![Community banner](https://lemmy.world/pictrs/image/2de2bb6c-cadf-4f95-b051-845e2fde1735.jpeg)
![nvidia](https://lemmy.world/pictrs/image/b659d27a-ce88-410b-8599-1d4982ad9219.png?format=webp&thumbnail=48)
nvidia
- Nvidia loses a cool $500B as market questions AI boomwww.theregister.com Nvidia share price drop prompts concerns of AI bubble burst
Cisco was briefly the world's most valuable company too, you know, just before the dot com bust
- Nvidia Conquers Latest AI Testsspectrum.ieee.org Nvidia Conquers Latest AI Tests
GPU maker tops new MLPerf benchmarks on graph neural nets and LLM fine-tuning
- Nvidia now has the best business brand reputation in Americaqz.com Nvidia now has the best business brand reputation in America
The chipmaker had the top score for categories including trajectory, vision, and growth, according to the 2024 Axios Harris Poll 100
- Nvidia’s Enormous Financial Success Becomes . . . Normalwww.nextplatform.com Nvidia’s Enormous Financial Success Becomes . . . Normal
For the past five years, since Nvidia acquired InfiniBand and Ethernet switch and network interface card supplier Mellanox, people have been wondering
- Nvidia shares pass $1,000 for first time on AI-driven sales surgewww.cnbc.com Nvidia shares pass $1,000 for first time on AI-driven sales surge
Nvidia reported earnings after the bell. Here are the results.
- Top500 Supers: This Is Peak Nvidia For Accelerated Supercomputerswww.nextplatform.com Top500 Supers: This Is Peak Nvidia For Accelerated Supercomputers
If high bandwidth memory was widely available and we had cheap and reliable fusion power, there never would have been a move to use GPU and other compute
- Green500 shows Nvidia's Grace-Hopper superchip is a power-efficiency beastwww.theregister.com Nvidia's GH200 is a power-efficiency beast, Green500 shows
Three most energy-friendly systems all driven by Nv's CPU-GPU chip. AMD's APU could soon change that
- NVIDIA GeForce RTX 5090 and RTX 5080 launch to be separated by a few weeksvideocardz.com NVIDIA GeForce RTX 5090 and RTX 5080 launch to be separated by a few weeks - VideoCardz.com
NVIDIA RTX 5090 & 5080 announcement later this year The company plans to launch its new GPUs one by one. According to XpeaGPU, NVIDIA is set to launch its new flagship RTX 5090 and high-end RTX 5080 GPUs within a few weeks of each other. Both SKUs are expected to be announced simultaneously in 2024...
- Valve Working On Explicit Sync Support For "NVK" NVIDIA Vulkan Driver
Looks like Valve is making progress towards releasing SteamOS on other platforms.
- NVIDIA RTX Remix: DLSS 3.5 With Ray Reconstruction Now Available To Further Enhance Fully Ray-Traced Modswww.nvidia.com NVIDIA RTX Remix: DLSS 3.5 With Ray Reconstruction Now Available To Further Enhance Fully Ray-Traced Mods
Also, check out the latest RTX Remix mods for Dark Messiah of Might and Magic, Deus Ex, and Half-Life 2 RTX: An RTX Remix Project.
- KDE On The Importance Of Wayland Explicit Sync
Original blog: https://zamundaaa.github.io/wayland/2024/04/05/explicit-sync.html
- Nvidia Tops Llama 2, Stable Diffusion Speed Trialsspectrum.ieee.org Speed Tests for Llama 2, Stable Diffusion
MLPerf releases new LLM and generative AI inferencing tests
- Q&A with Nvidia's Chief of DGX Systems on the DGX-GB200 Rack-scale Systemwww.hpcwire.com Q&A with Nvidia's Chief of DGX Systems on the DGX-GB200 Rack-scale System
Pictures of Nvidia’s new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to […]
- Tiny Corp launches Nvidia-powered AI computer because 'it just works'www.theregister.com Tiny Corp reluctantly launching Nvidia-powered AI PC
Startup slams AMD for buggy firmware
- One rack. 120kW of compute. Taking a closer look at Nvidia's DGX GB200 NVL72 beastwww.theregister.com A closer look at Nvidia's 120kW DGX GB200 NVL72 rack system
1.44 exaFLOPs of FP4, 13.5 TB of HBM3e, 2 miles of NVLink cables, in one liquid cooled unit
- Nvidia’s Blackwell Ignites the Generative AI Supernova! Expert Breakdown here.medium.com Nvidia’s Blackwell Ignites the Generative AI Supernova! Expert Breakdown here.
