chat: implement basic UI backend selection

Signed-off-by: Jared Van Bortel <jared@nomic.ai>

gpt4all-backend/llmodel.h

@@ -10,6 +10,8 @@
 #include <string_view>
 #include <vector>
 
+using namespace std::string_literals;
+
 #define LLMODEL_MAX_PROMPT_BATCH 128
 
 class Dlhandle;
@@ -51,6 +53,8 @@ class LLModel {
         GPUDevice(const char *backend, int index, int type, size_t heapSize, std::string name, std::string vendor):
             backend(backend), index(index), type(type), heapSize(heapSize), name(std::move(name)),
             vendor(std::move(vendor)) {}
+
+        std::string uiName() const { return backend == "cuda"s ? "CUDA: " + name : name; }
     };
 
     class Implementation {

gpt4all-chat/chatllm.cpp

@@ -302,19 +302,23 @@ bool ChatLLM::loadModel(const ModelInfo &modelInfo)
             QElapsedTimer modelLoadTimer;
             modelLoadTimer.start();
 
+            auto requestedDevice = MySettings::globalInstance()->device();
             auto n_ctx = MySettings::globalInstance()->modelContextLength(modelInfo);
             m_ctx.n_ctx = n_ctx;
             auto ngl = MySettings::globalInstance()->modelGpuLayers(modelInfo);
 
-            std::string buildVariant = "auto";
-#if defined(Q_OS_MAC) && defined(__arm__)
+            std::string backend = "auto";
+#if !defined(Q_OS_MAC)
+            if (requestedDevice.startsWith("CUDA: "))
+                backend = "cuda";
+#elif defined(__arm__)
             if (m_forceMetal)
-                buildVariant = "metal";
+                backend = "metal";
 #endif
             QString constructError;
             m_llModelInfo.model = nullptr;
             try {
-                m_llModelInfo.model = LLModel::Implementation::construct(filePath.toStdString(), buildVariant, n_ctx);
+                m_llModelInfo.model = LLModel::Implementation::construct(filePath.toStdString(), backend, n_ctx);
             } catch (const LLModel::MissingImplementationError &e) {
                 modelLoadProps.insert("error", "missing_model_impl");
                 constructError = e.what();
@@ -346,19 +350,19 @@ bool ChatLLM::loadModel(const ModelInfo &modelInfo)
 
                 // Pick the best match for the device
                 QString actualDevice = m_llModelInfo.model->implementation().buildVariant() == "metal" ? "Metal" : "CPU";
-                const QString requestedDevice = MySettings::globalInstance()->device();
                 if (requestedDevice == "CPU") {
                     emit reportFallbackReason(""); // fallback not applicable
                 } else {
                     const size_t requiredMemory = m_llModelInfo.model->requiredMem(filePath.toStdString(), n_ctx, ngl);
                     std::vector<LLModel::GPUDevice> availableDevices = m_llModelInfo.model->availableGPUDevices(requiredMemory);
                     LLModel::GPUDevice *device = nullptr;
 
+                    // NB: relies on the fact that Kompute devices are listed first
                     if (!availableDevices.empty() && requestedDevice == "Auto" && availableDevices.front().type == 2 /*a discrete gpu*/) {
                         device = &availableDevices.front();
                     } else {
                         for (LLModel::GPUDevice &d : availableDevices) {
-                            if (QString::fromStdString(d.name) == requestedDevice) {
+                            if (QString::fromStdString(d.uiName()) == requestedDevice) {
                                 device = &d;
                                 break;
                             }

gpt4all-chat/mysettings.cpp

@@ -68,7 +68,7 @@ MySettings::MySettings()
     std::vector<LLModel::GPUDevice> devices = LLModel::Implementation::availableGPUDevices();
     QVector<QString> deviceList{ "Auto" };
     for (LLModel::GPUDevice &d : devices)
-        deviceList << QString::fromStdString(d.name);
+        deviceList << QString::fromStdString(d.uiName());
     deviceList << "CPU";
     setDeviceList(deviceList);
 }