NVIDIA Posts Linux Scheduler Patches To Further Boost SMT Performance On NVIDIA Vera

Hitting the Linux kernel mailing list today is a new kernel scheduler patch series to enable preferred SMT siblings for NVIDIA Vera with its Olympus cores. NVIDIA engineer Andrea Righi summed up the situation well in the patch series cover letter: "NVIDIA Olympus implements SMT with two symmetric processing elements (PEs). When only one PE is active, the core operates in single-thread mode and that PE can use the full core resources. When both PEs are active, the core operates in two-thread mode and the PEs share those resources. This behavior is common to SMT implementations, but Olympus is particularly sensitive to brief sibling activations because returning from two-thread mode to single-thread mode after a sibling becomes idle is not immediate. As described by commit 293f9611ae735 ("sched/fair: Prefer fully idle cores for NOHZ balancing"): "Briefly activating an otherwise idle sibling can reduce the performance available to the other sibling and this effect does not necessarily end once the activated sibling becomes idle: after the ILB finishes and its CPU enters WFI, full single-thread performance is restored only after the sibling has remained idle for a qualification interval (10 Ki cycles on the tested Vera system)." That change prevents the NOHZ idle load balancer from unnecessarily waking a sibling of a busy PE. However, ordinary task placement can still select either sibling of an idle core and repeated changes of the active PE can keep Olympus cores in two-thread mode despite little or no useful overlap between the siblings. This series makes PE0 the preferred sibling of an Olympus core using SD_ASYM_PACKING and teaches the fair scheduler's idle-selection paths to honor asymmetric SMT priority. The scheduler first selects an idle core according to its existing placement and capacity rules, then chooses the highest-priority available sibling within that core. The generic scheduler behavior is enabled only when an architecture supplies an SD_ASYM_PACKING SMT domain." What's important is the end result that Righi summed up as delivering nice gains for SMT on NVIDIA two socket Vera: "The series was tested on a two-node Vera system using an 88-thread single-precision GEMM on the 88 physical cores of NUMA node 0. With the workload allowed to choose either sibling of every core, observed throughput improved from approximately 9.4 TFLOP/s on the baseline kernel to approximately 10.1 TFLOP/s with this series applied. Repeated runs also became more predictable because the workload consistently settled on PE0 while PE1 remained quiet." The 293f9611ae735 change was also authored by Righi and merged for Linux 7.3 to prefer fully idle cores for NOHZ balancing. This change already had boosted the GEMM benchmark from 6.2 TFLOP/s to 9.2 TFLOP/s on Vera. The patches specific to NVIDIA Vera / Olympus cores are now out for review on the Linux kernel mailing list.