Before the keynote speech, Nvidia’s CEO, Jensen Huang, said, “I hope you realize this is not a concert.” He sure left the audience at the…
#What is Nvidia Blackwell? Named after David Blackwell, the first black scholar inducted into the National Academy of Science, Blackwell (GB200, which stands for Grace-Blackwell 200) follows the success of its predecessor, Hopper(GH100, which stands for Grace-Hopper 100), which boosted Nvidia’s sales.
For anybody who is not familiar with the naming conventions of Nvidia’s chip world, G or Grace is the CPU used in Nvidia’s AI computing modules called DGX, and B or Blackwell is a GPU, which is an upgrade from the previous H or Hopper.
Nvidia Blackwell and the whole DGXGB200 SuperPOD AI computer Nvidia Blackwell and the whole DGXGB200 SuperPOD AI computer Apart from the apparent benchmark scores that Nvidia proudly wears on its chest, the most assuming part is that Nvidia has taken two dies or chips and just merged them without any loss of computational capabilities, giving birth to GB200, which is insane from an engineering perspective.
One of the direct competitors, AMD MI300, will indeed be shivering when they hear these numbers.
The new Blackwell ship is 4–5 times faster than its predecessor, Hopper. But when integrated into its environment of DGX Module, it becomes a beast and outputs almost 30x the inference performance, yes, not 30% but 30 times, and wait, wait, there is more, all while consuming 25X less power again that is not 25% its 25 times. This allows LLM’s models to train on over 27 trillion parameters. For comparison, the famous ChatGPT 4.0 was trained on 1.7 trillion parameters. Blackwell will be made in partnership with Nvidia’s best friend, TSMC, during a fantastic 3 nm fabrication process. It has 208 billion transistors, allowing data speeds up to 10Tb/sec. Amazing right?
#How is Blackwell(GB200) better than Hopper GPU(GH100)?
The previous generation of AI-optimized GPUs was called Hopper. Blackwell is between 2 and 30 times faster, depending on how you measure it. Huang explained that it took 8,000 GPUs, 15 megawatts, and 90 days to create the GPT-MoE-1.8T model. With the new system, you could use just 2,000 GPUs and 25% of the power.
Blackwell is genuinely a very successful replacement for the Hopper GPU because it builds upon GH100’s strengths and fills in the gaps or limitations, like improved communication bottlenecks for a larger number of racks. It has no Memory locality issues or cache storage issues.
This essentially means there are much lower bottlenecks when sharing data between servers, dramatically reducing the idle time for each GPU on each server. Thus, efficiency is boosted since more stuff gets done in the same clock cycles.
GB200 offers improved connectivity and data processing for AI tasks and is part of Nvidia’s “super chip” lineup, complementing its central processing unit, Grace.
Nvidia will definetly help AI to reach new heights. Click the link to visit and read the full blog.
- How Nvidia Blackwell Systems Attack 1 Trillion Parameter AI Modelswww.nextplatform.com How Nvidia Blackwell Systems Attack 1 Trillion Parameter AI Models
We like datacenter compute engines here at The Next Platform, but as the name implies, what we really like are platforms – how compute, storage,
- Nvidia turns up the AI heat with 1,200W Blackwell GPUswww.theregister.com Nvidia turns up the AI heat with 1,200W Blackwell GPUs
Five times the performance of the H100, but you'll need liquid cooling to tame the beast
- Authors file copyright lawsuit to torpedo Nvidia’s NeMowww.theregister.com Authors file copyright lawsuit to torpedo Nvidia’s NeMo
Claims allegedly pirated content from Books3 dataset trawled by its models
- Dell exec reveals Nvidia has a 1,000-watt GPU in the workswww.theregister.com Dell exec reveals Nvidia has a 1,000-watt GPU in the works
Hot and hungry, yet direct liquid cooling may not be required
- Nvidia posts revenue up 265% on booming AI businesswww.cnbc.com Nvidia posts revenue up 265% on booming AI business
Nvidia's total revenue rose 265% from a year ago, based on strong sales for AI chips for servers.
- The roast of Sam Altman: Nvidia CEO Jensen Huang's trillion dollar burnwww.digitimes.com The roast of Sam Altman: Nvidia CEO Jensen Huang's trillion dollar burn
Jensen Huang has thrown shade on Sam Altman's semiconductor ambitions.
- Startup retrofits Nvidia's GH200 into a €47.5K workstationwww.theregister.com Startup retrofits Nvidia's GH200 into a €47.5K workstation
Of course it's bristling with Noctuas – how else do you cool a 1kW desktop?
- Nvidia’s “Grace” Arm CPU Holds Its Own Against X86 For HPCwww.nextplatform.com Nvidia’s “Grace” Arm CPU Holds Its Own Against X86 For HPC
In many ways, the “Grace” CG100 server processor created by Nvidia – its first true server CPU and a very useful adjunct for extending the memory space of
- The Datacenter GPU Gravy Train That No One Will Derail - The Next Platformwww.nextplatform.com The Datacenter GPU Gravy Train That No One Will Derail - The Next Platform
We have five decades of very fine-grained analysis of CPU compute engines in the datacenter, and changes come at a steady but glacial pace when it comes
- Intel CEO: Nvidia Got 'Extraordinarily Lucky' in Dominating the AI Marketwww.extremetech.com Intel CEO: Nvidia Got 'Extraordinarily Lucky' in Dominating the AI Market
If it hadn't been for that meddling company, Intel would be rich by now!
- NVK holiday update: What we've achieved, and where we're headedwww.collabora.com NVK holiday update: What we've achieved, and where we're headed
A wrap up of NVK's progress in 2023 and what's to come for 2024.
- Nvidia's revenue triples as AI chip boom continueswww.cnbc.com Nvidia's revenue triples as AI chip boom continues
Nvidia called for accelerating revenue growth in the next quarter, even as executives warn that government limits on sales to some countries will hurt results.
- H100 Fading: Nvidia Touts 2024 Hardware with H200www.hpcwire.com H100 Fading: Nvidia Touts 2024 Hardware with H200
Supercomputing enthusiasts are speed demons, so it made sense for Nvidia to discuss its 2024 computing products at Supercomputing 2023. Nvidia’s next-generation products are being planned around H200 GPU, which […]
- NVIDIA TensorRT Extension for Stable Diffusion Performance Analysiswww.pugetsystems.com NVIDIA TensorRT Extension for Stable Diffusion Performance Analysis
NVIDIA has released a TensorRT extension for Stable Diffusion using Automatic 1111, promising significant performance gains. But does it work as advertised?
- Mellanox CEO's daughter killed in attack on Israelwww.theregister.com Mellanox CEO's daughter killed in attack on Israel
And engineer and girlfriend among those held hostage by Hamas
- Nvidia's yearly release cadence is trouble for Intel, AMDwww.theregister.com Nvidia's yearly release cadence is trouble for Intel, AMD
Or it could, just as soon as they figure out how to make the networking work
- Game Ready Driver 537.13 FAQ/Discussion
Game Ready Driver 537.13 has been released.
Article Here: https://www.nvidia.com/en-us/geforce/news/starfield-game-ready-driver/
Game Ready Driver Download Link: Link Here
New feature and fixes in driver 537.13:
Game Ready - This new Game Ready Driver provides the best gaming experience for the latest newgames including Starfield and the ICARUS: New Frontiers expansion.
