fix minor memory leak when querying by IVF_SQ8H (#4758)

* fix minor memory leak when querying by IVF_SQ8H

Signed-off-by: shengjun.li <shengjun.li@zilliz.com>

* fix mismatch malloc/free new/delete

Signed-off-by: shengjun.li <shengjun.li@zilliz.com>

CHANGELOG.md

@@ -6,6 +6,7 @@ Please mark all change in change log and use the issue from GitHub
 ## Bug
 -   \#4739 Fix mishards probe test problem
 -   \#4749 Fix minor memory leak when building IVF_SQ8 on GPU
+-   \#4757 Fix minor memory leak when querying by IVF_SQ8H
 
 ## Feature
 -   \#3977 Support logging to stdout

core/src/index/knowhere/knowhere/index/vector_index/gpu/IndexIVFSQHybrid.cpp

@@ -138,12 +138,12 @@ IVFSQHybrid::LoadData(const FaissIVFQuantizerPtr& quantizer_ptr, const Config& c
         if (ivf_quantizer == nullptr)
             KNOWHERE_THROW_MSG("quantizer type not faissivfquantizer");
 
-        auto index_composition = new faiss::IndexComposition;
-        index_composition->index = index_.get();
-        index_composition->quantizer = ivf_quantizer->quantizer;
-        index_composition->mode = 2;  // only 2
+        faiss::IndexComposition index_composition;
+        index_composition.index = index_.get();
+        index_composition.quantizer = ivf_quantizer->quantizer;
+        index_composition.mode = 2;  // only copy data
 
-        auto gpu_index = faiss::gpu::index_cpu_to_gpu(res->faiss_res.get(), gpu_id, index_composition, &option);
+        auto gpu_index = faiss::gpu::index_cpu_to_gpu(res->faiss_res.get(), gpu_id, &index_composition, &option);
         std::shared_ptr<faiss::Index> new_idx;
         new_idx.reset(gpu_index);
         auto sq_idx = std::make_shared<IVFSQHybrid>(new_idx, gpu_id, res);
@@ -162,17 +162,17 @@ IVFSQHybrid::LoadQuantizer(const Config& config) {
         faiss::gpu::GpuClonerOptions option;
         option.allInGpu = true;
 
-        auto index_composition = new faiss::IndexComposition;
-        index_composition->index = index_.get();
-        index_composition->quantizer = nullptr;
-        index_composition->mode = 1;  // only 1
+        faiss::IndexComposition index_composition;
+        index_composition.index = index_.get();
+        index_composition.quantizer = nullptr;
+        index_composition.mode = 1;  // only copy quantizer
 
-        auto gpu_index = faiss::gpu::index_cpu_to_gpu(res->faiss_res.get(), gpu_id, index_composition, &option);
+        auto gpu_index = faiss::gpu::index_cpu_to_gpu(res->faiss_res.get(), gpu_id, &index_composition, &option);
         delete gpu_index;
 
         auto q = std::make_shared<FaissIVFQuantizer>();
 
-        auto& q_ptr = index_composition->quantizer;
+        auto q_ptr = index_composition.quantizer;
         q->size = q_ptr->d * q_ptr->getNumVecs() * sizeof(float);
         q->quantizer = q_ptr;
         q->gpu_id = gpu_id;

core/src/index/unittest/test_gpuresource.cpp

@@ -42,14 +42,14 @@ class GPURESTEST : public DataGen, public TestGpuIndexBase {
 
         k = K;
         elems = nq * k;
-        ids = (int64_t*)malloc(sizeof(int64_t) * elems);
-        dis = (float*)malloc(sizeof(float) * elems);
+        ids = new int64_t[elems];
+        dis = new float[elems];
     }
 
     void
     TearDown() override {
-        delete ids;
-        delete dis;
+        delete[] ids;
+        delete[] dis;
         TestGpuIndexBase::TearDown();
     }