GeForce RTX 3060

[...] Yes, you can definitely build an entire fps game using WebGL for rendering. [...] My current WebGL project is rendering high definition terrain, high-poly animated models, thousands of particles, shaders, sound and more over 150 frames-per-second on a 10 year old PC with a RTX 3060. [...] Development is fast, re-loading is quick, errors and debugging is handled directly in the browsers which have great debug information and performance tracking. [...]

[...] Next, I found out that there is a powerPreference: 'high-performance' option. It can be enabled with chrome://flags/enable-webgpu-developer-features and I can now request a GPU with 'high-performance' power preference, although there is no way to check whether this actually selected the correct GPU, and it is useless to do it like that anyway, since no website actually sets this (experimental) option. Fortunately, there is the flag chrome://flags/force-high-performance-gpu which should solve this, but it is not available for my platform. I have a fairly standard RTX 3060, but apparently, the most common GPU (according to the steam hardware survey) is not supported. [...] work, but many other examples crash the browser. [...]

I have been using Qwen3.5-35B-A3B a lot in local testing, and it is by far the most capable model that could fit on my machine. [...] On an rtx 3060 12GB and 32GB of system ram, I went from slamming all my available memory to having like 14GB to spare.

[...] 8.6 min total on a single RTX 3060 Ti. [...] Hardware: one RTX 3060 Ti ($400). [...]

[...] I built to submit for Google Gemini Hackathon, and focused on an agent that would replace email content with regex to extract financial data. I used Gemini 3 Flash. [...] Both Ministral 3:3b and Qwen 3.5 Small:4b will fit on the smallest Mac Mini M4 or a RTX 3060 6GB (I have these devices). [...]

[...] The model trains in 15 minutes on an RTX 3060, the checkpoint is 22MB, and inference takes under a second. [...] THE APPROACH Instead of requiring gridded reanalysis, LILITH learns directly from sparse station observations: Transformer encoder on 30 days of historical data Autoregressive decoder for multi-day prediction Multi-timescale rollout: 6h steps for days 1-14, daily for 15-42, weekly for 43-90 Climate signal injection (ENSO, MJO) for extended range Total parameters: 1.87M. [...] The value is accessibility, not beating ECMWF. [...]

intel arc b580 (i think that's the latest one) isn't obnoxiously priced but you're going to have to face the fact that your PCIE is really very slow. [...] if you want to save even more money get the older Arc Battlemage GPUs. I used one it was comparable with an RTX 3060; i returned it because the machine i was running it in had a bug that was fixed 2 days before i returned it but i didn't know that. I was seriously considering getting a b580 or waiting until the b70 came out with more memory, although at this point i doubt it will be very affordable considering VRAM prices going up as well. [...]

[...] With GPU (tested on RTX 3060), inference runs at 30-50 tokens/sec. On CPU it drops to 5-10 tokens/sec, which is usable but noticeably slower for larger responses. [...] The lack of audio and the jumpy editing makes it hard to follow the actual workflow there are more videos on the website 5 in total I'll redo the demo with: - Narration or at least on-screen captions explaining each step - Slower pacing on the important bits (the parsing LLM rule generation flow) - A clear "here's the input here's what Syd does here's the output" structure - Maybe a side-by-side showing manual workflow vs. [...] what's planned) - Roadmap with realistic timelines - Maybe a shorter "state of the project" video or changelog - Clearer communication on what backers get and when you will recieve weekly or monthly updates and obviously ill answer any questions Current state for transparency: - Core features working: YARA, Nmap, Volatility, Metasploit, PCAP analysis with RAG-enhanced explanations - 356k chunk knowledge base indexed and searchable - Exploit/CVE database integrated - GUI and CLI both functional - Still refining: UX polish, additional tool integrations, documentatio

I wanted to dip my toe in the AI waters, so I bought a cheap Dell Precision 3620 Tower i7-7700, upgraded the RAM (sold what it came with on eBay) and ended up upgrading the power supply (this part wasn't planned) so I could install a RTX 3060 GPU. [...] The most useful thing I use it for is AI tagging and summaries for Karakeep, but I've also used it for a bunch of other applications including code I've written in Python to analyze driveway camera footage for delivery vehicles.

[...] The few times I’ve tried games on linux with decent hardware (eg radeon vii, rx580, rtx3060?) it was just trash. like straight up stuttering and nonsense. [...]

I have an RTX 3060 with 12GB VRAM. [...] If 14B doesn't answer correctly, I switch to Qwen 32B Q3KS, which will be slower because it needs to use the RAM. [...]

[...] I recently had to remove Windows completely from a few years old laptop with an 12th gen cpu and a Intel Iris / GeForce RTX 3060 Mobile combo because it was running very hot (90c+) and the fans were constantly running. [...]

[...] i’ve mainly tested it on MNIST for now, the CUDA backend trains one epoch in 0.4ms (batch size 1000, RTX 3060, as i mentioned in the post). It was primarily a deep dive into CUDA/C++, manual memory management, and building a dual backend architecture with a custom matrix lib (GPU-side completely from scratch). this was actually my 4th serious attempt at building a GPU-based MLP from scratch. [...] Turned out I was only updating the original weights matrix after backprop, but the transposed one was still holding stale values from earlier. this broke training completely, and I spent weeks trying to debug it, couldn’t figure it out until this current version. [...] when trying to cache 'smartly') -launching kernels in the right order while keeping data in sync -maintainingg modularity without sacrificing too much performance i avoided fused kernels/shared memory in this version to keep things clean and reusable, but now that the core works, I plan to start optimizing that layer too.

The most common GPU per the steam survey is the RTX 3060 at 170W TDP. [...] The SXM H100 has a 700W TDP, and will spend far more of its life at or near that value. [...] I think it will be a while before modern generative AI is even close to providing value in aggregate.

I use a Proxmox server with RTX 3060 to generate paintings (I have a couple of old jailbroken Amazon Kindle's attached to walls for that purpose), and to run ollama, which is connected to Home Assistant & their voice preview device, allowing me to talk with LLM without transmitting anything to cloud services. [...] It is pretty okay for telling me information, like temperature or value of some sensors I have connected to HA. [...]

[...] So a good CPU might read your question faster, but it will answer pretty much as slow as a cheap CPU with the same RAM. [...] CPU-only: 54.08 t/s prompt eval 2.69 t/s inference --- CPU + 52/65 layers offloaded to GPU (RTX 3060 12GB): 166.79 t/s prompt eval 6.62 t/s inference

For optimal performance, we recommend the sum of VRAM + RAM to be at least 80GB+. [...] My homesystem does run in a ryzen 5500 + 64gb RAM + 7x RTX 3060 12gb So 64gb RAM plus 84gb VRAM I dont want to brag around, but point to solutions for us tinkerers with a small budget and high energy costs. such system can be build for around 1600 euro. [...] I started with a AM4 Board (b450 Chipset) and one used RTX 3060 12gb which cost around 200 Euro used if you are patient. [...]

[...] Device 0 [NVIDIA GeForce RTX 3060] PCIe GEN 3@16x RX: 0.000 KiB/s TX: 55.66 MiB/s GPU 1837MHz MEM 7300MHz TEMP 43°C FAN 0% POW 43 / 170 W GPU[|| 5%] MEM[|||||||||||||||||||9.769Gi/12.000Gi] Device 1 [Tesla P40] PCIe GEN 3@16x RX: 977.5 MiB/s TX: 52.73 MiB/s GPU 1303MHz MEM 3615MHz TEMP 22°C FAN N/A% POW 50 / 250 W GPU[||| 9%] MEM[||||||||||||||||||18.888Gi/24.000Gi] Device 2 [Tesla P40] PCIe GEN 3@16x RX: 164.1 MiB/s TX: 310.5 MiB/s GPU 1303MHz MEM 3615MHz TEMP 32°C FAN N/A% POW 48 / 250 W GPU[|||| 11%] MEM[||||||||||||||||||18.966Gi/24.000Gi]

[...] The "Planner" is given your input question and the "scratchpad" and has to come up with DuckDuckGo search terms. Then the harness runs the DuckDuckGo search and gives the question, results, and scratchpad to the Synthesizer. The Synthesizer updates the scratchpad with new information that is learnt. This continues in a loop, with the Planner coming up with new search queries and the Synthesizer updating the scratchpad, until eventually the Planner decides to give a final answer, at which point the Finalizer summarises the information in a user-friendly final answer. [...] However I have found that the Synthesizer step is extremely slow on my RTX3060, and also I think it would cost me about £1/day extra to run the RTX3060 flat out vs idle. For the amount of work laconic can do in a day (not a lot!), I think I am better off just sending the money to OpenAI and getting the results more quickly. [...]

I used a similar budget and build something like this: 7x RTX 3060 - 12 GB which results in 84GB Vram AMD Ryzen 5 - 5500GT with 32GB Ram All in a 19-inch rack with a nice cooling solution and a beefy power supply. [...] (Added some 3d printed parts into the mix: if you think about building something similar) My power consumption is below 500 Watt at the wall, when using LLLMs,since I did some optimizations: Worked on power optimizations and after many weeks of benchmarking, the sweet spot on the RTX3060 12GB cards is a 105 Watt limit Created Patches for Ollama ( to group models to exactly memory allocation instead of spreading over all available GPUs (This also reduces the VRAM overhead) ensured that ASPM is used on all relevant PCI components (Powertop is your friend) It's not all shiny: I still use PCIe3 X1 for most of the cards, which limits their capability, but all I found so far (PCIe Gen4 x4 extender and bifurcation/special PCIE routers) are just too expensive to be used on such low powered cards Due to the slow PCIe bandwidth, the performance drops significantly Max VRAM per GPU is king. [...] Some RTX 3060 12GB Cards do idle at 11-15 Watt, which is unacceptable. [...]

On my RTX3060 - 12GB VRAM + 24GB RAM , with below command ollama run qwen3.8:27b --verbose "explain mmap”, I got 2.41 Tokens/s, Not sure if that can be improved considering VRAM doesn’t fit the entire, model. Additional details: total duration: 8m18.2870918s load duration: 612.105ms prompt eval count: 12 token(s) prompt eval duration: 2.900965s prompt eval rate: 4.14 tokens/s eval count: 1193 token(s) eval duration: 8m14.660618s eval rate: 2.41 tokens/s System spec: NVIDIA GeForce RTX3060 AMD Ryzen 5 1600 Six-Core B450 AORUS M Mother board. NVIDIA-SMI 620.02 Driver:620.02, CUDA Version: 13.2

Television HDR mode is set to FILMMAKER, OLED brightness 100%, Energy Saving Mode is off. [...] PC has Manjaro Linux with RTX 3060 12GB Graphic card driver: Nvidia 580.119.02 KDE Plasma Version 6.5.4 KDE Frameworks Version: 6.21.0 Qt Version: 6.10.1 Kernel Version 6.12.63-1-MANJARO Graphics Platform: Wayland Display Configuration High Dynamic Range: Enable HDR is checked There is a button for brightness calibration that I used for adjustment. [...] Brightness: 100% TV is reporting HDR signal. [...] The Blu-ray looks like it has the same brightness range but the color of the explosion (2:07:26) seems richer compared to the best playback on Linux (19s

I bought a new machine with an RTX 3060 Ti back in 2020 and it's still going strong, no reason to replace it.

I'm working with Unreal on PC, Engine Scalability Settings for game I've set it up like this now. I tried played build on my laptop (AMD Ryzen 7 4800H + NVIDIA GeForce RTX 3060), which is 4 years old, and the GPU showed only 15% usage. That's a bit encouraging.

I have an ASUS laptop that is borderline unusable in three years, the only saving grace is the RTX 3060 which I use for gaming and occasionally ollama.

My 6 year old mediocre gaming PC is still a mediocre gaming PC. I expect it to stay a mediocre gaming PC for the next 3 years maybe 5 years. 16GB RAM RTX 3060 TI (8 GB VRAM)

AMD Ryzen 7 5700G GeForce RTX 3060 8Gb 64 Gb of RAM It's about 3 years old, though it had 16GB of RAM at the time and I bought more after.

[...] And also what I was able to accomplish air-gapped with a Ryzen 7 and RTX 3060 12GB scared me. [...]

AI video generation can fairly easily choke anything that's not NVIDIA's flagship model. Even the latest local image gen models are so large that they can be frustratingly slow with non-optimal hardware even if they fit in the VRAM. IIRC when I had an M2, it was about 4x slower at running the venerable Stable Diffusion (and SDXL) than my meager RTX 3060.

I’ve paired my Z-Image Turbo with SeedVR2 upscale, running on a RTX3060 12gb, 32gb sysMEM, generates in 40sec. [...]

[...] It's, in my opinion, worth about $200, if it's relatively unused. As i mentioned, it's = RTX 3060 at least for compute loads, and the 16GB is nice to have for that. [...]

Data privacy and security don't matter? My secondhand RTX 3060 would buy a lot of cloud credits, but I don't want tons of highly personal data sent to the cloud. I can't imagine how it would be for healthcare and finance, at least if they properly shepherded their data.

my trusty RTX 3060 is gonna have its day in the sun... though I have run a bunch of 7B models fairly easily on Ollama.

I'll try it. It's cool that it's small and optimized to be a assistant, not sure it will work well with pi as a harness or if it really is only optimized with his harness, I'll find out, I only have a RTX3060 with 12 GB VRAM and would love to use it for assistant stuff.

context: I have had my computer off and power to it and the monitor off for fiveish days windows 11 23H2 nvidia geforce rtx 3060 game ready driver 566.36 (was an earlier version when lines started but updated to this and they are still there) have also restarted computer/monitor, reseated cables

[...] I have a M4 Mac Mini with 16GB unified memory and an RTX 3060 (Laptop) with 6GB VRAM. [...]

I guess you are doing offloading to system RAM? What tokens per second do you get? I've got an old gaming laptop with a RTX 3060, sounds like it could work well as a local inference server.

It's an NVIDIA GeForce RTX 3060 GPU. I know because I looked at my preferences in Blender, and that's what it says my laptop GPU is.

I have Interstellar on 4K UltraHD Blu-ray that features HDR on the cover, Sony 4K Blu-ray player (UBP-X700) and a LG G4 OLED television. I also have an AVR (Denon AVR-S760H 7.2 Ch) connecting both the Blu-ray and a PC running Linux with a RTX 3060 12GB graphic card to the television. [...]

[...] One is hard-put to buy a developer-power laptop with a sub-2K display these days, even in the Windows world I personally see a lot of 1080p screens on new gaming laptops too. [...] When I sold my RTX 3060 laptop with a 1080p screen, most buyers wanted it for professional work, according to them. I'm honestly not sure where all these hackernews commenters with low-dpi displays are coming from If anything, this is exactly the place where I'd expect a bunch of people to be rocking an older Thinkpad. [...]

My laptop has 2: Nvidia GeForce rtx 3060 laptop and Intel iris xe graphics

[...] Maybe VR doesn't need that many games because the small handful of good ones have so much depth and replay value. I guess I just talked myself into a $700 VR kit and possibly a $700 GPU upgrade, depending on whether or not my RTX 3060 is up to task.

[...] I got a new GeForce RTX 3060 last night. [...]

[...] I know I'd be interested in a lower powered card with a lot of vram just to get my foot in the door, that is why I bought a RTX 3060 12GB which is unimpressive for gaming but actually had the second most vram available in that generation. [...]

[...] People can play with "small" or "medium" models less powerfull and cheaper cards. A Nvidia Geforce RTX 3060 card with "only" 12Gb VRAM can be found around €200-250 on second hand market (and they are around 300350 new). [...]