Intel Corporation

Srinivas Pandruvada

Ashok Raj

Jacob Jun Pan

Efraim Rotem

Eliezer Weissmann

Russell Jerome Fenger

Sridhar Muthrasanallur

Asit K Mallick

Nadav Bonen

Hisham Abu Salah

Deepak Samuel Kirubakaran

Arjan Van De Ven

Vishwanath Somayaji

Prashant Kodali

Arthur Leonard Brown

Rafal J Wysocki

Jun Pan

Apparatus, method, and machine-readable medium to provide performance state matching between source and target processors based on inter-processor interrupts. An exemplary apparatus includes a target processor to execute a receiving task at a first performance level and a source processor to execute a sending task at a second performance level higher than the first performance level. The sending task is to store interrupt routing data indicating a pairing between the sending task and the receiving task into a memory location and that the sending task is to dispatch work to be processed by the receiving task. The apparatus further includes a performance management unit to detect the pairing between the sending task and the receiving task based on the interrupt routing data and responsively adjust the performance level of the target processor from the first performance level to the second performance level based, at least in part, on the pairing.

Apparatus and method for performance state matching between source and target processors based on interprocessor interrupts

Logical memory is divided into two regions. Data in the first region is always retained. The first region of memory is designated online (or powered on) and is not offlined during standby or low power mode. The second region is the rest of the memory which can be potentially placed in non-self-refresh mode during standby by offlining the memory region. Content in the second region can be moved to the first region or can be flushed to another memory managed by the operating system. When the first region does not have enough space to accommodate data from the second region, the operating system can increase the logical size of the first region. Retaining the content of the first region by putting that region in self-refresh and saving power in the second region by not putting it in self-refresh is performed by an improved Partial Array Self Refresh scheme.

Memory management to improve power performance

In an embodiment, a processor includes multiple processing engines and a power control unit. The power control unit is to: maintain a first utilization metric for a first processing engine; detect a thread transfer from a first processing engine to a second processing engine; and generate, using the first utilization metric for the first processing engine, a second utilization metric for a second processing engine. Other embodiments are described and claimed.

Utilization metrics for processing engines

In an embodiment, a processor for performance state adjustment includes a plurality of processing engines (PEs), a power control unit, and an input/output memory management unit (IOMMU). The IOMMU is to determine a destination PE for a user interrupt based on mapping data of the IOMMU, and to send a notification of the user interrupt to the power control unit. The notification indicates the destination PE for the user interrupt. The power control unit is to adjust a performance state of the destination PE in response to the notification of the user interrupt. Other embodiments are described and claimed.

Management of processor performance based on user interrupts

Growing community of inventors

Beaverton, OR, United States of America

Srinivas Pandruvada