SPEC CPU2006 Platform Settings for Intel-based systems

Firmware / BIOS / Microcode Settings

Hardware Prefetcher:

This BIOS option allows the enabling/disabling of a processor mechanism to prefetch data into the cache according to a pattern-recognition algorithm. This default setting is "Enabled".

In some cases, setting this option to Disabled may improve performance. Users should only disable this option after performing application benchmarking to verify improved performance in their environment.

Adjacent Cache Line Prefetch:

This BIOS option allows the enabling/disabling of a processor mechanism to fetch the adjacent cache line within a 128-byte sector that contains the data needed due to a cache line miss. This default setting is "Enabled".

In some cases, setting this option to Disabled may improve performance. Users should only disable this option after performing application benchmarking to verify improved performance in their environment.

Hyper-Threading:

Disabling Intel's Hyper-Threading Technology reduces the number of threads per core to 1. The default is Enabled; in this case each core provides additional resources for executing up to 2 threads in parallel.

Power Management Policy:

This option selects the power managment policy. If this option is set to "Custom", the processor power management options are selectable independently. This default setting is "Custom".

EIST:

Enabling this option allows the system to dynamically adjust processorvoltage and core frequency. This technology can result in decreased power consumption and decreased heat production. This default setting is "Enabled".

Energy Performance:

This option is displayed when the EIST is set to "Enabled". This option can only be selectable when the Power Management Policy is set to "Custom". This item decides the energy efficiency policy which is the energy per performance rate. If the highest performance is preferred, this option should be set to "Performance".

Patrol Scrub:

Patrol Scrub is a mechanism for memory controller to periodically read all memory. Corrected read data is written back to memory when a correctable error is detected. This default setting is "Enabled".

Demand Scrub:

Demand Scrub is a mechanism for memory controller to correct a correctable error in memory. Corrected read data is sent to the requestor and written back to memory. This default setting is "Enabled".

Cluster on Die:

Enable/Disable Cluster on Die mode. This default setting is "Disabled".

Cluster on Die mode logically splits a socket into 2 NUMA domains that are exposed to the OS with half the amount of cores and LLC assigned to each NUMA domain in a socket. This mode utilizes an on-die directory cache and in memory directory bits to determine whether a snoop needs to be sent. Use this mode for highly NUMA optimized workloads to get the lowest local memory latency and highest local memory bandwidth for NUMA workloads.

Snoop Mode:

This option can only be selectable when the Cluster on Die is set to "Disabled". This default setting is "Early Snoop". There are 3 snoop mode options for how to maintain cache coherency across the Intel QPI fabric, each with varying memory latency and bandwidth characteristics depending on how the snoop traffic is generated.

In Home Snoop and Early Snoop modes, snoops are always sent, they just originate from different places: the caching agent (earlier) in Early Snoop mode and the home agent (later) in Home Snoop mode.

In Home Snoop with Directory mode, the home agent does a speculative home snoop broadcast under very lightly loaded conditions, even before the directory information has been collected and checked.

QPI Frequency Limit:

This option selects the upper limit of the QPI link speed. This default setting is "Auto", which configures the highest supported QPI link speed automatically.