Modern DRAMs have multiple banks to serve multiple memory requests in parallel. However, when two requests go to the same bank, they have to be served serially, exacerbating the h...
Yoongu Kim, Vivek Seshadri, Donghyuk Lee, Jamie Li...
When multiple processor (CPU) cores and a GPU integrated together on the same chip share the off-chip main memory, requests from the GPU can heavily interfere with requests from t...
Rachata Ausavarungnirun, Kevin Kai-Wei Chang, Lava...
Most modern cores perform a highly-associative translation look aside buffer (TLB) lookup on every memory access. These designs often hide the TLB lookup latency by overlapping it...
Memory system reliability is a serious and growing concern in modern servers. Existing chipkill-level memory protection mechanisms suffer from several drawbacks. They activate a l...
Aniruddha N. Udipi, Naveen Muralimanohar, Rajeev B...
Hybrid processors are HW/SW co-designed processors that leverage blocked-execution, the execution of regions of instructions as atomic blocks, to facilitate aggressive speculative...