#rtx3090 — Public Fediverse posts

RT @cheenanet: Ist die Ära da, in der Qwen3.6-27B mit nur einer RTX 3090 120 bis 200 Token pro Sekunde verarbeitet?

mehr auf Arint.info

#AI #GitHub #LLM #MachineLearning #Qwen3 #RTX3090 #arint_info

https://x.com/cheenanet/status/2053134362749927641#m

RT @cheenanet: Ist die Ära da, in der Qwen3.6-27B mit nur einer RTX 3090 120 bis 200 Token pro Sekunde verarbeitet?

mehr auf Arint.info

#AI #GitHub #LLM #MachineLearning #Qwen3 #RTX3090 #arint_info

https://x.com/cheenanet/status/2053134362749927641#m

RT @cheenanet: Ist die Ära da, in der Qwen3.6-27B mit nur einer RTX 3090 120 bis 200 Token pro Sekunde verarbeitet?

mehr auf Arint.info

#AI #GitHub #LLM #MachineLearning #Qwen3 #RTX3090 #arint_info

https://x.com/cheenanet/status/2053134362749927641#m

RT @cheenanet: Ist die Ära da, in der Qwen3.6-27B mit nur einer RTX 3090 120 bis 200 Token pro Sekunde verarbeitet?

mehr auf Arint.info

#AI #GitHub #LLM #MachineLearning #Qwen3 #RTX3090 #arint_info

https://x.com/cheenanet/status/2053134362749927641#m

RT @cheenanet: Ist die Ära da, in der Qwen3.6-27B mit nur einer RTX 3090 120 bis 200 Token pro Sekunde verarbeitet?

mehr auf Arint.info

#AI #GitHub #LLM #MachineLearning #Qwen3 #RTX3090 #arint_info

https://x.com/cheenanet/status/2053134362749927641#m

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

https://github.com/Luce-Org/lucebox-hub

#HackerNews #Qwen3.5 #RTX3090 #tok/s #machinelearning #AIperformance

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

https://github.com/Luce-Org/lucebox-hub

#HackerNews #Qwen3.5 #RTX3090 #tok/s #machinelearning #AIperformance

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

https://github.com/Luce-Org/lucebox-hub

#HackerNews #Qwen3.5 #RTX3090 #tok/s #machinelearning #AIperformance

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

https://github.com/Luce-Org/lucebox-hub

#HackerNews #Qwen3.5 #RTX3090 #tok/s #machinelearning #AIperformance

We got 207 tok/s with Qwen3.5-27B on an RTX 3090

https://github.com/Luce-Org/lucebox-hub

#HackerNews #Qwen3.5 #RTX3090 #tok/s #machinelearning #AIperformance

🚀🤡 In what sounds like a fever dream concocted by a caffeinated coder, somebody managed to finagle a Llama 70B model onto an RTX 3090 using an NVMe-to-GPU #magic #trick. Meanwhile, #GitHub buzzwords are flying like confetti at a tech bro's birthday party. But hey, at least we can always count on these geniuses to overcomplicate the simple. 🧐💻
https://github.com/xaskasdf/ntransformer #Llama70B #RTX3090 #NVMe #techhumor #HackerNews #ngated

Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

https://github.com/xaskasdf/ntransformer

#HackerNews #Llama3.1 #RTX3090 #NVMe #GPU #bypass #CPU #AItechnology

Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

https://github.com/xaskasdf/ntransformer

#HackerNews #Llama3.1 #RTX3090 #NVMe #GPU #bypass #CPU #AItechnology

Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

https://github.com/xaskasdf/ntransformer

#HackerNews #Llama3.1 #RTX3090 #NVMe #GPU #bypass #CPU #AItechnology

Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

https://github.com/xaskasdf/ntransformer

#HackerNews #Llama3.1 #RTX3090 #NVMe #GPU #bypass #CPU #AItechnology

Bài đăng: Đang xây dựng máy chủ 8 GPU, đang cân nhắc giữa Radeon R9700 (VRAM lớn, giá rẻ) và RTX 3090 (CUDA, hiệu năng cao) cho công việc xử lý AI. Lo lắng về tình trạng GPU đã qua sử dụng RTX 3090 và khả năng nguồn ROCm. Câu hỏi: ai đã test 2 card này? Nên chọn VRAM+giá rẻ hay CUDA+hiệu năng?
#GPU #AI #R9700 #RTX3090 #OC #CôngNghệ #MáyHọc #NewsVie

