Platform Settings for Dell PowerEdge Servers

Operating System Tuning Parameters

kernel.randomize_va_space (ASLR)
This setting can be used to select the type of process address space randomization. Defaults differ based on whether the architecture supports ASLR, whether the kernel was built with the CONFIG_COMPAT_BRK option or not, or the kernel boot options used.
Possible settings: Disabling ASLR can make process execution more deterministic and runtimes more consistent. For more information see the randomize_va_space entry in the Linux sysctl documentation.

Transparent Hugepages (THP)
THP is an abstraction layer that automates most aspects of creating, managing, and using huge pages. It is designed to hide much of the complexity in using huge pages from system administrators and developers. Huge pages increase the memory page size from 4 kilobytes to 2 megabytes. This provides significant performance advantages on systems with highly contended resources and large memory workloads. If memory utilization is too high or memory is badly fragmented which prevents hugepages being allocated, the kernel will assign smaller 4k pages instead. Most recent Linux OS releases have THP enabled by default.
THP usage is controlled by the sysfs setting /sys/kernel/mm/transparent_hugepage/enabled. Possible values: THP creation is controlled by the sysfs setting /sys/kernel/mm/transparent_hugepage/defrag. Possible values: An application that "always" requests THP often can benefit from waiting for an allocation until those huge pages can be assembled.
For more information see the Linux transparent hugepage documentation.

Firmware / BIOS / Microcode Settings

C States:

C States allow the processor to enter lower power states when idle. When set to Enabled (OS controlled) or when set to Autonomous (if Hardware controlled is supported), the processor can operate in all available Power States to save power, but my increase memory latency and frequency jitter.

L3 cache as NUMA Domain:

This field specifies that each CCX within the processor will be declared as a NUMA Domain.

Dram Refresh Delay:

By enabling CPU memory controller to delay running the REFRESH commands, you can improve the performance for some workloads. By minimizing the delay time, it is ensured that the memory controller runs the REFRESH command at regular intervals. For Intel-based servers, this setting only affects systems configured with DIMMs which use 8Gb density DRAMs.

L1 Stream HW prefetcher, L2 Stream HW prefetcher:

Most workloads will benefit from the L1 and L2 Stream Hardware prefetchers gathering data and keeping the core pipeline busy. There are however some workloads that are very random in nature and will actually obtain better overall performance by disabling one or both of the prefetchers.

UPI Prefetch:

Enables you to get the memory read started early on DDR bus. The Ultra Path Interconnect (UPI) Rx path will spawn the speculative memory read to Integrated Memory Controller (iMC) directly.

Dynamic Link Width Management (DLWM):

DLWM reduces the XGMI link width between sockets from x16 to x8 (default), when no traffic is detected on the link. This feature is optimized to trade power between core and high IO/memory bandwidth workloads.
Forced = Force link width to x16, x8, or x2.

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). In some cases, setting this option to disabled can improve performance. Typically, setting this option to enable provides better performance. Disabled: Disables the LLC prefetcher. Enabled: Gives the core prefetcher the ability to prefetch data directly to the LLC.

Dead Line LLC Alloc:

In the Skylake cache scheme, mid-level cache (MLC) evictions are filled into the last level cache (LLC). If a line is evicted from the MLC to the LLC, the Skylake core can flag the evicted MLC lines as "dead". This means that the lines are not likely to be read again. This option allows dead lines to be dropped and never fill the LLC if the option is disabled. Disabled: Disabling this option can save space in the LLC by never filling dead lines into the LLC. Enabled: Opportunistically fill dead lines in LLC, if space is available.

Directory AtoS:

AtoS optimization reduces remote read latencies for repeat read accesses without intervening writes.

Algorithm performance Boost Disable (ApbDis):

When enabled a specific hard-fused Data Fabric (SOC) p-state is forced for optimizing workloads sensitive to latency or throughput. When disabled P-states will be automatically managed by the Application Power Management, allowing the processor to provide maximum performance while remaining within a specified power-delivery and thermal envelope.

