Hello, I recently purchased an NVIDIA Jetson Orin Developer Kitm but is showing only 6.3GB of RAM instead of 8GB. Can anyone help me with this issue? Thank you in advance.
Hi,
Pease run sudo tegrastats and check if you see similar output like:
$ sudo tegrastats
04-12-2023 07:57:07 RAM 1229/7337MB (lfb 1228x4MB) SWAP 0/3668MB (cached 0MB) CP
U [0%@729,0%@729,0%@729,0%@729,0%@729,0%@729] EMC_FREQ 0%@2133 GR3D_FREQ 0%@607
GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@47.625C SOC2@47.125C SOC0@44.
781C CV1@-256C GPU@47.937C tj@47.937C SOC1@47.468C CV2@-256C VDD_IN 4856mW/4856m
W VDD_CPU_GPU_CV 675mW/675mW VDD_SOC 1512mW/1512mW
04-12-2023 07:57:08 RAM 1229/7337MB (lfb 1228x4MB) SWAP 0/3668MB (cached 0MB) CP
U [1%@729,2%@729,2%@729,1%@729,2%@729,2%@729] EMC_FREQ 0%@2133 GR3D_FREQ 0%@607
GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@47.781C SOC2@47.281C SOC0@44.
812C CV1@-256C GPU@47.812C tj@47.812C SOC1@47.406C CV2@-256C VDD_IN 4896mW/4876m
W VDD_CPU_GPU_CV 715mW/695mW VDD_SOC 1512mW/1512mW
04-12-2023 07:57:09 RAM 1229/7337MB (lfb 1228x4MB) SWAP 0/3668MB (cached 0MB) CP
U [0%@849,0%@729,0%@729,0%@729,0%@729,0%@729] EMC_FREQ 0%@2133 GR3D_FREQ 0%@607
GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@47.562C SOC2@47.218C SOC0@44.
843C CV1@-256C GPU@47.968C tj@47.968C SOC1@47.5C CV2@-256C VDD_IN 4856mW/4869mW
VDD_CPU_GPU_CV 675mW/688mW VDD_SOC 1512mW/1512mW
04-12-2023 13:30:20 RAM 1711/6481MB (lfb 754x4MB) SWAP 0/3240MB (cached 0MB) CPU [0%@729,0%@729,11%@729,0%@729,1%@729,26%@729] EMC_FREQ 2%@665 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 435 APE 200 CV0@-256C CPU@43.156C SOC2@42.937C SOC0@41.656C CV1@-256C GPU@42.656C tj@43.156C SOC1@41.625C CV2@-256C VDD_IN 3858mW/3858mW VDD_CPU_GPU_CV 561mW/561mW VDD_SOC 1083mW/1083mW
That what I get.
Also the boot time is some how long.
Hi,
Please share the information for reference:
$ cat /etc/nv_boot_control.conf
And do you use Orin Nano module with or without SD card slot? There are two kinds of modules and not sure which one you are using.
TNSPEC 3767-300-0005-K.2-1-0-jetson-orin-nano-devkit-
COMPATIBLE_SPEC 3767–0005–1–jetson-orin-nano-devkit-
TEGRA_LEGACY_UPDATE false
TEGRA_BOOT_STORAGE mmcblk1
TEGRA_EMMC_ONLY false
TEGRA_CHIPID 0x23
TEGRA_OTA_BOOT_DEVICE /dev/mtdblock0
TEGRA_OTA_GPT_DEVICE /dev/mtdblock0
I am using an SD card for the system.
In the pre-production Jetson Orin Nano Developer Kit you need to flash the QSPI-NOR memory to update to the production JetPack release. You can do this by doing a full flash with the SDK Manager, or download the BSP and use the flash.sh utility.
This should fix the issue with the extra debug memory carve outs, after flashing the QSPI you will have ~ 7.2GB of memory
I have the same exact problem using the official SD card image. Can you elaborate more on how to do the full flash or download the BSP?
The SDK Manager is available here: SDK Manager | NVIDIA Developer
Download it and install on a host x86 Ubuntu machine.
You will need a USB cable, USB C for the Jetson side. The other end depends on what your x86 machine supports.