https://www.reddit.com/r/LocalLLaMA/comments/1qldsdx/8_radeon_r9700s_vs_8_rtx_3090_2_slot_blower_style/

Tự chế PC desktop tối đa: case Fractal Define 7, 3 x RTX 3090 + RTX 3060, 128 GB DDR4, PSU 1500 W. Nhiệt độ ổn, VRAM 72 GB cho các mô hình LLM lớn (gpt‑oss‑120b, glm‑4.5‑air, qwen3‑vl‑235b, …). 3060 trên riser dùng cho tạo ảnh & TTS. Setup mạnh mẽ cho AI, thích hợp cho người đam mê phần cứng. #AI #Hardware #PCBuild #RTX3090 #RTX3060 #Vietnam #CôngNghệ

https://www.reddit.com/r/LocalLLaMA/comments/1qgx83t/3x3090_3060_in_a_mid_tower_case/

🗣️ Muốn xây dựng trợ lý AI offline “Jarvis” với độ trễ TTFT (Time To First Token) thấp? 💡 So sánh RTX 3090 (≈700 €) vs RTX 4090 (≈1700 €):
- 4090 mạnh hơn, nhưng TTFT thường bị chi phối bởi tải mô hình, khởi động kernel, và giao tiếp CPU‑GPU → lợi thế thời gian phản hồi thực tế có thể không quá lớn.
- Nếu ngân sách hạn chế và muốn mở rộng nhiều GPU, 3090 vẫn đủ cho mô hình 7‑13B.
#AI #LLM #RTX3090 #RTX4090 #Latency #CôngNghệ #TríTuệNhânTạo #AIOffline

https://www.reddit.com/r/LocalLLaMA/

Một người dùng đang cân nhắc nâng cấp từ RTX 4080 lên 2x RTX 3090 cũ (khoảng $800) để chạy các mô hình ngôn ngữ lớn (LLM) 70B cục bộ. Các mô hình 14B hiện tại không đủ mạnh, đặc biệt với tiếng Hàn. Người này lo ngại liệu cấu hình 2x 3090 có bền vững lâu dài khi công nghệ đang chuyển dịch sang bộ nhớ hợp nhất.
#LLM #AI #Hardware #RTX3090 #RTX4080 #DeepLearning #TríTuệNhânTạo #PhầnCứng #HọcSâu

https://www.reddit.com/r/LocalLLaMA/comments/1pvacv8/thoughts_on_picking_up_dual_rtx_3090s_at_this/

Oh wow, yet another blog post in the series that refuses to die 🙄. Because who doesn't love the idea of using an RTX 3090 to train a base model for what feels like the 3090th time? 😂 Spoiler alert: it still won't cook your breakfast. 🍳🤦‍♂️
https://www.gilesthomas.com/2025/12/llm-from-scratch-28-training-a-base-model-from-scratch #blogpost #RTX3090 #AItraining #techhumor #machinelearning #gaming #HackerNews #ngated

LLM from scratch, part 28 – training a base model from scratch on an RTX 3090

https://www.gilesthomas.com/2025/12/llm-from-scratch-28-training-a-base-model-from-scratch

#HackerNews #LLMfromScratch #RTX3090 #BaseModel #AITraining #MachineLearning

Black Friday & Cyber Monday - Angebot!

https://www.ebay.de/itm/257219114144

Tauchen Sie ein in die Welt der High-End-Gaming mit der NVIDIA GeForce RTX 3090 Founders 24GB GDDR6X Grafikkarte, einer wahnsinnigen Leistungsmaschine.
#geforce #rtx3090 #wasserkühlung