Fan Speed:

Selecting this option allows additional cooling to the server. In case hardware is added (example, new PCIe cards), it may require additional cooling. A fan speed offset causes fan speeds to increase (by the offset % value) over baseline fan speeds calculated by the Thermal Control algorithm. Maximum — Drives fan speeds to full speed.

Determinism Slider:

It controls whether BIOS will enable determinism to control performance. Performance: BIOS will enable 100% deterministic performance control. Power: BIOS will not enable deterministic performance control.

CPU Power Management set to Maximum Performance:

Allows selection of CPU power management methodology. Maximum Performance is typically selected for performance-centric workloads where it is acceptable to consume additional power to achieve the highest possible performance for the computing environment. This mode drives processor frequency to the maximum across all cores (although idled cores can still be frequency reduced by C-state enforcement through BIOS or OS mechanisms if enabled). This mode also offers the lowest latency of the CPU Power Management Mode options, so is always preferred for latency-sensitive environments. OS DBPM is another performance-per-watt option that relies on the operating system to dynamically control individual cores in order to save power.

Memory Frequency set to Maximum Performance:

Governs the BIOS memory frequency. The variables that govern maximum memory frequency include the maximum rated frequency of the DIMMs, the DIMMs per channel population, the processor choice, and this BIOS option. Additional power savings can be achieved by reducing the memory frequency, at the expense of reduced performance. Read-only unless System Profile is set to Custom.

Efficiency Optimized Mode Disabled:

This field enables/disabled Efficiency Optimized Mode. Efficiency Optimized Mode maximizes Performance-per-Watt by opportunistically reducing frequency/power.

NUMA Nodes Per Socket:

NUMA nodes per socket (NPS) field allows you to configure the memory NUMA domains per socket. The configuration can consist of one whole domain (NPS1), two domains (NPS2), or four domains (NPS4). In the case of a two-socket platform, an additional NPS profile is available to have whole system memory to be mapped as single NUMA domain (NPS0).

CCX as NUMA Domain:

In addition to selecting the number of NUMA domains via NPS option, the processor allows for making memory per CCX as NUMA domain. In the processor each CCD has a maximum of two CCXs with each CCX having a shared last-level cache (LLC, or L3 cache) for all cores. The CCX as NUMA domain option allows for each LLC to be configured as a NUMA domain so that for certain workloads pinning execution to a single NUMA domain can be done.

Adaptive Double DRAM Device Correction (ADDDC):

When Adaptive Double DRAM Device Correction (ADDDC) is enabled, failing DRAM’s are dynamically mapped out. When set to enabled, it can have some impact to system performance under certain workloads. This feature is applicable for x4 DIMMs only.

DCU Streamer Prefetcher:

Enables or disables Data Cache Unit (DCU) Streamer Prefetcher. This setting can affect performance, depending on the application running on the server. DCU streamer prefetchers detect multiple reads to a single cache line in a certain period of time and choose to load the following cache line to the L1 data caches. Recommended for High Performance Computing applications.

DCU IP Prefetcher:

Enables or disables Data Cache Unit (DCU) IP Prefetcher. DCU IP Prefetcher looks for sequential load history to determine whether to prefetch the data to the L1 caches.

Memory Frequency:

Governs the BIOS memory frequency. The variables that govern maximum memory frequency include the maximum rated frequency of the DIMMs, the DIMMs per channel population, the processor choice, and this BIOS option. Additional power savings can be achieved by reducing the memory frequency, at the expense of reduced performance.

Turbo Boost:

Governs the Boost Technology. This feature allows the processor cores to be automatically clocked up in frequency beyond the advertised processor speed. The amount of increased frequency (or 'turbo upside') one can expect from an EPYC processor depends on the fewer cores being exercised with work the higher the potential turbo upside. The potential drawback for Boost are mainly centered on increased power consumption and possible frequency jitter that can affect a small minority of latency-sensitive environments.

C1E:

When set to Enabled, the processor is allowed to switch to minimum performance state when idle.

CPU Interconnect Bus Link Power Management:

When Enabled, CPU interconnect bus link power management can reduce overall system power a bit while slightly reducing system performance.

