ActivePointers: a case for software address translation on GPUs
… for running GPU programs. We implement ActivePointers, a software address translation layer
… for page faults and virtual address space management to GPU programs, and enables the …
… for page faults and virtual address space management to GPU programs, and enables the …
ActivePointers: A case for software address translation on GPUs
… for running GPU programs. We implement ActivePointers, a software address translation layer
… for page faults and virtual address space management to GPU programs, and enables the …
… for page faults and virtual address space management to GPU programs, and enables the …
Supporting address translation for accelerator-centric architectures
… ActivePointers [51] introduces a software address translation layer on GPUs that supports
page fault handling without CPU involvement. However, system abstractions for GPUs are …
page fault handling without CPU involvement. However, system abstractions for GPUs are …
Improving address translation in multi-gpus via sharing and spilling aware tlb design
… in multi-GPU systems receives little attention. In this paper, we conduct a … of the
address translation efficiency in both “single-application-multi-GPU” and “multi-application-multi-GPU” …
address translation efficiency in both “single-application-multi-GPU” and “multi-application-multi-GPU” …
IDYLL: Enhancing Page Translation in Multi-GPUs via Light Weight PTE Invalidations
… page table invalidation requests to GPUs that hold the corresponding address translation (ii)
… , we provide a detailed discussion of address translations in multi-GPU systems (Section 3) …
… , we provide a detailed discussion of address translations in multi-GPU systems (Section 3) …
Enhancing address translations in throughput processors via compression
… Particularly, we observe that GPU execution suffers from … mechanisms for address translations
to improve GPU TLB hit … cution of GPU applications to design efficient address translation …
to improve GPU TLB hit … cution of GPU applications to design efficient address translation …
[PDF][PDF] BaM: A case for enabling fine-grain high throughput GPU-orchestrated access to storage
… software overhead involved in the virtual address translation … from the GPU memory and
needs to be brought into GPU from … ActivePointers [41] added a memory-map like abstraction on …
needs to be brought into GPU from … ActivePointers [41] added a memory-map like abstraction on …
GPU-initiated on-demand high-throughput storage access in the BaM system architecture
… Next, we compare BaM and ActivePointers [57]. … ActivePointers, the file is pinned in the
Linux page cache in CPU memory, favoring ActivePointers as any misses in the ActivePointers …
Linux page cache in CPU memory, favoring ActivePointers as any misses in the ActivePointers …
Eleos: ExitLess OS services for SGX enclaves
… memory access due to the extra level of address translation. In our solution, we retrofit the
ActivePointers software address translation mechanism used in GPUs [25]. Spointers provide …
ActivePointers software address translation mechanism used in GPUs [25]. Spointers provide …
DRAGON: breaking GPU memory capacity limits with direct NVM access
… Silberstein, “Activepointers: a case for software address translation on gpus,” in Proceedings
of the 43rd International Symposium on Computer Architecture, 2016. [29] M. Silberstein, B. …
of the 43rd International Symposium on Computer Architecture, 2016. [29] M. Silberstein, B. …