Advanced Driver Settings for Intel® Ethernet 10 Gigabit Server Adapters

Documentation

Install & Setup

000005783

01/28/2020

The settings in the Advanced tab customize how the adapter handles QoS packet tagging, Jumbo Frames, Offloading, and other capabilities.

This page does not apply to adapters that have PRO/10GbE in their name. For those adapters, see Advanced Settings for Intel® PRO/10GbE Server Adapters.

Note
  • Some features might not be available, depending on the operating system and the specific adapters installed.
  • Advanced tab layout and feature availability are different when Intel® PROSet for Windows Device Manager* is installed.
  • Onscreen help is available when Intel® PROSet for Windows Device Manager is installed.

Click or the topic for details:

Flow control

Flow control enables adapters to generate or respond to flow control frames, which help regulate network traffic. You can configure flow control under Performance Options properties when Intel® PROSet for Windows Device Manager is installed.

A link partner can become overloaded if incoming frames arrive faster than the device can process them. Frames are discarded until the overload condition passes. The flow control mechanism overcomes this problem and eliminates the risk of lost frames.

If an overload situation occurs, the device generates a flow control frame. It forces the transmitting link partner to stop transmitting before trying to retransmit.

Note For adapters to benefit from this feature, link partners must support flow control frames.
Default RX and TX Enabled
Range Disabled: No flow control.
RX and TX Enabled: The adapter generates and responds to flow control frames.
RX Enabled: The adapter pauses transmission when it receives a flow control frame from a link partner.
TX Enabled: The adapter generates a flow control frame when its receive queue reaches a predefined limit.
 
Header data split

Header data split enables the adapter to split the header information from an incoming packet so the CPU does not have to process the whole packet. This splitting action reduces CPU utilization.

OS Windows Server 2008*
Default Enabled
Range Disabled, Enabled
 
Interrupt moderation

Interrupt moderation allows the adapter to moderate interrupts.

When a packet arrives, the adapter generates an interrupt that allows the driver to handle the packet. At greater link speeds, it creates more interrupts, and CPU utilization also increases. It results in poor system performance. When you enable interrupt moderation, the interrupt rate is lower, and the result is better system performance.

OS Windows Server 2008
Default Enabled
Range Disabled, Enabled
 
Interrupt moderation rate

Interrupt moderation rate sets the rate at which the controller moderates or delays the generation of interrupts. Moderation helps optimize network throughput and CPU utilization. The default setting (Adaptive) adjusts the interrupt rates dynamically depending on traffic type and network usage. Choosing a different setting can improve network and system performance in certain configurations. You can configure interrupt moderation rate under Performance Options properties when Intel® PROSet for Windows Device Manager is installed.

Without interrupt moderation, CPU utilization increases at higher data rates because the system must handle a larger number of interrupts. Interrupt moderation causes the network driver to accumulate interrupts and send one interrupt rather than a series of interrupts. At higher data rates, a high interrupt moderation setting may improve system performance. At low data rates, a lower interrupt moderation setting is preferred as delayed interrupts cause latency.

Default Adaptive
Range Adaptive, High, Low, Medium, Off
 
IPv4 checksum offload

IPv4 checksum offload enables the adapter to verify the TCP/IP checksum on received packets (RX) and compute checksum on transmitted packets (TX). You can configure IPv4 checksum offload under TCP/IP Offloading Options properties when Intel® PROSet for Windows Device Manager is installed. Enabling this feature can improve TCP/IP performance and reduce CPU utilization. With offloading disabled, the operating system calculates and verifies the TCP/IP checksum.

Default RX and TX Enabled
Range Disabled, RX and TX Enabled, RX Enabled, TX Enabled
 
Jumbo packet

Enables jumbo packet capability for TCP/IP packets. When large packets make up most traffic and more latency can be tolerated, jumbo packets can reduce CPU utilization and improve wire efficiency. The standard Ethernet frame size is 1514 bytes, while Jumbo frames can contain 4088 or 9014 bytes. Available settings vary depending on the specific adapter.

Enable jumbo packets only if devices across the network support them and are configured to use the same frame size. When setting up jumbo frames on other network devices, be aware that network devices calculate jumbo frame sizes differently. Some devices include the header information in the frame size while others do not.

Note Intel adapters do not include header information in the frame size.
Default Disabled
Range Disabled (1514), 4088, and 9014 bytes
(Set the switch 4 bytes higher for CRC, plus 4 bytes if using VLANs.or QoS packet tagging)
 
Large receive offload (IPv4)

Large receive offload (IPv4) sets the adapter to offload the task of segmenting TCP messages into valid Ethernet frames. You can configure large receive offload under TCP/IP Offloading Options properties when Intel® PROSet for Windows Device Manager is installed.

Because the adapter hardware can complete data segmentation much faster than operating system software, this feature can improve transmission performance. The adapter also uses fewer CPU resources.

OS Windows Server 2008
Default Enabled
Range Disabled, Enabled
 
Large send offload (IPv4)

Large send offload (IPv4) and large send offload (IPv6) enable the adapter to offload the task of segmenting TCP messages into valid Ethernet frames.

Because the adapter hardware can complete data segmentation much faster than operating system software, this feature can improve transmission performance. The adapter also uses fewer CPU resources.

OS Windows Server 2008
Default Enabled
Range Disabled, Enabled
 
Locally administered address

Locally administered address overrides the initial MAC address with a user-assigned MAC address. To enter a new network address, type a 12-digit hexadecimal number in this box.