CPU Performance:

Maximum Performance is typically selected for performance-centric workloads where it is acceptable to consume additional power to achieve the highest possible performance for the computing environment. This mode drives processor frequency to the maximum across all cores (although idled cores can still be frequency reduced by C-state enforcement through BIOS or OS mechanisms if enabled). This mode also offers the lowest latency of the CPU Power Management Mode options, so is always preferred.

Energy Efficient Policy:

The CPU uses the setting to manipulate the internal behavior of the processor and determines whether to target higher performance or better power savings. The possible settings are: Performance, Balanced Performance, Balanced Energy, Energy Efficient.

Energy Efficient Turbo:

Permits Energy Efficient Turbo to be Enabled or Disabled.
Energy Efficient Turbo (EET) is a mode of operation where a processor's core frequency is adjusted within the turbo range based on workload.

Logical Processor:

Each processor core supports up to two logical processors. When set to Enabled, the BIOS reports all logical processors. When set to Disabled, the BIOS only reports one logical processor per core. Generally, higher processor count results in increased performance for most multi-threaded workloads and the recommendation is to keep this enabled. However, there are some floating point/scientific workloads, including HPC workloads, where disabling this feature may result in higher performance.

Memory Patrol Scrub:

Patrol Scrubbing searches the memory for errors and repairs correctable errors to prevent the accumulation of memory errors. When set to Disabled, no patrol scrubbing will occur. When set to Standard Mode, the entire memory array will be scrubbed once in a 24 hour period. When set to Extended Mode, the entire memory array will be scrubbed more frequently to further increase system reliability.

Memory Refresh Rate:

The memory controller will periodically refresh the data in memory. The frequency at which memory is normally refreshed is referred to as 1X refresh rate. When memory modules are operating at a higher than normal temperature or to further increase system reliability, the refresh rate can be set to 2X, but may have a negative impact on memory subsystem performance under some circumstances.

PCI ASPM L1 Link Power Management:

When Enabled, PCIe Advanced State Power Management (ASPM) can reduce overall system power a bit while slightly reducing system performance.

NOTE: Some devices may not perform properly (they may hang or cause the system to hang) when ASPM is enabled, for this reason L1 will only be enabled for validated qualified cards.

System Profile:

When set to Custom, you can change setting of each option. Under Custom mode when C States is enabled, Monitor/Mwait should also be Enabled.

Monitor/Mwait:

Specifies whether Monitor/Mwait instructions are enabled. Monitor/Mwait is only active when C States is set to Disabled.

Sub NUMA Cluster:

When Enabled, Sub NUMA Clustering (SNC) is a feature for breaking up the LLC into disjoint clusters based on address range, with each cluster bound to a subset of the memory controllers in the system. It improves average latency to the LLC.

Uncore Frequency:

Selects the Processor Uncore Frequency.
Dynamic mode allows processor to optimize power resources across the cores and uncore during runtime. The optimization of the uncore frequency to either save power or optimize performance is influenced by the setting of the Energy Efficient Policy.

Virtualization Technology:

When set to Enabled, the BIOS will enable processor Virtualization features and provide the virtualization support to the Operating System (OS) through the DMAR table. In general, only virtualized environments such as VMware(r) ESX (tm), Microsoft Hyper-V(r) , Red Hat(r) KVM, and other virtualized operating systems will take advantage of these features. Disabling this feature is not known to significantly alter the performance or power characteristics of the system, so leaving this option Enabled is advised for most cases.

nohz_full:

This kernel option sets adaptive tick mode (NOHZ_FULL) to specified processors. Since the number of interrupts is reduced to ones per second, latency-sensitive applications can take advantage of it.

Memory Interleaving:

When Enabled, memory interleaving is supported if a symmetric memory configuration is installed. When set to Disabled, the system supports Non-Uniform Memory Access (NUMA) (asymmetric) memory configurations. Channel interleaving is available with all configurations and is the intra-die memory interleave option. With channel interleaving the memory behind each UMC will be interleaved and seen as 1 NUMA domain per die.