MRL Sample Workload
The data streams optimizer provides a sample workload called MMIO Read Latency (MRL).
MRL is a workload created to measure MMIO read latency performance. The sample models a situation in which the core requests data from the network via PCIe device, such as the integrated Ethernet controller for Time-Sensitive Networking (TSN), or Intel® Ethernet Controller I225. The PCIe device returns the data to the core. The stream exercised in MRL is core from PCIe (MMIO reads).
Tuning this stream will impact the latency
reported by MRL.
Source Files
The tool provides MRL source files in the following directory:
- Host system:${TCC_TOOLS_PATH}/demo/workloads/src/mrl_workload
- Yocto Project*-based image:/usr/share/tcc_tools/tools/demo/workloads/src/mrl_workload
The following table provides a short description of each file:
File | Description |
---|---|
src/mrl_userspace_app.c | Main source for executable, sets up the PCIe device for timestamping and runs the workload |
The source files are compiled into the binary
./build/mrl
.