N
I upgraded my MSI sparks from stock 1tb to 2tb Corsair MP600 Micro 2242 (pcie gen4). I installed the new drive, downloaded the recovery image from the MSI website and wrote the image to a USB drive and followed the instructions to install the recovery image on the sparks. Very easy and everything went smoothly.
N
hi: 另外我想确认下,设备树配置的分辨率对不对 还有驱动里面这里要同步设置成1537呢 2022 static const struct camera_common_frmfmt ar0234_frmfmt[] = { 2023 {{1920, 1536+1}, ar0234_30fps, 1, 0, AR0234_MODE_1920X1080_CROP_30FPS}, 2024 };
N
Platform and Pin Information Platform: Jetson AGX Orin, JetPack 6.2.1 GPIO:GPIO3_PH.05 Pin function:UART4_RTS Pinmux name:uart4_rts_ph5 Signal name:GP120_UART4_RTS_N_BOOT0_STRAP Usage: GPIO output Observed Behavior A brief high pulse is observed at the output of the external level shifter immediately after power-on . The signal subsequently transitions to low. We suspect that the transition to low occurs when MB1 applies the GPIO configuration, but this has not yet been verified against the…
N
So then the eager indie masses will be resorting to external solutions? Is that the gist of it? Because if history shows anything, it that the timelines on this sort of stuff is… well it’ll be a long time. I’m super-stoked you responded here with even that informative, albeit not quite what anyone was hoping to hear, news… It seems like “unsanctioned” groups like RenoDX are having much faster iterative development and release cycles with DLSS NR, and there is already an API, and tons of other…
N
I am strongly considering going in the opposite direction: rather than replacing the OEM 1TB in my Asus with a faster Corsair drive, rather replace it with something like the 2230 WD Black SN770M 2TB. It will be a tad slower but it is TLC, has good endurance and 2TB gives quite a bit more breathing room. It looks like it can sustain its speeds much longer than the other 2230 options. It is also by far the cheapest 2TB TLC drive I can find, roughly half the price of the Corsair. The OEM 1TB…
N
Thanks for providing r36_5_TEST_yuv444_lossless.zip. We run on Syslogic carrier boards (RML A4NX / RML A4AGX) with D3 Embedded GMSL2 camera drivers. Syslogic does not provide an L4T R36.5 (JetPack 6.2.2) BSP for these boards—their supported releases jump directly from JetPack 6.2.1 (L4T 36.4.4 / 36.4.7) to JetPack 7.2.1 (L4T 39.2.1). When we test r36_5_TEST_yuv444_lossless/{libtegrav4l2.so,libnvvideo.so} on our current L4T R36.4.7 stack (nvidia-l4t-multimedia / nvidia-l4t-gstreamer…
N
Hello EricGggg, Thank you for confirming that the three QC stickers must be removed before applying TIM and installing the heatsink. I have reviewed Section 4.4 of the Jetson Orin NX/Nano Thermal Design Guide v1.5, but I could not find instructions for removing these stickers safely. Could you please clarify the recommended removal procedure for the stickers directly on the SoC contact surface? What tool or method should be used to lift and peel them off without scratching or damaging the SoC…
N
Hi, We have verified enabling iwlwifi on Jetpack 6.2.2. Please check Making sure you're not a bot! Enable iwlwifi driver If there is further query, @emmanuel.grumbach may be able to provide suggestion.
N
I Maybe spoke too soon. i have just reviewed the logs for the boot. it seems the NVIDIA kernel flip timeout that i was experiencing is now resolved but i then had a crash caused by what i think is my IP KVM being plugged in which is causing a race condition just as the display manager starts. so allthough the symptom looks the same it is in fact no longer being caused by the kernel flip. the issue i now see is now bieng caused by https://bugs.kde.org/show_bug.cgi?id=528890 so thankyou for your…
N
I can confirm for me that there was no change in behaviour. I have just realised however if i shutdown and do essentially a cold boot and power on again before my monitors goto sleep then the system boots fine. but if i wait for my monitors to goto sleep, then the system will fail to boot
N
Thanks man ! that was exactly the information I was looking for, I owe you a beer :)
N
Hi All, As a diagnostic experiment, could someone try setting the nvidia_modeset.opportunistic_flip_sync=0 kernel module parameter and report whether it changes the behavior?
N
Hi All, As a diagnostic experiment, could someone try setting the nvidia_modeset.opportunistic_flip_sync=0 kernel module parameter and report whether it changes the behavior?
