SPRADG0A April   2024  – August 2024 AM62P , AM62P-Q1

 

  1.   Abstract
  2.   2
  3.   Trademarks
  4. 1Introduction
    1. 1.1 Change Cortex-A53 Clock Frequency
  5. 2Processor Core and Compute Benchmarks
    1. 2.1 Dhrystone
    2. 2.2 CoreMark-Pro
    3. 2.3 Fast Fourier Transform
    4. 2.4 Cryptographic Benchmarks
    5. 2.5 IPC Mailbox Latency
  6. 3Memory System Benchmarks
    1. 3.1 Memory Bandwidth and Latency
      1. 3.1.1 LMBench
      2. 3.1.2 STREAM
    2. 3.2 Critical Memory Access Latency
    3. 3.3 UDMA: DDR to DDR Data Copy
  7. 4Graphics Processing Unit Benchmarks
    1. 4.1 Glmark2
    2. 4.2 GFXBench5
  8. 5Video Codec
  9. 6References
  10. 7Revision History

IPC Mailbox Latency

The AM62Px device contains a Mailbox IP as one of the primary methods of inter-processor communication (IPC). The Mailbox module serves to facilitate the communication between the various on-chip processors of the device by providing a queued mailbox-interrupt mechanism.

The queued mailbox-interrupt mechanism allows the software to establish communication channels between multiple processors (users) through a set of registers and associated interrupt signals with a small 32-bit payload.(1) The Mailbox consists of 8 groups of FIFOs (clusters) that support bidirectional communication between a maximum of 4 users.

Each cluster contains a series of 16 FIFOs, each supporting unidirectional communication between up to 4 users. Each FIFO holds up to 4 32-bit messages.

The measurements were made on the AM62Px platform using bare-metal silicon verification tests. The R5F cores operated out of local TCM, while A53 cores operated out of DDR. Each test included a loop of 32 send/receive iterations, whose results were averaged. Two methods of servicing received messages were used: one using processor interrupts (shown in Table 2-6), and the other using polling (shown in Table 2-7).

Table 2-6 IPC 32-Bit Latency with Interrupts

32-Bit Send/Receive Average Latency (ns)

Sending Cores

Receiving Cores

A53

R5F MCU

R5F WKUP

A53

693

409

340

R5F MCU

745

309

R5F WKUP

695

331

Table 2-7 IPC 32-Bit Latency with Polling

32-Bit Send/Receive Average Latency (ns)

Sending Cores Receiving Cores
A53 R5F MCU R5F WKUP
A53 521 502 471
R5F MCU 445 497
R5F WKUP 508 361
The Mailbox interrupts are statically routed as shown in Table 4-48 of the AM62Px Sitara Processors Technical Reference Manual.