     if (res->contains("error")) {
         out << "Error: " << (*res)["error"].get<std::string>() << std::endl;
         setExitCodeByJson(*res);
@@ -248,7 +327,10 @@ void ComletDiagnostic::getTableResult(std::ostream &out) {
         auto devices = (*json)["device_list"].get<std::vector<nlohmann::json>>();
         bool cont = false;
         for (auto device : devices) {
-            showDeviceDiagnostic(out, std::make_shared<nlohmann::json>(device), false, cont);
+            if (this->opts->level >= 1 && this->opts->level <= 3)
+                showDeviceDiagnostic(out, std::make_shared<nlohmann::json>(device), LEVEL_TEST, cont);
+            else
+                showDeviceDiagnostic(out, std::make_shared<nlohmann::json>(device), SINGLE_TEST, cont);
             cont = true;
         }
         return;
@@ -292,12 +374,15 @@ void ComletDiagnostic::getTableResult(std::ostream &out) {
     }
 
     if (isDeviceOperation()) {
-        showDeviceDiagnostic(out, json);
+        if (this->opts->level >= 1 && this->opts->level <= 3)
+            showDeviceDiagnostic(out, json, LEVEL_TEST);
+        else
+            showDeviceDiagnostic(out, json, SINGLE_TEST);
         return;
     }
 
     if (this->opts->preCheck) {
-        showDeviceDiagnostic(out, json, true);
+        showDeviceDiagnostic(out, json, PRE_CHECK);
         return;
     }
 }

cli/src/comlet_diagnostic.h

@@ -20,12 +20,20 @@ struct ComletDiagnosticOptions {
     uint32_t groupId = UINT_MAX;
 #endif
     int level = INT_MIN;
-    bool rawComponentTypeStr = true;
+    int singleTestId = INT_MIN;
+    bool rawJson = true;
     bool preCheck = false;
+    bool onlyGPU = false;
     uint32_t stressTime = 0;
     bool stress = false;
 };
 
+enum ShowMode {
+    LEVEL_TEST,
+    SINGLE_TEST,
+    PRE_CHECK
+};
+
 class ComletDiagnostic : public ComletBase {
    public:
     ComletDiagnostic() : ComletBase("diag", "Run some test suites to diagnose GPU.") {

cli/src/comlet_dump.cpp

@@ -53,6 +53,17 @@ bool ComletDump::dumpEUMetrics() {
     return false;
 }
 
+bool ComletDump::dumpRASMetrics() {
+    for (auto id : this->opts->metricsIdList) {
+        if ((id >= xpum_dump_type_t::XPUM_DUMP_RAS_ERROR_CAT_RESET && id <= xpum_dump_type_t::XPUM_DUMP_RAS_ERROR_CAT_CACHE_ERRORS_UNCORRECTABLE)
+         || id == xpum_dump_type_t::XPUM_DUMP_RAS_ERROR_CAT_NON_COMPUTE_ERRORS_CORRECTABLE
+         || id == xpum_dump_type_t::XPUM_DUMP_RAS_ERROR_CAT_NON_COMPUTE_ERRORS_UNCORRECTABLE) {
+            return true;
+        }
+    }
+    return false;
+}
+
 static std::string getFileValue(std::string file_name) {
     std::ifstream ifs(file_name);
     std::string content((std::istreambuf_iterator<char>(ifs)), (std::istreambuf_iterator<char>()));

cli/src/comlet_dump.h

@@ -81,6 +81,8 @@ class ComletDump : public ComletBase {
 
     bool dumpEUMetrics();
 
+    bool dumpRASMetrics();
+
     bool dumpIdlePowerOnly();
 
     std::unique_ptr<nlohmann::json> combineTileAndDeviceLevel(nlohmann::json rawJson);

cli/src/config.h.in

@@ -4,4 +4,4 @@
 #define CLI_VERSION_GIT_COMMIT "@GIT_COMMIT@"
 #define CLI_VERSION_IN_HELP "@PROJECT_VERSION_IN_HELP@"
 
-#define XPUM_CONFIG_DIR "@CPACK_PACKAGING_INSTALL_PREFIX@/config/"
+#define XPUM_CONFIG_DIR "@XPUM_CONFIG_DIR@/"