This section provides the performance numbers of Ethernet drivers using CPSW peripheral in MCU+ SDK
| SOC Details | Values |
|---|---|
| Core | R5F |
| Core Operating Speed | 400 MHz |
| Memory Type | MSRAM |
| Cache status | Enabled |
| Configuration | Value |
|---|---|
| Processing Core | Main R5F0 Core 0 |
| Core Frequency | 400 MHz |
| Ethernet Interface Type | RGMII at 1 Gbps |
| Packet buffer memory | MSRAM (un-cached) |
| Scatter-gather TX | Yes |
| Scatter-gather RX | Yes |
| CPDMA interrupt pacing | Yes |
| RTOS | FreeRTOS |
| RTOS application | Modified Enet CPSW Loopback Example example |
| Host PC tool version | nload |
| Rx packet length | 200 B |
| Tx packet length | 200 B |
| Parameter | CPU<->CPSW Latency Value (ns) | PHY (DP83869HM)Latency (from datasheet) in ns | Total Latency (ns) |
| RX Latency | 4756 | 193 | 4949 |
| TX Latency | 13225 | 384 | 13609 |
| Configuration | Value |
|---|---|
| EVM Type | AM261x-SOM |
| CPSW CPPI Frequency | 320 MHz |
| Ethernet Interface Type | RGMII |
| PHY model number | DP83869 |
| Ethernet Cable Spec | CAT-6 |
| Sample size | 10 Million Packets |
| RTOS application | Modified Enet Layer 2 CPSW SWITCH Example example |
| Datagram Length | Latency @ 1Gbps link (in us) | Latency @ 100Mbps link (in us) | Latency @ 10Mbps link (in us) | |||
| Without Cut-Through | With Cut-Through | Without Cut-Through | With Cut-Through | Without Cut-Through | With Cut-Through | |
| 64 Bytes | 1.8 | 1.7 | 7.9 | 7.8 | 73.4 | 71.9 |
| 256 Bytes | 3.1 | 2.0 | 23.0 | 13.8 | 227.3 | 135.6 |
| 512 Bytes | 5.2 | 2.0 | 43.5 | 13.8 | 432.0 | 135.6 |
| 1518 Bytes | 13.2 | 2.0 | 124.0 | 13.8 | 1237.0 | 135.6 |
| Configuration | Value |
|---|---|
| Processing Core | Main R5F0 Core 0 |
| Core Frequency | 400 MHz |
| Ethernet Interface Type | RGMII at 1 Gbps |
| Packet buffer memory | MSRAM (cached) |
| Hardware checksum offload | Enabled on Tx side Enabled on Rx Side |
| Scatter-gather TX | Yes |
| Scatter-gather RX | Yes |
| CPDMA interrupt pacing | Yes |
| RTOS | FreeRTOS |
| RTOS application | Enet Lwip CPSW Example example |
| TCP/IP stack | LwIP version STABLE-2_2_0_RELEASE |
| Host PC tool version | iperf v2.0.10 |
| Number of Rx packet buffers | 32 |
| Number of Tx packet buffers | 16 |
| Test | Bandwidth (Mbps) | CPU Load (%) |
| TCP RX | 195 | 98 |
| TCP TX | 158 | 100 |
| TCP Bidirectional | RX=73 TX=87 | 100 |
Host PC commands:
iperf -c <evm_ip> -riperf -c <evm_ip> -d| Test | Datagram Length = 64B | Datagram Length = 256B | Datagram Length = 512B | Datagram Length = 1470B | ||||||||
| Bandwidth (Mbps) | CPU Load (%) | Packet Loss (%) | Bandwidth (Mbps) | CPU Load (%) | Packet Loss (%) | Bandwidth (Mbps) | CPU Load (%) | Packet Loss (%) | Bandwidth (Mbps) | CPU Load (%) | Packet Loss (%) | |
| UDP RX | 5 | 41 | 0.0 | 5 | 20 | 0.0 | 25 | 48 | 0.0 | 50 | 42 | 0.0 |
| 10 | 72 | 0.0 | 15 | 53 | 0.0 | 50 | 85 | 0.7 | 60 | 48 | 0.0 | |
| 15 | 91 | 7.3 | 25 | 81 | 0.15 | 55 | 95 | 0.4 | 95 | 72 | 0.0 | |
| UDP RX (Max) | 13 | 91 | 1.3 | 30 | 96 | 3.4 | 58 | 97 | 1.1 | 130.8 | 97 | 1.4 |
| UDP TX (Max) | 23.7 | 100 | 0.0 | 54.1 | 100 | 0.0 | 108 | 100 | 0.0 | 311 | 100 | 0.0 |
Host PC commands:
iperf -c <evm_ip> -u -l64 -b <bw> -riperf -c <evm_ip> -u -l256 -b <bw> -riperf -c <evm_ip> -u -l512 -b <bw> -riperf -c <evm_ip> -u -b <bw> -rLwip Application size: 793KB
FLASH: 198KB
| Region | Memory |
| .text | 156KB |
| .rodata | 42KB |
OCRAM: 595KB
| Region | Memory |
| .bss | 471KB |
| .sysmem | 34KB |
| .data | 23KB |
| .stack | 8KB |
| misc | 59KB |
| Board POV | MSRAM - OOB | XIP : text and ro data in flash | Opti-flash: RL2 8KB | Opti-flash: RL2 16KB | Opti-flash: RL2 32KB | Opti-flash: RL2 64KB | Opti-flash: RL2 128KB |
| UDP Tx | 273Mbps at 100% CPU load | 145Mbps at 100% CPU load | 224Mbps at 100% CPU load | 252Mbps at 100% CPU load | 270Mbps at 100% CPU load | 273Mbps at 100% CPU load | 273Mbps at 100% CPU load |
| UDP Rx | 116Mbps ~1% pkt drop 86% CPU load | 45Mbps ~1% pkt drop 98% CPU load | 64Mbps ~1% pkt drop 99% CPU load | 92Mbps ~1% pkt drop 97% CPU load | 108Mbps ~1% pkt drop 89% CPU load | 110Mbps ~1% pkt drop 86% CPU load | 115Mbps ~1% pkt drop 89% CPU load |