Monday, February 3rd 2025

NVIDIA GeForce RTX 50 Series AI PCs Accelerate DeepSeek Reasoning Models
The recently released DeepSeek-R1 model family has brought a new wave of excitement to the AI community, allowing enthusiasts and developers to run state-of-the-art reasoning models with problem-solving, math and code capabilities, all from the privacy of local PCs. With up to 3,352 trillion operations per second of AI horsepower, NVIDIA GeForce RTX 50 Series GPUs can run the DeepSeek family of distilled models faster than anything on the PC market.
A New Class of Models That Reason
Reasoning models are a new class of large language models (LLMs) that spend more time on "thinking" and "reflecting" to work through complex problems, while describing the steps required to solve a task. The fundamental principle is that any problem can be solved with deep thought, reasoning and time, just like how humans tackle problems. By spending more time—and thus compute—on a problem, the LLM can yield better results. This phenomenon is known as test-time scaling, where a model dynamically allocates compute resources during inference to reason through problems. Reasoning models can enhance user experiences on PCs by deeply understanding a user's needs, taking actions on their behalf and allowing them to provide feedback on the model's thought process—unlocking agentic workflows for solving complex, multi-step tasks such as analyzing market research, performing complicated math problems, debugging code and more.The DeepSeek Difference
The DeepSeek-R1 family of distilled models is based on a large 671-billion-parameter mixture-of-experts (MoE) model. MoE models consist of multiple smaller expert models for solving complex problems. DeepSeek models further divide the work and assign subtasks to smaller sets of experts. DeepSeek employed a technique called distillation to build a family of six smaller student models—ranging from 1.5-70 billion parameters—from the large DeepSeek 671-billion-parameter model. The reasoning capabilities of the larger DeepSeek-R1 671-billion-parameter model were taught to the smaller Llama and Qwen student models, resulting in powerful, smaller reasoning models that run locally on RTX AI PCs with fast performance.
Peak Performance on RTX
Inference speed is critical for this new class of reasoning models. GeForce RTX 50 Series GPUs, built with dedicated fifth-generation Tensor Cores, are based on the same NVIDIA Blackwell GPU architecture that fuels world-leading AI innovation in the data center. RTX fully accelerates DeepSeek, offering maximum inference performance on PCs.
Throughput performance of the Deepseek-R1 distilled family of models across GPUs on the PC:Experience DeepSeek on RTX in Popular Tools
NVIDIA's RTX AI platform offers the broadest selection of AI tools, software development kits and models, opening access to the capabilities of DeepSeek-R1 on over 100 million NVIDIA RTX AI PCs worldwide, including those powered by GeForce RTX 50 Series GPUs. High-performance RTX GPUs make AI capabilities always available—even without an internet connection—and offer low latency and increased privacy because users don't have to upload sensitive materials or expose their queries to an online service.
Experience the power of DeepSeek-R1 and RTX AI PCs through a vast ecosystem of software, including Llama.cpp, Ollama, LM Studio, AnythingLLM, Jan.AI, GPT4All and OpenWebUI, for inference. Plus, use Unsloth to fine-tune the models with custom data.
Source:
NVIDIA
A New Class of Models That Reason
Reasoning models are a new class of large language models (LLMs) that spend more time on "thinking" and "reflecting" to work through complex problems, while describing the steps required to solve a task. The fundamental principle is that any problem can be solved with deep thought, reasoning and time, just like how humans tackle problems. By spending more time—and thus compute—on a problem, the LLM can yield better results. This phenomenon is known as test-time scaling, where a model dynamically allocates compute resources during inference to reason through problems. Reasoning models can enhance user experiences on PCs by deeply understanding a user's needs, taking actions on their behalf and allowing them to provide feedback on the model's thought process—unlocking agentic workflows for solving complex, multi-step tasks such as analyzing market research, performing complicated math problems, debugging code and more.The DeepSeek Difference
The DeepSeek-R1 family of distilled models is based on a large 671-billion-parameter mixture-of-experts (MoE) model. MoE models consist of multiple smaller expert models for solving complex problems. DeepSeek models further divide the work and assign subtasks to smaller sets of experts. DeepSeek employed a technique called distillation to build a family of six smaller student models—ranging from 1.5-70 billion parameters—from the large DeepSeek 671-billion-parameter model. The reasoning capabilities of the larger DeepSeek-R1 671-billion-parameter model were taught to the smaller Llama and Qwen student models, resulting in powerful, smaller reasoning models that run locally on RTX AI PCs with fast performance.
