![]() ![]() ![]() Although it would be interesting to measure the performance when the whole system running directly on the NVMM, we cannot boot the operating system without any DRAM. The kernel considers Optane DC as slower memory and DRAM as faster memory, and puts them in two separate NUMA nodes. Each socket can access 192 GB DRAM and around 1.5 TB Optane DC. (In report they say "To build this configuration, we configured the Optane DC PMM into App Direct mode and let the Linux kernel consider Optane DC to be DRAM. Only the same memory medium is installed on the NUMA node of a CPU, so that the memory can be clearly allocated to the designated memory medium. Although the details of implementation are missing in their work, we guess they partitioned Optane DC memory and traditional DRAM installation. We follow the way in "Basic Performance Measurements of the Intel Optane DC Persistent Memory Module" and choose to use numa with its numactl interface (numctl -membind) for memory allocation in AD mode testing. The second is that we do not want to test the memory allocation interface based on the DAX implementation (also without persistent). ![]() The first is that our application is not easy to transplant to the interface provided by Intel. We hope to test our application in AD (app direct) mode without using Intel Persistent Memory Development Kit ( ). ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |