CPU2017 Flag Description - Platform settings for New H3C systems

Operating System Tuning Parameters

cpupower

Manage the frequency scaling of your CPU.We use "cpupower frequency-set -g performance" to set to performance mode.

• performance: Run the CPU at the maximum frequency.
• powersave(default): Run the CPU at the minimum frequency.

tuned-adm

Tuning profile delivery mechanism for Linux.We use "tuned-adm profile throughput-performance" to set to throughput-performance profile.

• accelerator-performance: Throughput-performance based tuning with disabled higher latency STOP states.
• balanced: General non-specialized tuned profile.
• desktop: Optimize for the desktop use-case.
• hpc-compute: Optimize for HPC compute workloads.
• intel-sst: Configure for Intel Speed Select Base Frequency.
• latency-performance: Optimize for deterministic performance at the cost of increased power consumption.
• network-latency: Optimize for deterministic performance at the cost of increased power consumption, focused on low latency network performance.
• network-throughput: Optimize for streaming network throughput, generally only necessary on older CPUs or 40G+ networks.
• optimize-serial-console: Optimize for serial console use.
• powersave: Optimize for low power consumption.
• throughput-performance(default): Broadly applicable tuning that provides excellent performance across a variety of common server workloads.
• virtual-guest: Optimize for running inside a virtual guest.
• virtual-host: Optimize for running KVM guests.

Firmware / BIOS / Microcode Settings

Enable LP [Global]

Select whether all logical processors are enabled.

• ALL LPs(Default Setting):Enable all logical processors.
• Single LP:Enable only a single logical processor.

Package C State

Package C state is one of energy saving options of the processor, which not only allow the individual cores of a processor, but the entire processor chip to be put into a type of sleep state. As a result, power consumption is even further reduced. But the waking up time that is required to change from the lower package C state to the active state is even longer in comparison with the CPU core C state. The higher the package C state, the lower the power usage of that idle package state. If the "C0/C1" is set, the processor chip always remains active, and it can improve the performance of latency sensitive workloads.

• C0/C1 state:When the CPU is idle, the system slightly reduces power consumption. This option requires less power than C0 and allows the server to return quickly to high-performance mode.
• C2 state:When the CPU is idle, the system reduces power consumption more than with the C1 option. This option requires less power than C1 or C0, but the server takes slightly longer to return to high-performance mode.
• C6(non Retention) state:When the CPU is idle, the system reduces power consumption more than with the C3 option. This option saves more power than C0, C1, or C3, but the system may experience performance problems until the server returns to full power.
• No Limit:all C-States supported by the processor are available.
• Auto(Default Setting):Initialize to deepest available package C state limit.

SNC

Sub‑NUMA Clustering (SNC) is a feature that partitions the Last Level Cache (LLC) into disjoint clusters based on address ranges, with each cluster bound to a subset of the memory controllers in the system. This can improve the average latency to the LLC for NUMA‑aware workloads. The number of clusters is automatically determined by the BIOS, based on the detected CPU SKU, core count, cache hierarchy, and IMC (Integrated Memory Controller) layout.

For example:

SNC2 is supported on Granite Rapids‑SP XCC SKUs, but not supported on Granite Rapids‑SP HCC and LCC SKUs.
SNC3 is supported on Granite Rapids‑AP UCC SKUs.

• Auto (Default Setting): System decides based on Si Compatibility. For Granite Rapids CPUs, Auto is treated as Enable.
• Disabled: supports 1-cluster and 2-way IMC interleave, the LLC is treated as one cluster.
• Enable: Utilizes LLC capacity efficiently and reduces latency due to core/IMC proximity. This may provide performance improvement on NUMA-aware operating systems.

Latency Optimized Mode

Prioritize low latency and consistent performance over energy efficiency or other optimizations.

• Disabled(Default Setting):Out-of-box mode for newer disaggregated SoC arch with significant power savings across the load line. Better performance per watt across the load line. No significant difference in performance compared to Latency Optimized Mode at 100% Load Level.
• Enabled: Uncore frequencies will run up to their maximum limits within the RAPL budget. This mode is not performance-per-watt optimized across the load line.

VMX

This BIOS option enables or disables additional virtualization functions of the CPU.If the server is not used for virtualization, this option should be set to "Disabled".This can result in energy savings. Default setting is "Enabled".

DCU Streamer Prefetcher

DCU Streamer Prefetcher is an L1 data cache prefetcher. When enabled, this parameter fetches the next cache line into the L1 data cache when multiple loads from the same cache line are executed in a certain time limit. Lightly threaded applications and some benchmarks can benefit from having the DCU streamer prefetcher enabled. Values for this BIOS option can be: Auto/Enable/Disable. Current default is Auto (Same as Enable).

Adjacent Cache Prefetch

Adjacent Cache Prefetch is an L2 cache prefetcher. When enabled, fetches both cache lines that make up a 128 byte cache line pair even if the requested data is only in the first cache line. Lightly threaded applications and some benchmarks can benefit from having the adjcent cache line prefetch enabled. Values for this BIOS option can be: Enable/Disable. Current default is Enable.

LLC Prefetch

Last Level Cache (LLC) Prefetch.This option configures the processor last level cache (LLC) prefetch feature as a result of the non-inclusive cache architecture. The LLC prefetcher exists on top of other prefetchers that can prefetch data into the core data cache unit (DCU) and mid-level cache (MLC).

• Disabled(Default Setting):not allowed to prefetch data into the core data cache unit and L2 cache.
• Enabled: Enabling LLC prefetch gives the core prefetcher the ability to prefetch data directly into the LLC

ACPI C6x Enumeration

Use this feature to configure C6 state or C6 P-state as ACPI C2 or ACPI C3 state.

• Disabled:Don't enumerate any C6S state in ACPI.
• C6 as ACPI C2:Enumerate C6 as ACPI C2 state. PkgC6 is not allowed.
• C6 as ACPI C3:Enumerate C6 as ACPI C3 state. PkgC6 is not allowed..
• C6-P as ACPI C2:Enumerate C6-P as ACPI C2 state. PkgC6 is allowed.
• C6-P as ACPI C3:Enumerate C6-P as ACPI C3 state. PkgC6 is allowed.
• Auto(Default Setting):Maps to C6 or C6-P as ACPI C2.

Virtual NUMA

Divide physical NUMA nodes into evenly sized virtual NUMA nodes in ACPI table.

• Disabled(Default Setting):Physical NUMA nodes will not be devided into virtual NUMA nodes.
• Enabled:Physical NUMA nodes will be devided into evenly sized virtual NUMA nodes.

Fan mode:

To Select system Fan policy on BMC WebUI.

• Balanced mode(Default Setting): : makes fan speed self-adjust actively according to the reading of the system temperature sensors.
• Silent mode : makes fan speed self-adjust actively according to the reading of the system temperature sensors.but fan speed is lower and quieter when the server is under low load.
• powerful mode: makes fan run at full speed to provide best cooling performance.
• Custom mode: allow users to customize fan speed settings.