Peak Performance on RTX
Inference speed is critical for this new class of reasoning models. GeForce RTX 50 Series GPUs, built with dedicated fifth-generation Tensor Cores, are based on the same NVIDIA Blackwell GPU architecture that fuels world-leading AI innovation in the data center. RTX fully accelerates DeepSeek, offering maximum inference performance on PCs.
Throughput performance of the Deepseek-R1 distilled family of models across GPUs on the PC:Experience DeepSeek on RTX in Popular Tools
NVIDIA's RTX AI platform offers the broadest selection of AI tools, software development kits and models, opening access to the capabilities of DeepSeek-R1 on over 100 million NVIDIA RTX AI PCs worldwide, including those powered by GeForce RTX 50 Series GPUs. High-performance RTX GPUs make AI capabilities always available—even without an internet connection—and offer low latency and increased privacy because users don't have to upload sensitive materials or expose their queries to an online service.
Experience the power of DeepSeek-R1 and RTX AI PCs through a vast ecosystem of software, including Llama.cpp, Ollama, LM Studio, AnythingLLM, Jan.AI, GPT4All and OpenWebUI, for inference. Plus, use Unsloth to fine-tune the models with custom data.
24 Comments on NVIDIA GeForce RTX 50 Series AI PCs Accelerate DeepSeek Reasoning Models
Aren't the 50 series vaporware right now, why promote this if there is no way to (independently) do it?
Now, in the above benchmarks from Nvidia the Radeon card is running Vulkan. Is this optimum, or does Nvidia sabotaging the 7900 here?
Also even with the above results from Nvidia, 7900 wins on performance/dollar easily.
The bigger ones require more hardware.
No consumer hardware is going to run the actual big MoE model tho. Vulkan is quite a bit slower, but it's way easier to get up and running than ROCm.
Nonetheless, those results from AMD were really weird, as even a 3090 usually beats a 7900XTX:
source
DeepSeek is a open-source software which can be run anywhere in the world.
So many people still think that AI has to run in the 'cloud' because of the compute requirements, or require high end nvidia cards. Training an AI does, but most people aren't training AIs, they are just running a pre-trained model (inferencing). Any semi-recent GPU or CPU can do that for the small to mid-sized AI models. If you want to run the full 670b model then yeah you will need a high-end workstation (because of the ram requirements mainly), but a 14b distill can meet the majority of people's LlM needs and will run on consumer hardware.
If you were a billionare, would you invest billions into a business that has the potential to be free for all ?
In other words: any marketing that says, “deepseek runs on X consumer card” is a load of BS.
If that's the discussion in this thread, I assume the person is confused about where deepseek can be used.
2) If it is open source, is it the property of anybody?
So, again, if we are going to comment about 'great, China is going to scalp all our GPUs', why not post in the thread dedicated to that topic? Posting that in every thread is counter-productive.
The person I quoted clearly was clueless. No need to carry on with this devils advocate thing.
Carry on that sort on conversation here: www.techpowerup.com/forums/threads/us-investigates-possible-singapore-loophole-in-chinas-access-to-nvidia-gpus.331912/
Where it belongs.
If someone thinks that deepseek means that China will steal GPUs, they are clueless. Simple as that. China will steal GPUs for a variety of reasons, not for something that has already been publicly released and creates no income.
Let people call out nonsense without playing this weird devils advocate crap.