New AGX Orin 64GB - no GPU?

Autonomous Machines Jetson & Embedded Systems Jetson AGX Orin
1

I just got the AGX Orin 64GB Development kit and it’s supposed to come with a gpu with 2048 cuda cores. I’ve run the update and install procedure for the JetPack SDK and installed everything I can think of that might be needed but I get nothing reported about a display device, period. lshw shows cpu cores but nothing else. Installing PyTorch and using torch.cuda.device_count() results in 0.

The one thing I do see mention of gpu is in tegrastats: 07-02-2023 13:15:03 RAM 1369/62797MB (lfb 14879x4MB) SWAP 0/31398MB (cached 0MB) CPU [0%@729,0%@729,0%@729,0%@729,0%@729,0%@729,0%@729,0%@729,off,off,off,off] EMC_FREQ 0% GR3D_FREQ 0% GR3D2_FREQ 0%@0 CV0@-256C CPU@37.687C Tboard@26C SOC2@34.593C Tdiode@27.5C SOC0@35.062C CV1@-256C GPU@-256C tj@37.562C SOC1@34.562C CV2@-256C

lshw output:

orin
description: Computer
product: Jetson AGX Orin
vendor: Unknown
version: Not Specified
serial: 1420723021530
width: 64 bits
capabilities: smbios-3.0.0 dmi-3.0.0 smp cp15_barrier setend swp tagged_addr_disabled
configuration: boot=normal family=Unknown sku=Unknown
*-core
description: Motherboard
product: Jetson AGX Orin
vendor: Unknown
physical id: 0
version: Not Specified
serial: 1420723021530
slot: Unknown
*-cache:0 DISABLED
description: L1 cache
physical id: 0
slot: L1 Instruction Cache
size: 64KiB
capacity: 64KiB
capabilities: internal instruction
configuration: level=1
*-cache:1 DISABLED
description: L1 cache
physical id: 1
slot: L1 Data Cache
size: 64KiB
capacity: 64KiB
capabilities: internal data
configuration: level=1
*-cache:2 DISABLED
description: L2 cache
physical id: 2
slot: L2 Cache
size: 256KiB
capacity: 256KiB
capabilities: internal unified
configuration: level=2
*-cache:3 DISABLED
description: L3 cache
physical id: 3
slot: L3 Cache
size: 2MiB
capacity: 2MiB
capabilities: internal unified
configuration: level=3
*-cpu:0
description: CPU
product: ARMv8 (Not Specified)
vendor: Not Specified
physical id: 4
bus info: cpu@0
version: Not Specified
serial: Not Specified
slot: CPU01
size: 729MHz
capacity: 2201MHz
clock: 31MHz
capabilities: lm cpufreq
configuration: cores=12 enabledcores=12 threads=1
*-firmware
description: BIOS
vendor: EDK II
physical id: 5
version: 2.1-32413640
date: 01/24/2023
size: 1MiB
capabilities: pci pnp upgrade shadowing cdboot bootselect acpi uefi
*-memory
description: Generic Memory
physical id: b
size: 61GiB
capacity: 62GiB
*-cpu:1 DISABLED
description: CPU
product: cpu
physical id: 6
bus info: cpu@0
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:2 DISABLED
description: CPU
product: cpu
physical id: 7
bus info: cpu@1
size: 729MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:3 DISABLED
description: CPU
product: cpu
physical id: 8
bus info: cpu@2
size: 729MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:4 DISABLED
description: CPU
product: cpu
physical id: 9
bus info: cpu@3
size: 729MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:5 DISABLED
description: CPU
product: cpu
physical id: a
bus info: cpu@4
size: 1728MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:6 DISABLED
description: CPU
product: cpu
physical id: c
bus info: cpu@5
size: 1728MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:7 DISABLED
description: CPU
product: cpu
physical id: d
bus info: cpu@6
size: 1728MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:8 DISABLED
description: CPU
product: cpu
physical id: e
bus info: cpu@7
size: 1728MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:9 DISABLED
description: CPU
product: cpu
physical id: f
bus info: cpu@8
size: 115MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:10 DISABLED
description: CPU
product: cpu
physical id: 10
bus info: cpu@9
size: 115MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:11 DISABLED
description: CPU
product: cpu
physical id: 11
bus info: cpu@10
size: 115MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:12 DISABLED
description: CPU
product: cpu
physical id: 12
bus info: cpu@11
size: 115MHz
capacity: 2201MHz
capabilities: cpufreq
*-cache
description: L1 Cache
physical id: 0
size: 64KiB
*-cpu:13 DISABLED
description: CPU
product: cpu_core_power_states
physical id: 13
bus info: cpu@12
*-cpu:14 DISABLED
description: CPU
product: cpu-map
physical id: 14
bus info: cpu@13
*-pci
description: PCI bridge
product: NVIDIA Corporation
vendor: NVIDIA Corporation
physical id: 100
bus info: pci@0001:00:00.0
version: a1
width: 32 bits
clock: 33MHz
capabilities: pci pm msi pciexpress msix normal_decode bus_master cap_list
configuration: driver=pcieport
resources: irq:51 ioport:1000(size=4096) memory:20a8000000-20a80fffff
*-network
description: Wireless interface
product: RTL8822CE 802.11ac PCIe Wireless Network Adapter
vendor: Realtek Semiconductor Co., Ltd.
physical id: 0
bus info: pci@0001:01:00.0
logical name: wlan0
version: 00
serial: 90:e8:68:84:4a:47
width: 64 bits
clock: 33MHz
capabilities: pm msi pciexpress bus_master cap_list ethernet physical wireless
configuration: broadcast=yes driver=rtl88x2ce driverversion=5.10.104-tegra latency=0 multicast=yes wireless=unassociated
resources: irq:311 ioport:1000(size=256) memory:20a8000000-20a800ffff
*-usbhost:0
product: xHCI Host Controller
vendor: Linux 5.10.104-tegra xhci-hcd
physical id: 1
bus info: usb@1
logical name: usb1
version: 5.10
capabilities: usb-2.00
configuration: driver=hub slots=4 speed=480Mbit/s
*-usb:0
description: Bluetooth wireless interface
product: Bluetooth Radio
vendor: Realtek
physical id: 3
bus info: usb@1:3
version: 0.00
serial: 00e04c000001
capabilities: bluetooth usb-1.00
configuration: driver=rtk_btusb maxpower=500mA speed=12Mbit/s
*-usb:1
description: USB hub
product: 4-Port USB 2.0 Hub
vendor: Generic
physical id: 4
bus info: usb@1:4
version: 1.40
capabilities: usb-2.10
configuration: driver=hub slots=4 speed=480Mbit/s
*-usbhost:1
product: xHCI Host Controller
vendor: Linux 5.10.104-tegra xhci-hcd
physical id: 2
bus info: usb@2
logical name: usb2
version: 5.10
capabilities: usb-3.10
configuration: driver=hub slots=4 speed=10000Mbit/s
*-usb
description: USB hub
product: 4-Port USB 3.0 Hub
vendor: Generic
physical id: 3
bus info: usb@2:3
version: 1.40
capabilities: usb-3.20
configuration: driver=hub slots=4 speed=10000Mbit/s
*-network:0 DISABLED
description: Ethernet interface
physical id: 3
logical name: dummy0
serial: 0e:46:fa:50:83:aa
capabilities: ethernet physical
configuration: broadcast=yes driver=dummy driverversion=5.10.104-tegra
*-network:1
description: Ethernet interface
physical id: 4
logical name: eth0
serial: 48:b0:2d:dc:fa:1d
size: 100Mbit/s
capacity: 10Gbit/s
capabilities: ethernet physical 100bt 100bt-fd 1000bt-fd 10000bt-fd autonegotiation
configuration: autonegotiation=on broadcast=yes driver=nvethernet driverversion=5.10.104-tegra duplex=full ip=192.168.2.85 link=yes multicast=yes port=twisted pair speed=100Mbit/s
*-network:2
description: Ethernet interface
physical id: 5
logical name: rndis0
serial: 8a:90:60:3e:34:6d
capabilities: ethernet physical
configuration: broadcast=yes driver=g_ether driverversion=29-May-2008 firmware=tegra-xudc link=yes multicast=yes
*-network:3
description: Ethernet interface
physical id: 6
logical name: usb0
serial: 8a:90:60:3e:34:6f
capabilities: ethernet physical
configuration: broadcast=yes driver=g_ether driverversion=29-May-2008 firmware=tegra-xudc link=yes multicast=yes

