Activity
-
Elon Musk says SpaceX has committed to using Nvidia GPUs exclusively, calling them "the best" in a post on X. The statement reinforces Nvidia's…
Elon Musk says SpaceX has committed to using Nvidia GPUs exclusively, calling them "the best" in a post on X. The statement reinforces Nvidia's…
Liked by Wei Liu
-
10 million downloads! 🎉 An incredible milestone for NVIDIA Cosmos—and a testament to the team’s vision, hard work, and commitment to advancing…
10 million downloads! 🎉 An incredible milestone for NVIDIA Cosmos—and a testament to the team’s vision, hard work, and commitment to advancing…
Liked by Wei Liu
-
10x tokens per megawatt. First measured silicon numbers on NVIDIA Vera Rubin NVL72, from CoreWeave. Not projections — real results from live…
10x tokens per megawatt. First measured silicon numbers on NVIDIA Vera Rubin NVL72, from CoreWeave. Not projections — real results from live…
Shared by Wei Liu
Experience
-
-
-
-
-
Education
-
-
-
-
Publications
-
HiRe: Hint & Release for Efficient Synchronization of Speculative Threads
ACM/IEEE International Conference on Supercomputing
-
A Balanced Programming Model for Emerging Heterogeneous Multicore Systems
HotPar 2010: USENIX Workshop on Hot Topics in Parallelism
-
TAO - Two Level Atomicity for Dynamic Binary Optimizations
CGO
-
Pillar: A Parallel Implementation Language
LCPC 2007: International Workshop on Languages and Compilers for Parallel Computing
More activity by Wei
-
For all my Agents/Agent Creator connections out there(shouldnt that be all of you or will be all of you soon?), here is a great deep dive look on the…
For all my Agents/Agent Creator connections out there(shouldnt that be all of you or will be all of you soon?), here is a great deep dive look on the…
Liked by Wei Liu
-
Retirement begins today. I spent last week in Santa Clara, CA with many of my direct colleagues, finishing the trip by going to the World Cup Game…
Retirement begins today. I spent last week in Santa Clara, CA with many of my direct colleagues, finishing the trip by going to the World Cup Game…
Liked by Wei Liu
-
Today is, with mixed emotions, my last day at the most wonderful company in the world. Nearly 32 years ago, Jensen cold-called me to join his small…
Today is, with mixed emotions, my last day at the most wonderful company in the world. Nearly 32 years ago, Jensen cold-called me to join his small…
Liked by Wei Liu
-
I was in Sherman, Texas to celebrate the groundbreaking for Coherent’s new factory – where indium phosphide lasers will be built to connect hundreds…
I was in Sherman, Texas to celebrate the groundbreaking for Coherent’s new factory – where indium phosphide lasers will be built to connect hundreds…
Liked by Wei Liu
-
Re-posting Carole-Jean Wu’s wonderful summary of ISCA-53, which just concluded. It has been a great honor to serve our research community as program…
Re-posting Carole-Jean Wu’s wonderful summary of ISCA-53, which just concluded. It has been a great honor to serve our research community as program…
Liked by Wei Liu
-
It is a great honor to bring ISCA'26 to Raleigh. Big thanks to our industry sponsors, organization committee, steering committee, and student…
It is a great honor to bring ISCA'26 to Raleigh. Big thanks to our industry sponsors, organization committee, steering committee, and student…
Liked by Wei Liu
-
It was an honor to be inducted into the Fellow rank at the ACM Awards event last week in San Francisco. It was an even bigger honor to meet with many…
It was an honor to be inducted into the Fellow rank at the ACM Awards event last week in San Francisco. It was an even bigger honor to meet with many…
Liked by Wei Liu
-
Today marks my last day at NVIDIA. I have decided to take a career break to spend meaningful time with my aging parents in India. Throughout our…
Today marks my last day at NVIDIA. I have decided to take a career break to spend meaningful time with my aging parents in India. Throughout our…
Liked by Wei Liu
-
🚀 Major leap in Mixture-of-Experts (MoE) training! Our latest blog reveals how custom fused MLP kernels—built with CUTLASS—are eliminating memory…
🚀 Major leap in Mixture-of-Experts (MoE) training! Our latest blog reveals how custom fused MLP kernels—built with CUTLASS—are eliminating memory…
Liked by Wei Liu
-
Cosmos 3 is a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a…
Cosmos 3 is a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a…
Liked by Wei Liu
-
After ~20.6 years, it's apparently time for something else. This week I turned in my Intel materials as they requested – laptops, badges, etc. After…
After ~20.6 years, it's apparently time for something else. This week I turned in my Intel materials as they requested – laptops, badges, etc. After…
Liked by Wei Liu
-
I'm excited to announce Nemotron 3 Ultra, our largest and most capable open model to date. Nemotron 3 Ultra is a 550B Hybrid-MoE model with a 1M…
I'm excited to announce Nemotron 3 Ultra, our largest and most capable open model to date. Nemotron 3 Ultra is a 550B Hybrid-MoE model with a 1M…
Liked by Wei Liu
-
Poster session humming at #MLSys2026 tonight.