Documentation for the SDK Manager is here: NVIDIA SDK Manager Documentation This includes how to download and install on the x86 host.
Installing software on the Jetson in particular: Install Jetson Software with SDK Manager :: NVIDIA SDK Manager Documentation
Here’s a video, you will need to adjust it for your needs: NVIDIA SDK Manager Tutorial: Installing Jetson Software Explained - YouTube
If you flash the Jetson SD card using the SDK Manager, it will also flash the QSPI at the same time.
Good luck!
Thanks! I have experience with TX2 NX and it sounds similar.
Is there any way to do it without the SDK manager? I am doing it mainly for education, and it will be hard to ask users to install Linux on the host machine. It will be easier (for developers as well) to get the whole RAM from the official SD card image method.
I am curious. Did you get this from a retailer, or did NVIDIA give you one for review? The reason I ask is that the review units typically don’t have the production version of the software, including the QSPI, installed.
The QSPI is not part of the SD card image. QSPI is a type of flash memory, a chip on the Jetson module. The firmware code within is used during the boot process. Here’s an article on how everything fits together on JetsonHacks.
With that said, there’s currently no way to flash the QSPI from a running Jetson system. I would guess that this is mostly to due to security reasons. Being able to exploit a boot loader makes bad guys happy.
The NVIDIA engineers probably know more about this.
1 Like
Hi,
We see identical result for Orin NX 8GB without SD and Orin NX 8GB with SD modules:
$ cat /etc/nv_boot_control.conf
TNSPEC 3767-RC1-0005-B.4-1-1-jetson-orin-nano-devkit-
COMPATIBLE_SPEC 3767--0005--1--jetson-orin-nano-devkit-
TEGRA_LEGACY_UPDATE false
TEGRA_BOOT_STORAGE mmcblk1
TEGRA_EMMC_ONLY false
TEGRA_CHIPID 0x23
TEGRA_OTA_BOOT_DEVICE /dev/mtdblock0
TEGRA_OTA_GPT_DEVICE /dev/mtdblock0
$ sudo tegrastats
04-17-2023 04:49:01 RAM 1026/7337MB (lfb 1261x4MB) SWAP 0/3668MB (cached 0MB) CPU [1%@1510,1%@729,0%@729,10%@729,9%@729,0%@729] EMC_FREQ 0%@204 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@42.656C SOC2@42.218C SOC0@40.031C CV1@-256C GPU@43.156C tj@43.156C SOC1@41.531C CV2@-256C VDD_IN 3542mW/3542mW VDD_CPU_GPU_CV 522mW/522mW VDD_SOC 1046mW/1046mW
04-17-2023 04:49:02 RAM 1026/7337MB (lfb 1261x4MB) SWAP 0/3668MB (cached 0MB) CPU [26%@1510,14%@729,4%@729,1%@729,0%@729,0%@729] EMC_FREQ 0%@204 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@42.656C SOC2@42.187C SOC0@40.031C CV1@-256C GPU@43.031C tj@43.031C SOC1@41.562C CV2@-256C VDD_IN 3542mW/3542mW VDD_CPU_GPU_CV 522mW/522mW VDD_SOC 1046mW/1046mW
04-17-2023 04:49:03 RAM 1026/7337MB (lfb 1261x4MB) SWAP 0/3668MB (cached 0MB) CPU [5%@1510,0%@1510,26%@1510,0%@1510,0%@729,0%@729] EMC_FREQ 0%@665 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@42.75C SOC2@42.187C SOC0@40.062C CV1@-256C GPU@42.906C tj@42.906C SOC1@41.687C CV2@-256C VDD_IN 3542mW/3542mW VDD_CPU_GPU_CV 562mW/535mW VDD_SOC 1046mW/1046mW
04-17-2023 04:49:04 RAM 1026/7337MB (lfb 1261x4MB) SWAP 0/3668MB (cached 0MB) CPU [1%@729,0%@729,0%@729,0%@729,0%@729,0%@729] EMC_FREQ 0%@204 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@42.562C SOC2@42.093C SOC0@39.906C CV1@-256C GPU@42.875C tj@42.875C SOC1@41.5C CV2@-256C VDD_IN 3341mW/3491mW VDD_CPU_GPU_CV 482mW/522mW VDD_SOC 1046mW/1046mW
04-17-2023 04:49:05 RAM 1026/7337MB (lfb 1261x4MB) SWAP 0/3668MB (cached 0MB) CPU [3%@1510,14%@1510,7%@1510,0%@1510,0%@729,14%@729] EMC_FREQ 0%@204 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 115 APE 200 CV0@-256C CPU@42.843C SOC2@42.25C SOC0@40.062C CV1@-256C GPU@42.843C tj@42.843C SOC1@41.75C CV2@-256C VDD_IN 3462mW/3485mW VDD_CPU_GPU_CV 522mW/522mW VDD_SOC 1046mW/1046mW
Please try to re-flash system image and check if you can get RAM 1026/7337MB
Hi,
I bought the development kit from an official reseller for nvidia “openzeka”. I downloaded the latest sd card image from nvidia website and ran it.