2

deviceQuery in the cuda samples is able to detect the cpu as being cuda capable.

The pytorch thing is probably because pycuda is not installed and won’t–which is just something to figure out. The other is, how do I specify that I want to specifically run a kernel on the cuda cores rather than the native ones with SYCL–another thing to learn.

So I think I’m good. Anyone else runs into this just compile /usr/local/cuda/samples/1_Utilities/deviceQuery and run it. Should get this:

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: “Orin”
CUDA Driver Version / Runtime Version 11.4 / 11.4
CUDA Capability Major/Minor version number: 8.7
Total amount of global memory: 62797 MBytes (65847095296 bytes)
(008) Multiprocessors, (128) CUDA Cores/MP: 1024 CUDA Cores
GPU Max Clock rate: 1300 MHz (1.30 GHz)
Memory Clock rate: 612 Mhz
Memory Bus Width: 128-bit
L2 Cache Size: 4194304 bytes
Maximum Texture Dimension Size (x,y,z) 1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
Maximum Layered 1D Texture Size, (num) layers 1D=(32768), 2048 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(32768, 32768), 2048 layers
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 49152 bytes
Total shared memory per multiprocessor: 167936 bytes
Total number of registers available per block: 65536
Warp size: 32
Maximum number of threads per multiprocessor: 1536
Maximum number of threads per block: 1024
Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)
Maximum memory pitch: 2147483647 bytes
Texture alignment: 512 bytes
Concurrent copy and kernel execution: Yes with 2 copy engine(s)
Run time limit on kernels: No
Integrated GPU sharing Host Memory: Yes
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
Device supports Unified Addressing (UVA): Yes
Device supports Managed Memory: Yes
Device supports Compute Preemption: Yes
Supports Cooperative Kernel Launch: Yes
Supports MultiDevice Co-op Kernel Launch: Yes
Device PCI Domain ID / Bus ID / location ID: 0 / 0 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.4, CUDA Runtime Version = 11.4, NumDevs = 1
Result = PASS

4

Hi,

Based on the deviceQuery log, your GPU’s functionality is good.
For PyTorch, you can find our package which already has CUDA enabled:

https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform/index.html

Thanks.

1 Like
5

Incidentally, any program which uses nivida-smi to detect the GPU will fail (that’s for discrete GPUs over PCI, but the Jetson’s GPU is integrated directly to the memory controller).

1 Like
7

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.