Default None
Range 0000 0000 0001 - FFFF FFFF FFFD
Exceptions:
  • Do not use a multicast address (Least Significant Bit of the high byte = 1).
  • Do not use all zeros or all F's.
Note Present Setting: Reverts to original network address of the adapter.
Note

In a team, Intel® PROSet uses either:

  • The team LAA if the team has an LAA configured
  • The permanent MAC address of the main adapter, if the team does not have an LAA configured

Intel PROSet does not use the adapter LAA if:

  • The adapter is the main adapter in a team
  • The team has an LAA
 
Log link state event

Log link state event enables the logging of the following link state changes to the system event log:

  • LINK_UP_CHANGE
    Indicates that the adapter has established link.
  • LINK_DOWN_CHANGE
    Indicates that the adapter has lost link.
  • LINK_DUPLEX_MISMATCH
    Indicates a mismatch in duplex between the adapter and the link partner

To investigate connection problems, click the Link Speed tab and run diagnostics. The Link Speed tab and diagnostics are only available when Intel® PROSet for Windows Device Manager is installed.

Low latency interrupts

The low latency interrupts option enables adapters to bypass interrupt moderation and generate an interrupt when certain TCP packets arrive. The system can then handle the packet faster. Certain applications have faster access to network data because of the reduced data latency.

Intel PROSet for Windows Device Manager must be installed to enable configuration of low latency interrupts. You can configure low latency interrupt properties in the Performance Options properties. To use low latency interrupts, select the Use Low Latency Interrupts check box and choose one of these options:

  • Use for packets with TCP PSH flag: Any incoming packet with the TCP PSH flag triggers an immediate interrupt. The sending device sets the PSH flag.
  • Use for these TCP ports: Every packet received on the specified ports triggers an immediate interrupt. You can specify up to eight ports.
Note If you enable low latency interrupts, system CPU utilization can increase.
 
Offload TCP segmentation

Offload TCP segmentation enables the adapter to offload the task of segmenting TCP messages into valid Ethernet frames. You can configure offload TCP segmentation under TCP/IP Offloading Options properties when Intel® PROSet for Windows Device Manager is installed.

Because the adapter hardware completes data segmentation much faster than operating system software, this feature can improve transmission performance. The adapter also uses fewer CPU resources.

Default On
Range Off, On
 
Priority and VLAN

Priority and VLAN enables sending and receiving of IEEE 802.3ac tagged frames, including:

  • 802.1p QoS (Quality of Service) tags for priority-tagged packets
  • 802.1Q tags for VLANs

When this feature is enabled, tagged packets use the queue settings defined by the operating system Priority Level Definition. Priority & VLAN is automatically enabled when you set up a VLAN on the VLAN tab. You cannot disable tagging because tagging is required for VLANs.

Note Intel® PROSet for Windows Device Manager and Advanced Networking Services (ANS) must be installed to configure VLANs on Intel® network adapters.
 
Default Priority & VLAN Enabled
Range Priority & VLAN Disabled, Priority & VLAN Enabled, Priority Enabled, VLAN Enabled
 
Receive buffers

Sets the number of receive buffers used by the adapter when copying data to memory. Increasing this value can enhance receive performance, but consumes system memory. You can configure receive buffers under Performance Options properties when Intel® PROSet for Windows Device Manager is installed.

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.

Default 512
Range 128–2048 in increments of 64
  
Receive side scaling

Receive side scaling (RSS) enables processing for a TCP connection across multiple processors or processor cores. If your adapter does not support RSS, or if your operating system does not support it, the RSS setting does not display.

Default Enabled
Range Disabled, Enabled
 
Receive side scaling queues

Receive side scaling queues allocates queue space to buffer transactions between the network adapter and processors. Configuration of queues is not supported on all adapters that support RSS. The number of queues supported varies depending on the adapter and operating system. Only settings allowed for your adapter and operating system are displayed.

Default 2 Queues
Range 1, 2, or 4 Queues

See the User Guide for more information on receive side scaling.

TCP checksum offload (IPv4)

TCP checksum offload (IPv4) and TCP Checksum Offload (IPv6) enable the adapter to compute (TX) or verify (RX) the TCP checksum of packets. You can configure TCP checksum offload under TCP/IP Offloading Options properties when Intel® PROSet for Windows Device Manager is installed. This feature can improve performance and reduce CPU utilization. With Offloading enabled, the adapter computes or verifies the checksum for the operating system.

OS All for IPv4, Windows Server 2008 for IPv6
Default RX and TX Enabled
Range Disabled, RX and TX Enabled, RX Enabled, TX Enabled
 
Transmit buffers

Sets the number of Transmit Buffers used by the adapter when copying data to memory. Increasing this value can enhance transmission performance, but also consumes system memory. You can configure transmit buffers under Performance Options properties when Intel® PROSet for Windows Device Manager is installed.

You can choose to increase the number of transmit buffers if you notice a significant decrease in the performance of transmitted traffic. If transmission performance is not an issue, use the default setting.

Default 512
Range 128–16384 in increments of 64
UDP checksum offload (IPv4)

UDP checksum offload (IPv4) and UDP checksum offload (IPv6) enable the adapter to compute (TX) or verify (RX) the UDP checksum of packets. You can configure UDP checksum offload under TCP/IP Offloading Options properties when Intel® PROSet for Windows Device Manager is installed. This feature can improve performance and reduce CPU utilization. With offloading enabled, the adapter computes or verifies the checksum for the operating system.

Default RX and TX Enabled
Range Disabled, RX and TX Enabled, RX Enabled, TX Enabled
Related topic
Jumbo Frames and Jumbo Packets Notes