Anyway, right now I used the sdk manager to flash the system and now everything is working correct.
2 Likes
I got mine from seeed studio.
Here’s information from nv_boot_control.conf
cat /etc/nv_boot_control.conf
TNSPEC 3767-300-0005-K.2-1-0-jetson-orin-nano-devkit-
COMPATIBLE_SPEC 3767--0005--1--jetson-orin-nano-devkit-
TEGRA_LEGACY_UPDATE false
TEGRA_BOOT_STORAGE mmcblk1
TEGRA_EMMC_ONLY false
TEGRA_CHIPID 0x23
TEGRA_OTA_BOOT_DEVICE /dev/mtdblock0
TEGRA_OTA_GPT_DEVICE /dev/mtdblock0
I will try SDK manager as well.
1 Like
Hi,
The factory image may not be stable. One you receive the developer kits, we would suggest re-flash to a production release through SDKManager first.
Just received my jetson orin nano 8GB developer kit and can confirm this as well with the official SD image.
04-20-2023 22:31:22 RAM 2947/6481MB (lfb 3x1MB) SWAP 1168/3240MB (cached 5MB) CPU [67%@1510,58%@1510,91%@1510,64%@1510,62%@1510,92%@1510] EMC_FREQ 16%@665 GR3D_FREQ 0%@305 GR3D2_FREQ 0%@0 VIC_FREQ 435 APE 200 CV0@-256C CPU@47.593C SOC2@46.218C SOC0@44.375C CV1@-256C GPU@46.937C tj@47.593C SOC1@45.75C CV2@-256C VDD_IN 6089mW/6089mW VDD_CPU_GPU_CV 2400mW/2400mW VDD_SOC 1201mW/1201mW
Does re-installing the debian package nvidia-l4t-bootloader flash the QSPI chip as well? https://repo.download.nvidia.com/jetson/t234/pool/main/n/nvidia-l4t-bootloader/nvidia-l4t-bootloader_35.3.1-20230319081403_arm64.deb
sudo apt install --reinstall nvidia-l4t-bootloader
How else would the QSPI be updated as necessary through debian based OTA updates?
@DaneLLL I reinstalled the nvidia-l4t-bootloader package which downloaded the latest version from nvidia apt repos. I saw logging that the bootloader capsule was being updated in apt logs
Starting bootloader post-install procedure.
INFO. Copy capsule payload: TEGRA_BL.Cap.
INFO. Copy capsule payload to /opt/nvidia/esp/EFI/UpdateCapsule/ and l4tlauncher to /opt/nvidia/esp/EFI/BOOT/ done.
INFO. Set capsule UEFI variable /sys/firmware/efi/efivars/OsIndications-8be4df61-93ca-11d2-aa0d-00e098032b8c done.
however the issue is NOT fixed and firmware is still dated 2023-02-17. Are you certain that the debs have the latest bootloader with the debug carveouts removed? or can the QSPI not get flashed through an OTA on the orin?
Hi,
Are you able to re-flash the system through SDKManager? This method should work well. We would suggest do complete re-flash to replace factory image with production image.
@DaneLLL yes I was able to flash using SDKManager and am now on the latest 2023-03-19. However there are issues as I detailed here and I would like to go back for verfication Significantly LOWER performance when using exclusive fullscreen - #5 by theofficialgman