Poster session humming at #MLSys2026 tonight.
Liked by Wei Liu
-
We have multiple position opening at NVIDIA Cosmos, including - Research Scientist of different seniority. - System Software Engineer of different…
We have multiple position opening at NVIDIA Cosmos, including - Research Scientist of different seniority. - System Software Engineer of different…
Liked by Wei Liu
Other similar profiles
Explore more posts
-
Ramine Roane
Kicked off AMD DevDay 2025 with an inspiring & insightful fireside chat Ion Stoica, UC Berkeley Professor, Director of the Sky Computing Lab, co-creator of Spark, Ray, and vLLM… cofounder of LMArena, Anyscale, Databricks, Conviva… Fellow at both IEEE & ACM! Key takeaways: • LLMs are not so quadratic anymore. With all the tricks, they effectively scale quasi-linearly • did we run out of data? Synthetic data & CoT still boost intelligence (LLMs reached IMO gold level) • LLMs can generate optimized GPU kernels guided by simulated and realtime execution feedback #AMDDevDay #AI #AMD
1 Comment -
Milo Riano
🤖 AI is moving at lightning speed, and most people aren’t even paying attention. Jensen Huang, CEO at NVIDIA, just announced that their latest generation of processors is already in “full production,” signaling a major leap in AI hardware capacity. This means more powerful AI models are coming sooner than we think, and if you’re not prepared, you risk falling behind in this rapidly evolving game. This is a clear sign that the AI race is heating up, with big tech companies like NVIDIA pushing the boundaries of what’s possible. These processors power the AI that shapes industries, disrupts jobs, and creates new wealth opportunities. For entrepreneurs, investors, and everyday users, understanding how AI hardware advances can give you an edge in making smarter moves and generating income. 1 ⚡ AI hardware upgrades will fuel faster, more efficient AI tools that can transform how you work or create new revenue streams. 2 🔥 Companies racing to develop the most advanced processors will dominate future markets—timing is everything to stay ahead. 3 💰 Investing in AI-focused hardware or services now could unlock significant financial gains as AI adoption accelerates. If NVIDIA’s processors are already in full production, imagine what that means for the future of AI-powered income streams and cybersecurity defenses. Are you leveraging the latest tech, or are you still stuck using outdated tools? Try learning more about AI hardware developments and how they impact the tools you use daily. Test new AI platforms that utilize the latest processors for your business or projects. Build a plan to incorporate AI into your income streams to stay competitive. Audit your current cybersecurity measures to ensure they’re ready for increasingly sophisticated AI threats. Protect your digital assets by staying informed on hardware and software innovations that could threaten or boost your security. ✅ I understand how AI hardware advancements are shaping the future of tech and business. ✅ I can identify at least one way to leverage new AI processors for income or security. ✅ I am willing to test or learn about cutting-edge AI tools this week. In my view, ignoring this AI hardware revolution could leave you vulnerable or unprofitable as the big players take the lead—are you ready to adapt? #artificialintelligence #cybersecurity
-
Adnan Khaleel
TACC’s exploration of FP64 emulation on Horizon marks an interesting pivot in High Performance Computing. With hardware vendors aggressively optimizing for low-precision AI workloads (FP8/FP16), the density of native FP64 logic is relatively diminishing. By utilizing the Ozaki scheme to emulate double-precision on Blackwell’s tensor cores, TACC expects to outperform native FP64 execution by 2-3X. This effectively turns 'AI hardware' into high-fidelity scientific instruments, proving that software complexity can bridge the gap between GenAI infrastructure and rigorous physical simulation. https://lnkd.in/gb3-g238
4 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content