Abstract | ||
---|---|---|
We compare five existing dynamic memory allocators optimized for GPUs and show their strengths and weaknesses. In the measurements, we use three generic evaluation tests proposed in the past and we add one with a real workload, where dynamic memory allocation is used in building the k-d tree data structure. Following the performance analysis we propose a new dynamic memory allocator and its variants that address the limitations of the existing dynamic memory allocators. The new dynamic memory allocator uses few resources and is targeted towards large and variably sized memory allocations on massively parallel hardware architectures. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1111/cgf.12666 | Computer Graphics Forum |
Keywords | Field | DocType |
dynamic memory allocation,many-core architecture,GPU,CUDA | Computer architecture,Uniform memory access,Shared memory,Physical address,Computer science,Parallel computing,Memory management,Flat memory model,Allocator,Distributed shared memory,CUDA Pinned memory | Journal |
Volume | Issue | ISSN |
34 | 8 | 0167-7055 |
Citations | PageRank | References |
3 | 0.43 | 12 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Marek Vinkler | 1 | 22 | 3.14 |
Vlastimil Havran | 2 | 353 | 34.71 |