2399642_en-US

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

2399642_en-US

2399642_en-US

imx93 cannot load elf in mcimx93evk_ethosu_apps (sdk2.14)

I need try sdk2.14 demo:
mcimx93evk_ethosu_apps
after the elf is get .I load this elf to m33.
it failed:
                        remoteproc remoteproc0: Registered carveout doesn't fit da request
[ 236.970879] rproc-virtio: probe of rproc-virtio.1.auto failed with error -12
[ 236.984471] remoteproc remoteproc0: Registered carveout doesn't fit da request
[ 236.992036] rproc-virtio: probe of rproc-virtio.2.auto failed with error -12

I donnot know what happen.

Re: imx93 cannot load elf in mcimx93evk_ethosu_apps (sdk2.14)

Hi @bighero77,

Thank you for contacting NXP Support.

According to the README file, the application required for your use case is ethosu_apps_rpmsg, not ethosu_apps.

README extract:

4. Test process
---------------
4.1 Example ethosu_apps_rpmsg is built as .out or .elf and installed in rootfs. In Linux command line,
    user can run the following command to do inference for the model.
    #./inference_runner -d /dev/ethosu0 -n mobilenet_v1_1.0_224_pb_int8_vela.tflite -i stopwatch.bmp -l labels.txt -o output_data_detection -t 1080000000

    When the examples runs, the log and inference result would be seen on the terminal as below.

    Send Ping
    Send version request
    Send capabilities request
    Capabilities:
            version_status:1
            version:{ major=0, minor=0, patch=0 }
            product:{ major=6, minor=0, patch=0 }
            architecture:{ major=1, minor=0, patch=6 }
            driver:{ major=0, minor=16, patch=0 }
            macs_per_cc:8
            cmd_stream_version:0
            custom_dma:false
    Create network
    Create inference
    Wait for inferences
    Inference status: success

    Detected: jinrikisha, confidence:70
    Detected: horse cart, confidence:70
    Detected: academic gown, confidence:70
    Detected: howler monkey, confidence:70

    For other demos, such as face recognition, please refer to the user manual.

4.2 Example ethous_apps is built as .bin. In U-Boot prompt, user can run the following command to do
    inference for the conv2d op model.
    u-boot=> tftp 0x80000000 ethosu_apps.bin;cp.b 0x80000000 0x201e0000 0x20000;bootaux 0x201e0000 0

    When the example runs, the log and inference result would be seen on the terminal as below.

    Initialize Arm Ethos-U
    Inference status: success

Based on the documentation, if your goal is to run inference from Linux using /dev/ethosu0, then ethosu_apps_rpmsg is the correct application to use.

Best regards,
Alejandro Garcia

Tags (1)
No ratings
Version history
Last update:
3 weeks ago
Updated by: