Intel XL710-Q2 User Manual page 42

Ethernet adapters and devices
Hide thumbs Also See for XL710-Q2:
Table of Contents

Advertisement

Receive Buffers
Defines the number of Receive Buffers, which are data segments. They are allocated in the host memory and
used to store the received packets. Each received packet requires at least one Receive Buffer, and each
buffer uses 2KB of memory.
You might choose to increase the number of Receive Buffers if you notice a significant decrease in the
performance of received traffic. If receive performance is not an issue, use the default setting appropriate to
the adapter.
Default
Range
Recommended Value
Receive Side Scaling
When Receive Side Scaling (RSS) is enabled, all of the receive data processing for a particular TCP
connection is shared across multiple processors or processor cores. Without RSS all of the processing is
performed by a single processor, resulting in less efficient system cache utilization. RSS can be enabled for a
LAN or for FCoE. In the first case, it is called "LAN RSS". In the second, it is called "FCoE RSS".
LAN RSS
LAN RSS applies to a particular TCP connection.
NOTE: This setting has no effect if your system has only one processing unit.
LAN RSS Configuration
RSS is enabled on the Advanced tab of the adapter property sheet. If your adapter does not support RSS, or
if the SNP or SP2 is not installed, the RSS setting will not be displayed. If RSS is supported in your system
environment, the following will be displayed:
Port NUMA Node. This is the NUMA node number of a device.
l
Starting RSS CPU. This setting allows you to set the p referred starting RSS processor. Change this
l
setting if the current processor is dedicated to other processes. The setting range is from 0 to the num-
ber of logical CPUs - 1. In Server 2008 R2, RSS will only use CPUs in group 0 (CPUs 0 through 63).
Max number of RSS CPU. This setting allows you to set the maximum number of CPUs assigned to
l
an adapter and is primarily used in a Hyper-V environment. By decreasing this setting in a Hyper-V
environment, the total number of interrupts is reduced which lowers CPU utilization. The default is 8 for
Gigabit adapters and 16 for 10 Gigabit adapters.
Preferred NUMA Node. This setting allows you to choose the preferred NUMA (Non-Uniform
l
Memory Access) node to be used for memory allocations made by the network adapter. In addition the
system will attempt to use the CPUs from the preferred NUMA node first for the purposes of RSS. On
NUMA platforms, memory access latency is dependent on the memory location. Allocation of memory
from the closest node helps improve performance. The Windows Task Manager shows the NUMA
512, for the 10 Gigabit Server Adapters.
256, for all other adapters depending on the features selected.
128-4096, in intervals of 64, for the 10 Gigabit Server Adapters.
80-2048, in intervals of 8, for all other adapters.
Teamed adapter: 256
Using IPSec and/or multiple features: 352

Hide quick links:

Advertisement

Table of Contents
loading

Table of Contents