Viewing profile — blelbach

> I get the impression nvidia puts out a lot of their hardware supporting software themselves because they are hostile to open source community collaboration in general. This could…

I'm not sure how making a large contribution to the C++ Standard is a problem?

That involves a few diagrams, but essentially, we have two layers: - the libcu++ layer, which has some of our extensions and implementations specific to our platform. - the libc++ …

We change the mangling of all the symbols by changing the inline namespace that they are in, regardless of whether or not functional ABI breaks occurred. That's why it says the ABI…

Not your fault, we don't make it easy. The acroynms are terrible! That's why I typically spell out the full term. My first week at NVIDIA: Me, to very senior engineer: something so…

> I expect you don't want to have to recompile an entire c++ stdlib every time you recompile your own code. That's basically our current model, I discussed this on Twitter recently…

> It’s a gift. I wouldn't say it's a gift, though; it's part of what you pay for when you buy one of our products. Sure, it's not listed as a spec on the box, but users expect that…

You'll notice some people from NVIDIA are authors on those papers too! :)

Yep, that's correct. My team develops Thrust, CUB, and libcu++.

Since Unified Memory. UVA, or Unified Virtual Addressing, just ensured that a GPU-private object wouldn't have the same address as a CPU-private object.

The majority of software engineers at NVIDIA write software that runs on the host CPU. The majority of software written at NVIDIA (by any metric, lines of code, number of projects,…

> I did not say the library controls anything, Nvidia controls the library : its features, its roadmap, its bugs corrections, development efforts (people) etc. All these choices ar…

> A lot of hardware has builtin software, either inside a firmware or as a driver. Correct. > Keeping the software part in firmware lets customer free to use any kind of OS. Do you…

> To be clear, I don't think nvidia-paid developers should be able to write C++ Code for a nvidia-sold GPU. I'm not sure what you're saying here? You think another company or organ…

> IMHO, the question is not that we need code to run on CPUs and GPUs , we do need that, The question is whether the GPU seller has to control both sides. The question is not about…

> It is not their role to do that. You are incorrect. NVIDIA employs more software engineers than hardware engineers. > We need smarter and more shining GPUs from nvidia, not softw…

> apparently only a small bit of the library is actually implemented. Yep. It's an incremental project. But stay tuned. > I'm somewhat suspicious of the presumption of us using NVI…

> A pathetic attempt to lock developers into their hardware Ah-ha, you've caught us! Our plan is to lock you into our hardware by implementing Standard C++. Once you are all writin…

> What do you mean about running in the same memory space? Your operating system doesn’t allow that. Is your concern about using host memory? Actually, the basis of our modern GPU …

NVIDIA employs more software engineers than hardware engineers. > Why do they need to run software in the same memory space as my mail reader ? It is a lot more expensive to build …

"cu" is a pretty common prefix for CUDA libraries. cuBLAS, cuTENSOR, CUTLASS, CUB, etc. It gets worse if you try to spell libcu++ without pluses: libcuxx libcupp (I didn't hate thi…

Yep, this is a good summary (good enough that perhaps I should put something similar in the docs).

https://youtu.be/75LcDvlEIYw https://youtu.be/VogqOscJYvk

Today, you can use the library with NVCC, and the subset is small. We'll be focusing on expanding that subset over time. Our end goal is to enable the full C++ Standard Library. Th…

GPUs are parallel processors. So, yes, synchronization primitives are the highest priority. We focused on things that require /different/ implementations in host and device code. T…