Fixed Gaming Bugs
- [Battlefield 2042] Game stability can decrease when applying GeForce Experience Freestyle filters [4170804]
- [Ratchet & Clank: Rift Apart] Performance fluctuations due to issues between DirectStorage and NVIDIA Reflex [4212649]
Fixed General Bugs
- [Vegas Pro FX] Preview not working on some notebook configurations [3752618]
Open Issues
- [Halo Infinite] Significant performance drop is observed on Maxwell-based GPUs. [4052711]
- [DaVinci Resolve] This driver implements a fix for creative application stability issues seen during heavy memory usage. We’ve observed some situations where this fix has resulted in performance degradation when running DaVinci Resolve. This will be addressed in an upcoming driver release. [4172676]
Driver Downloads and Tools
Driver Download Page: Nvidia Download Page
Latest Game Ready Driver: 537.13 WHQL
Latest Studio Driver: 536.99 WHQL
DDU Download: Source 1 or Source 2
DDU Guide: Guide Here
DDU/WagnardSoft Patreon: Link Here
Documentation: Game Ready Driver 537.13 Release Notes | Studio Driver 536.99 Release Notes
Submit driver feedback directly to NVIDIA: Link Here
Having Issues with your driver? Read here!
Before you start - Make sure you Submit Feedback for your Nvidia Driver Issue
There is only one real way for any of these problems to get solved, and that’s if the Driver Team at Nvidia knows what those problems are. So in order for them to know what’s going on it would be good for any users who are having problems with the drivers to Submit Feedback to Nvidia. A guide to the information that is needed to submit feedback can be found here.
Additionally, if you see someone having the same issue you are having in this thread, reply and mention you are having the same issue. The more people that are affected by a particular bug, the higher the priority that bug will receive from NVIDIA!!
Common Troubleshooting Steps
- Be sure you are on the latest build of Windows 10 or 11
- Please visit the following link for DDU guide which contains full detailed information on how to do Fresh Driver Install.
- If your driver still crashes after DDU reinstall, try going to Go to Nvidia Control Panel -> Managed 3D Settings -> Power Management Mode: Prefer Maximum Performance
If it still crashes, we have a few other troubleshooting steps but this is fairly involved and you should not do it if you do not feel comfortable. Proceed below at your own risk:
- A lot of driver crashing is caused by Windows TDR issue. There is a huge post on GeForce forum about this here. This post dated back to 2009 (Thanks Microsoft) and it can affect both Nvidia and AMD cards.
- Unfortunately this issue can be caused by many different things so it’s difficult to pin down. However, editing the windows registry might solve the problem.
- Additionally, there is also a tool made by Wagnard (maker of DDU) that can be used to change this TDR value. Download here. Note that I have not personally tested this tool.
If you are still having issue at this point, visit GeForce Forum for support or contact your manufacturer for RMA.
Common Questions
- Is it safe to upgrade to ? Fact of the matter is that the result will differ person by person due to different configurations. The only way to know is to try it yourself. My rule of thumb is to wait a few days. If there’s no confirmed widespread issue, I would try the new driver.
Bear in mind that people who have no issues tend to not post on Reddit or forums. Unless there is significant coverage about specific driver issue, chances are they are fine. Try it yourself and you can always DDU and reinstall old driver if needed.
- My color is washed out after upgrading/installing driver. Help! Try going to the Nvidia Control Panel -> Change Resolution -> Scroll all the way down -> Output Dynamic Range = FULL.
- My game is stuttering when processing physics calculation Try going to the Nvidia Control Panel and to the Surround and PhysX settings and ensure the PhysX processor is set to your GPU
- What does the new Power Management option “Optimal Power” means? How does this differ from Adaptive? The new power management mode is related to what was said in the Geforce GTX 1080 keynote video. To further reduce power consumption while the computer is idle and nothing is changing on the screen, the driver will not make the GPU render a new frame; the driver will get the one (already rendered) frame from the framebuffer and output directly to monitor.
Remember, driver codes are extremely complex and there are billions of different possible configurations. The software will not be perfect and there will be issues for some people. For a more comprehensive list of open issues, please take a look at the Release Notes. Again, I encourage folks who installed the driver to post their experience here... good or bad.
Did you know NVIDIA has a Developer Program with 150+ free SDKs, state-of-the-art Deep Learning courses, certification, and access to expert help. Sound interesting? Learn more here.