N
Boost the clocks to try. sudo su echo 1 > /sys/kernel/debug/bpmp/debug/clk/vi/mrq_rate_locked echo 1 > /sys/kernel/debug/bpmp/debug/clk/isp/mrq_rate_locked echo 1 > /sys/kernel/debug/bpmp/debug/clk/nvcsi/mrq_rate_locked echo 1 > /sys/kernel/debug/bpmp/debug/clk/emc/mrq_rate_locked cat /sys/kernel/debug/bpmp/debug/clk/vi/max_rate |tee /sys/kernel/debug/bpmp/debug/clk/vi/rate cat /sys/kernel/debug/bpmp/debug/clk/isp/max_rate | tee /sys/kernel/debug/bpmp/debug/clk/isp/rate cat…
N
You may reference to below sample. #include <Argus/Argus.h> #include <Argus/Ext/SensorPrivateMetadata.h> #include <iostream> #include <vector> using namespace Argus; bool enablePrivateMetadata(CameraDevice *cameraDevice, Request *request) { // 1. 檢查 CameraDevice 是否支援私有詮釋資料 Ext::ISensorPrivateMetadataCaps *metaCaps = interface_cast<Ext::ISensorPrivateMetadataCaps>(cameraDevice); if (!metaCaps) { std::cerr << "Device does not support ISensorPrivateMetadataCaps" << std::endl; return false; } // 2.…
N
openrouter is good, but 50rpd is kinda harsh
N
usually, the new model wont last long
N
This can because the newer version is about to come soon
N
Thanks everyone for the reports on “Flip event timeout” / nv_flip == NULL / page-flip timeout issue on driver 620.70.06 with multiple displays We’ve filed internal bug 7126445 , and we’re working on reproducing it internally. As a diagnostic experiment, could someone try setting the nvidia_modeset.opportunistic_flip_sync=0 kernel module parameter and report whether it changes the behavior?
N
Hi whitesscott, Not intentional. Those headers should go under the install prefix, e.g. /usr/local/include/hololink/core, not /hololink. The SIPL CMake only adds hololink/src/hololink/core. That install() uses ${CMAKE_INSTALL_INCLUDEDIR}/hololink/core. The SIPL top-level CMakeLists.txt never includes GNUInstallDirs, so CMAKE_INSTALL_INCLUDEDIR is empty and CMake treats /hololink/core as an absolute path. libhololink_core.a still goes to /usr/local/lib because the library rule falls back to lib…
N
hi: 看了文档没看明白,有没有demo提供一版呢,看了这个介绍不知道如何读取,谢谢!
N
Dear NV team, This is to report an issue with the intel Wi-Fi driver used for jetson agx orin devkit. 1) JetPack 6.2.2 The Wi-Fi firmware crashes, and approximately 100–190 ms later, nvargus-daemon also starts failing, preventing my ZED X camera from being brought up successfully. I have reproduced this issue three times, and the behavior is consistent. Evidence: a) Wi-Fi kernel log (November 19): 18:58:49.372019 wlP1p1s0: cannot understand ECSA IE operating class, 10, ignoring 18:58:49.476218…
N
Same multi-monitor page-flip issue on RTX 4090 with 620.73.04 I’m observing the same NVIDIA DRM problem on a different GPU generation and Linux configuration. My setup consists of an RTX 4090, Arch Linux kernel 7.3.5-arch1-1, NVIDIA Open 620.73.04, KDE Plasma 6.8.3, and two 2560×1440 DisplayPort monitors. During some boots, SDDM’s Wayland greeter stalls while initializing the displays. The kernel reports Flip event timeout on heads 0 and 1, followed by nv_flip == NULL. The error persists with…
N
I prepared a Hyperflare (at 1a53275) runtime package for GLM-5.3-Flash that runs across two DGX Sparks (GB10 / SM121), and a detailed (human-written) guide on how to serve or reproduce it. The package can be served with vision support, a full 1,048,576-token context, and around 4.5 million KV tokens. TLDR on why this may interest you: llama-benchy results. You don’t enjoy using Docker and fancy a host-native solution. You don’t enjoy messing with an endless amount of environment variables. vLLM…
N
Additional evidence for the NVIDIA DRM event-handling regression in 620.73.04 I have been investigating intermittent NVIDIA DRM failures on an RTX 4090 running Arch Linux, kernel 7.3.5-arch1-1, NVIDIA Open 620.73.04, and KDE Plasma 6.8.3 on Wayland. I noticed the recently published patch addressing the NVIDIA DRM event-handling deadlock: Commit gulafaran/open-gpu-kernel-modules nvidia-drm: handle display events off the NVKMS event thread committed 04:28AM - 19 Nov 26 UTC VeilSilence +463 -69…
N
Thanks everyone for the reports and for sharing logs and workarounds. We’ve filed internal bug 7126445 to track the “Flip event timeout” / nv_flip == NULL / page-flip timeout issue on driver 620.70.06 with multiple displays, and we’re working on reproducing it internally.
N
Thanks everyone for the reports and for sharing logs and workarounds. We’ve filed internal bug 7126445 to track the “Flip event timeout” / nv_flip == NULL / page-flip timeout issue on driver 620.70.06 with multiple displays, and we’re working on reproducing it internally.
N
RTX 4090 update: regression persists with 620.73.04, including at default GPU settings I performed further testing on Arch Linux after upgrading to NVIDIA Open Kernel Modules 620.73.04 and Linux 7.3.5-arch1-1. The original DRM regression remains reproducible. The same Flip event timeout and nv_flip == NULL errors still occur, although KDE Plasma has successfully initialized during my recent tests. System configuration GPU: NVIDIA GeForce RTX 4090 Kernel: 7.3.5-arch1-1 NVIDIA: nvidia-open…
N
Thanks. Yes, this patch is already in my patch collection, too. During review and testing, I found a few more fixes which I pushed to my branch.
N
For Argus using below API for it. docs.nvidia.com Jetson Linux API Reference: Argus::Ext::ISensorPrivateMetadata Class Reference...