[feature] Graceful termination of background threads in LlamaV2 (#458)

* cuda allocator fix

* graceful termination

* lint and compilation fix

src/turbomind/models/llama/LlamaV2.cc

@@ -126,6 +126,7 @@ LlamaV2<T>::LlamaV2(size_t                       head_num,
 template<typename T>
 LlamaV2<T>::~LlamaV2()
 {
+    shared_state_->request_queue.close();
     internal_thread_.join();
 
     delete decoder_;
@@ -448,12 +449,24 @@ void LlamaV2<T>::internalThreadEntry(int device_id)
 
             request_queue.dequeue(stop_requests, infer_requests, free_slot_count, is_empty);
 
+            // request queue was closed
+            // and there are no unprocessed requests in the queue
+            if (is_empty && infer_requests.empty() && stop_requests.empty()) {
+                // rank 0 sets flag
+                shared_state_->should_stop = true;
+            }
+
             batch_.verifyRequests(stop_requests, infer_requests);
         }
 
         // wait while rank-0 is dequeueing
         shared_state_->barrier->wait();
 
+        // exit if job is done
+        if (shared_state_->should_stop) {
+            return;
+        }
+
         bool modified = false;
 
         if (!(batch_.finishedCount() == 0 && stop_requests.empty() && infer_requests.empty())) {
@@ -486,8 +499,6 @@ void LlamaV2<T>::internalThreadEntry(int device_id)
             batch_.finish();
         }
     }
-
-    FT_CHECK(0);
 }
 
 template<typename T>

src/turbomind/models/llama/LlamaV2.h

@@ -46,6 +46,9 @@ class LlamaV2 {
         std::vector<std::shared_ptr<Request>> stop_requests;
         RequestQueue                          request_queue;
         std::shared_ptr<Barrier>              barrier;
+
+        // rank 0 sets flag to true if there are no more tasks in the request_queue
+        bool should_stop = false;
     };
 
     ~LlamaV2();

src/turbomind/models/llama/Request.h

@@ -44,6 +44,11 @@ class RequestQueue {
         futures.reserve(requests.size());
         {
             std::lock_guard<std::mutex> lock(mutex_);
+
+            if (closed_) {
+                throw std::runtime_error("Queue is closed");
+            }
+
             for (auto& r : requests) {
                 futures.push_back(r->signal.get_future());
                 if (r->stop_flag) {
@@ -65,7 +70,7 @@ class RequestQueue {
     {
         std::unique_lock<std::mutex> lock(mutex_);
         if (blocking) {
-            cv_.wait(lock, [this] { return !(stop_queue_.empty() && infer_queue_.empty()); });
+            cv_.wait(lock, [this] { return !(stop_queue_.empty() && infer_queue_.empty() && closed_ == false); });
         }
 
         stop_requests.clear();
@@ -81,11 +86,19 @@ class RequestQueue {
         }
     }
 
+    void close()
+    {
+        std::lock_guard<std::mutex> lock(mutex_);
+        closed_ = true;
+        cv_.notify_all();
+    }
+
 private:
     std::queue<std::shared_ptr<Request>> stop_queue_;
     std::queue<std::shared_ptr<Request>> infer_queue_;
     std::mutex                           mutex_;
     std::condition_variable              cv_;
+    bool                                 closed_ = false;
 };
 
 }  // namespace turbomind

src/turbomind/utils/allocator.h

@@ -125,9 +125,15 @@ class Allocator;
 template<>
 class Allocator<AllocatorType::CUDA>: public IAllocator {
 private:
-    const int                          device_id_;
-    cudaStream_t                       stream_ = 0;  // initialize as default stream
-    std::unordered_map<void*, size_t>* pointer_mapping_;
+    enum class MemoryType
+    {
+        HOST,
+        DEVICE
+    };
+
+    const int                                                 device_id_;
+    cudaStream_t                                              stream_ = 0;  // initialize as default stream
+    std::unordered_map<void*, std::pair<size_t, MemoryType>>* pointer_mapping_;
 
     bool isExist(void* address) const
     {
@@ -136,10 +142,10 @@ class Allocator<AllocatorType::CUDA>: public IAllocator {
     ReallocType isReMalloc(void* address, size_t size) const
     {
         FT_CHECK(isExist(address));
-        if (pointer_mapping_->at(address) < size) {
+        if (pointer_mapping_->at(address).first < size) {
             return ReallocType::INCREASE;
         }
-        else if (pointer_mapping_->at(address) == size) {
+        else if (pointer_mapping_->at(address).first == size) {
             return ReallocType::REUSE;
         }
         else {
@@ -151,7 +157,7 @@ class Allocator<AllocatorType::CUDA>: public IAllocator {
     Allocator(int device_id): device_id_(device_id)
     {
         TM_LOG_DEBUG(__PRETTY_FUNCTION__);
-        pointer_mapping_ = new std::unordered_map<void*, size_t>();
+        pointer_mapping_ = new std::unordered_map<void*, std::pair<size_t, MemoryType>>();
 #if defined(CUDA_MEMORY_POOL_DISABLED)
         TM_LOG_WARNING(
             "Async cudaMalloc/Free is not supported before CUDA 11.2. Using Sync cudaMalloc/Free."
@@ -188,7 +194,9 @@ class Allocator<AllocatorType::CUDA>: public IAllocator {
     {
         TM_LOG_DEBUG(__PRETTY_FUNCTION__);
         while (!pointer_mapping_->empty()) {
-            free((void**)(&pointer_mapping_->begin()->first));
+            auto ptr           = pointer_mapping_->begin()->first;
+            auto size_and_type = pointer_mapping_->begin()->second;
+            free(&ptr, size_and_type.second == MemoryType::HOST);
         }
         delete pointer_mapping_;
     }
@@ -229,18 +237,19 @@ class Allocator<AllocatorType::CUDA>: public IAllocator {
         check_cuda_error(getSetDevice(o_device));
         TM_LOG_DEBUG("malloc buffer %p with size %ld", ptr, size);
 
-        pointer_mapping_->insert({getAddress(ptr), size});
+        pointer_mapping_->insert({getAddress(ptr), {size, is_host ? MemoryType::HOST : MemoryType::DEVICE}});
 
         return ptr;
     }
 
-    void free(void** ptr, bool is_host = false) const
+    void free(void** ptr, bool _ = false) const
     {
         TM_LOG_DEBUG(__PRETTY_FUNCTION__);
         void* address = getAddress(*ptr);
         if (*ptr != nullptr) {
             int o_device = 0;
             if (pointer_mapping_->count(address)) {
+                const auto is_host = pointer_mapping_->at(address).second == MemoryType::HOST;
                 TM_LOG_DEBUG("Free buffer %p", address);
                 check_cuda_error(getSetDevice(device_id_, &o_device));
                 if (is_host) {
@@ -361,7 +370,7 @@ class Allocator<AllocatorType::TF>: public IAllocator {
     {
         while (!pointer_mapping_->empty()) {
             void* ptr = pointer_mapping_->begin()->second.flat<uint8>().data();
-            free((void**)(&ptr));
+            free(&ptr);
         }
         pointer_mapping_->clear();
         delete pointer_mapping_;
@@ -454,7 +463,7 @@ class Allocator<AllocatorType::TH>: public IAllocator {
         TM_LOG_DEBUG(__PRETTY_FUNCTION__);
         while (!pointer_mapping_->empty()) {
             void* ptr = pointer_mapping_->begin()->second.data_ptr();
-            free((void**)(&ptr));
+            free(&ptr);
         }
         pointer_mapping_->clear();
         delete pointer_mapping_;