Cài đặt 4x RTX 3090 cho mô hình Wan2.2-TI2V-5B gặp lỗi VRAM OOM trong bước decode. #Wan2.2 #RTX3090 #VRAM #OOM #CUDA #PyTorch #MachineLearning #AI #GPUs #Docker

https://www.reddit.com/r/LocalLLaMA/comments/1ojdx3y/4x_rtx_3090_setup_for_wan22ti2v5b_fp16/

L.SpringBootApplication xây dựng cluster GPU: Tư vấn thêm card GPU RTX 3090 cho hệ thống hiện tại. Hệ thống hiện tại: Ryzen 5950X, bo mạch chủ X570 Dark Hero III, dual RTX 3090 NVLink.dereferenceLOCAL Cần thêm 2 card RTX 3090 để tạo cluster 4 GPU. Các lựa chọn: sử dụng card bifurcation PCIe và Oculink, lo lắng về công suất nguồn và hiệu suất. Cân nhắc xây dựng hệ thống mới cho tương lai. #GPU #PCBuilding #RTX3090 #NVLink #Hardware

https://www.reddit.com/r/LocalLLaMA/comments/1o1smha/advice_for_

Người dùng đạt tốc độ 70 t/s khi chạy mô hình Qwen3-Next-80B-A3B-Instruct-exl3 4.06bpw trên 2 GPU RTX 3090! Bản cập nhật exl3 0.0.7 mang lại cải thiện đáng kể về hiệu suất, duy trì tốc độ ổn định tới 14k context. Ấn tượng với khả năng xử lý LLM trên phần cứng cá nhân.
#LLM #AI #Qwen3Next #LocalLLaMA #RTX3090 #DeepLearning #HiệuNăngAI #GPU #MáyHọc #MôHìnhNgônNgữ

https://www.reddit.com/r/LocalLLaMA/comments/1nyb1x2/getting_70_ts_on_qwen3next80ba3binstructexl3/

YouTuber Receives Four NVIDIA RTX 4090 Cards; Discover Three Are Fake #fakertx4090 #graphics #hardware #rtx3080 #rtx3090 #rtx4090 #scammers

https://www.lowyat.net/2025/355942/youtuber-rtx-4090-fake-cards/

YouTuber Receives Four NVIDIA RTX 4090 Cards; Discover Three Are Fake #fakertx4090 #graphics #hardware #rtx3080 #rtx3090 #rtx4090 #scammers

https://www.lowyat.net/2025/355942/youtuber-rtx-4090-fake-cards/

YouTuber Receives Four NVIDIA RTX 4090 Cards; Discover Three Are Fake #fakertx4090 #graphics #hardware #rtx3080 #rtx3090 #rtx4090 #scammers

https://www.lowyat.net/2025/355942/youtuber-rtx-4090-fake-cards/

YouTuber Receives Four NVIDIA RTX 4090 Cards; Discover Three Are Fake #fakertx4090 #graphics #hardware #rtx3080 #rtx3090 #rtx4090 #scammers

https://www.lowyat.net/2025/355942/youtuber-rtx-4090-fake-cards/

YouTuber Receives Four NVIDIA RTX 4090 Cards; Discover Three Are Fake #fakertx4090 #graphics #hardware #rtx3080 #rtx3090 #rtx4090 #scammers

https://www.lowyat.net/2025/355942/youtuber-rtx-4090-fake-cards/

Google Releases Gemma 3 QAT AI Models for Consumer GPUs

#AI #AIModels #GoogleAI #Gemma3 #LLM #OpenSourceAI #GPUs #QAT #Quantization #DeepLearning #MachineLearning #NVIDIA #RTX3090 #Kaggle

https://winbuzzer.com/2025/04/20/google-releases-gemma-3-qat-ai-models-for-consumer-gpus-xcxwbn/

