This is a draft document that was built and uploaded automatically. It may document beta software and be incomplete or even incorrect. Use this document at your own risk.

Jump to contentJump to page navigation: previous page [access key p]/next page [access key n]
SUSE Telco Cloud Documentation|Telco features configuration|vRAN Acceleration (Intel ACC100/VRB1/VRB2)

41 vRAN Acceleration (Intel ACC100/VRB1/VRB2)

 

As communications service providers move from 4G to 5G networks, many are adopting virtualized radio access network (vRAN) architectures for higher channel capacity and easier deployment of edge-based services and applications. vRAN solutions are ideally located to deliver low-latency services with the flexibility to increase or decrease capacity based on the volume of real-time traffic and demand on the network.

One of the most compute-intensive 4G and 5G workloads is RAN layer 1 (L1) FEC, which resolves data transmission errors over unreliable or noisy communication channels. FEC technology detects and corrects a limited number of errors in 4G or 5G data, eliminating the need for retransmission. Since the FEC acceleration transaction does not contain cell state information, it can be easily virtualized, enabling pooling benefits and easy cell migration.

Historically, Intel provided the ACC100 vRAN Accelerator card to rapidly execute Layer 1 FEC algorithms, freeing up host processing power for the main CPU. Intel has since integrated this technology directly into newer CPUs, starting with Sapphire Rapids, under the name Intel vRAN Boost (VRB). Intel vRAN Boost acts as an offload accelerator on the CPU itself, eliminating the need for a separate hardware card. This section details configuration of SUSE Telco Cloud for workloads to leverage ACC100 or Intel vRAN Boost.

41.1 Kernel parameters

 

To enable the vRAN acceleration, we need to enable the following kernel parameters (if not present yet):

parametervaluedescription

iommu

pt

This option enables the use of vfio for the DPDK interfaces.

intel_iommu or amd_iommu

on

This option enables the use of vfio for VFs.

Modify the GRUB file /etc/default/grub to add them to the kernel command line:

GRUB_CMDLINE_LINUX="BOOT_IMAGE=/boot/vmlinuz-6.4.0-9-rt root=UUID=77b713de-5cc7-4d4c-8fc6-f5eca0a43cf9 skew_tick=1 rd.timeout=60 rd.retry=45 console=ttyS1,115200 console=tty0 default_hugepagesz=1G hugepagesz=1G hugepages=40 hugepagesz=2M hugepages=0 ignition.platform.id=openstack net.ifnames=1 intel_iommu=on iommu=pt irqaffinity=0,31,32,63 isolcpus=domain,nohz,managed_irq,1-30,33-62 nohz_full=1-30,33-62 nohz=on mce=off nosoftlockup nowatchdog nmi_watchdog=0 quiet rcu_nocb_poll rcu_nocbs=1-30,33-62 rcupdate.rcu_cpu_stall_suppress=1 rcupdate.rcu_expedited=1 rcupdate.rcu_normal_after_boot=1 rcupdate.rcu_task_stall_timeout=0 rcutree.kthread_prio=99 security=selinux selinux=1 idle=poll"

Update the GRUB configuration and reboot the system to apply the changes:

$ transactional-update grub.cfg
$ reboot

To verify that the parameters are applied after the reboot, check the command line:

$ cat /proc/cmdline

41.2 Configure SR-IOV on FEC Accelerators

 

  • Load vfio-pci kernel module to enable vRAN acceleration. First argument indicates vfio-pci module to support SR-IOV, and second argument prevents the PCI device from entering a low-power state when it’s idle.

$ modprobe vfio-pci enable_sriov=1 disable_idle_d3=1

  • Retrieve the PCI device address of FEC accelerator:

$ lspci -DPPnn | grep -i acc
0000:07:00.0 Processing accelerators [1200]: Intel Corporation Device [8086:57c2]
0000:0a:00.0 Processing accelerators [1200]: Intel Corporation Device [8086:57c2]

  • Bind the physical interface (PF) with vfio-pci driver:

$ dpdk-devbind.py -b vfio-pci 0000:07:00.0

  • Create the virtual functions (VFs) from the physical interface (PF).

Check the maximum VF capacity of the accelerator card. Next, configure the card to expose the desired number of VFs, not exceeding the maximum number. In this example, we configure the card for its full capacity of 16 VFs:

$ cat /sys/bus/pci/devices/0000:07:00.0/sriov_totalvfs
64
$ echo 16 > /sys/bus/pci/devices/0000:07:00.0/sriov_numvfs

  • Configure the accelerator card and its virtual functions with a 4G or 5G profile, selecting the right vRAN Boost device type depending on the exact Intel processor generation: VRB1 (formerly known as ACC200) if Sapphire Rapids Edge Enhanced (SPR-EE), VRB2 for Granite Rapids-D (GNR-D). A unique VF token (UUID) must be provided (the workload will consume this VF token to utilize the card’s FEC acceleration capabilities).

Warning
Warning

Deprecation from pf-bb-config: new rpm release supporting VRB2 device type is now installing the configuration example files in /usr/share/pf-bb-config/examples location; the old /opt/pf-bb-config/ path is still supported but to be removed in future releases. Please, update your scripts and documentation to reflect the new path (as done in the example below).

$ pf_bb_config VRB2 -c /usr/share/pf-bb-config/examples/vrb2/vrb2_config_vf_5g.cfg -f /usr/share/pf-bb-config/examples/vrb2/srs_fft_windows_coefficient.bin -v 00112233-4455-6677-8899-aabbccddeeff -p 0000:07:00.0
== pf_bb_config Version 25.11 ==
VRB2 PF [0000:07:00.0] configuration complete!
Log file = /var/log/pf_bb_cfg_0000:07:00.0.log

  • Verify the new VFs created from the FEC PF. Note that the VFs got 0d5d device ID. This information is required in next step to expose these VFs as Kubernetes resource:

$ # dpdk-devbind.py -s | grep -A18 Baseband
Baseband devices using DPDK-compatible driver
=============================================
0000:07:00.0 'Device 57c2' numa_node=0 drv=vfio-pci unused=
0000:07:00.1 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.2 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.3 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.4 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.5 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.6 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:00.7 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.0 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.1 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.2 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.3 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.4 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.5 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.6 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:01.7 'Device 57c3' numa_node=0 drv=vfio-pci unused=
0000:07:02.0 'Device 57c3' numa_node=0 drv=vfio-pci unused=
--
Other Baseband devices
======================
0000:0a:00.0 'Device 57c2' numa_node=0 unused=vfio-pci

Other Crypto devices
====================
0000:01:00.0 '420xx Series QAT 4946' numa_node=0 unused=vfio-pci
0000:03:00.0 'Device 2714' numa_node=0 unused=vfio-pci

DMA devices using kernel driver
===============================
0000:00:01.0 'Device 11fb' numa_node=0 drv=idxd unused=vfio-pci

Other Eventdev devices
======================
0000:03:00.0 'Device 2714' numa_node=0 unused=vfio-pci

No 'Mempool' devices detected
=============================

41.3 Configure Kubernetes for FEC Acceleration

 

The final step is exposing the VFs to Kubernetes with help of SR-IOV device plugin. Create a ConfigMap using the VFs' deviceID gathered from previous step, and install SR-IOV device plugin. Once Kubernetes nodes display the FEC VFs as Allocatable resources, the cluster is ready for the workloads to enjoy the FEC acceleration.

Follow the steps of Option 1 in Chapter 39, SR-IOV chapter. SRIOV Network Operator isn’t applicable for FEC Accelerator, so Option 2 is not applicable.