NVLink, every mention
9 scenes · ← back to NVLink
tap a year for its mentions
every year anyone Kyle Kranen 3George Hotz 3Soumith Chintala 2Jonathan Frankle 2Varun Mohan 1Philip Kiely 1Eugene Cheah 1
Verbatim, from the transcripts: the passages where NVLink comes up
Next 100x in AI: Inference, Networking, & Self-Optimizing Models — Philip Kiely & Ali Taha, Baseten
- ▶ 51:29 Philip Kiely Uh, tensor parallelism is not a good fit for local AI because it assumes a very high bandwidth interconnects like NVLink, uh, was, you know,
Agent Inference at the "Speed of Light" — How NVIDIA moves like a $4.3 Trillion Startup
- ▶ 30:26 Kyle Kranen Let's say you're on an H 100, uh, the maximum NVLink domain, domain for H 100 for most DGX H 100 3 times in the scene
🕰️ The Oral History of Windsurf (ft. Varun Mohan, Scott Wu, Jeff Wang, Kevin Hou, Anshul R)
- ▶ 45:55 Varun Mohan On top of that, within the machine, they've created this, this fabric called NVLink that allows you to communicate at ultra low latency.
[LLM Paper Club] Llama 3.1 Paper: The Llama Family of Models
- ▶ 20:11 Eugene Cheah Yeah, so I'll, I'll draw the distinction between pipeline parallelism and techniques like deep, deep speed tree, which is essentially where the GPU has, uh, uh, NV-link connectivity to other, other GPUs to actually read the, the model…
State of the Art: Training 70B LLMs on 10,000 H100 clusters
- ▶ 42:20 Jonathan Frankle You know, for, for both the GH 200 and the GB 200 is this hybrid networking where you'll have blocks of NV link networked chips. 2 times in the scene
Supervise the Process of AI Research — with Jungwon Byun and Andreas Stuhlmüller of Elicit
- ▶ 52:39 unnamed speaker I don't know if that's like the best name to, to think about it, but I was using one of these chat which are documents features, and I put the AMD and my 300 specs on the, um, new, you know, Blackwell chips from NVIDIA, and I was asking… 2 times in the scene
Open Source AI is AI we can Trust — with Soumith Chintala of Meta AI
- ▶ 7:13 Soumith Chintala Um, similarly, GPU has several levels of memory, and then you have, like, different levels of network hierarchies, NVLink plus, like, uh, InfiniBand or Rocky or something like that, right?
- ▶ 19:16 Soumith Chintala So, I'm, like, I'm, like, I'm sure the other hardware providers are working on it, but.
Ep 18: Petaflops to the People — with George Hotz of tinycorp
- ▶ 41:53 George Hotz Um, so, the bandwidth is the, is roughly 10 X less than what you can get with NV-linked A-Hundreds. 3 times in the scene