Scammers Are Now Trying To Pass Off RTX 3090 Cards As RTX 4090 #fakertx4090 #graphics #hardware #nvidia #rtx3090 #scammers

https://www.lowyat.net/2025/346671/scammers-are-now-trying-to-pass-off-rtx-3090-cards-as-rtx-4090/

Scammers Are Now Trying To Pass Off RTX 3090 Cards As RTX 4090 #fakertx4090 #graphics #hardware #nvidia #rtx3090 #scammers

https://www.lowyat.net/2025/346671/scammers-are-now-trying-to-pass-off-rtx-3090-cards-as-rtx-4090/

Scammers Are Now Trying To Pass Off RTX 3090 Cards As RTX 4090 #fakertx4090 #graphics #hardware #nvidia #rtx3090 #scammers

https://www.lowyat.net/2025/346671/scammers-are-now-trying-to-pass-off-rtx-3090-cards-as-rtx-4090/

Scammers Are Now Trying To Pass Off RTX 3090 Cards As RTX 4090 #fakertx4090 #graphics #hardware #nvidia #rtx3090 #scammers

https://www.lowyat.net/2025/346671/scammers-are-now-trying-to-pass-off-rtx-3090-cards-as-rtx-4090/

Scammers Are Now Trying To Pass Off RTX 3090 Cards As RTX 4090 #fakertx4090 #graphics #hardware #nvidia #rtx3090 #scammers

https://www.lowyat.net/2025/346671/scammers-are-now-trying-to-pass-off-rtx-3090-cards-as-rtx-4090/

Mit etwas Glück könnte Frame Generation bald auch auf der #GeForce #RTX3090 und den anderen #GPUs der 3000-Serie von #Nvidia landen. Aussagen des Unternehmens im Rahmen eines Interviews machen jetzt Hoffnung. https://winfuture.de/news,148274.html?utm_source=Mastodon&utm_medium=ManualStatus&utm_campaign=SocialMedia

#Benchmarks show even an old #Nvidia #RTX3090 is enough to serve #LLM to thousands
For 100 concurrent users, the card delivered 12.88 tokens per second—just slightly faster than average human reading speed
https://www.theregister.com/2024/08/23/3090_ai_benchmark/ #AI #ML

This is still the best looking graphics card ever made.

Show me one better. I’ll wait.

#nvidia #rtx #rtx3090 #amd #graphicscard

Tom Clancy's Ghost Recon Breakpoint (Ubuntu 23.04 Proton-Experimental) [2160p] {RTX3090}

https://www.youtube.com/watch?v=OwGTurox2UM

#GamingOnLinux #Gaming #Linux #RTX3090

Forza Horizon 5 (Ubuntu 23.04 Proton 8.0-2) [2160p] {RTX3090}

https://www.youtube.com/watch?v=RyV3CHGxxfk

#GamingOnLinux #Gaming #Linux #RTX3090

Mythforce Demo (Ubuntu 23.04 Proton 8.0-2) [2160p] {RTX3090}

https://www.youtube.com/watch?v=62UpSrmWHJE

#GamingOnLinux #Gaming #Linux #RTX3090

I’m building a high-end #gaming PC. The #CPU is #intel core #i912900k, on #ASUS ROG MAXIMUS HERO motherboard.
But when it comes to the #GPU I have two ideas, either I go with one of #RTX3090 / #RTX3080 Ti / #rx6900xt or wait the new #Nvidia & #AMD series.

It should be built by 2022/2/15.

#radeon #pcbuild #pcbuilder

$340,000 of Nvidia RTX 3090 graphics cards were stolen in China - Enlarge / The GPU Grinch doesn't care about your lists or whether you've been naughty or nice. (cre... - https://arstechnica.com/?p=1728217 #rtx3090 #nvidia #theft #tech #gpu

#RTX3090 is ridiculously expensive...

https://blog.seedly.sg/rtx-3080-3090-singapore-price/

Também quero... 😍 #rtx3090 @NVIDIA @NVIDIAGeForce

Технологии свернули куда-то не туда

#rtx3090