Marek Kolodziej

Sr Staff Engineer, Google. He/Him

[Marek Kolodziej @marekinfo]

@deliprao Best question I’ve seen in a long time. As Nvidia’s chief scientist Bill Dally used to ask (paraphrasing): “What problem will you have if you succeed?”

[@[email protected] @CompileExplore]

We now support executing GPU code compiled with nvc++ - https://t.co/OxXiIvWV0K - we're not yet quite sure why it's so very slow but it's a start! -mg

[Marek Kolodziej @marekinfo]

@relistan Verticals seem to be noisier than EFHWs. Not surprising really, most man made noise is vertically polarized.

[Marek Kolodziej @marekinfo]

@relistan Yes the unun grounding helps, as does a CMC.

[Marek Kolodziej @marekinfo]

@relistan FTDX-10. Staying within 20-40W. Very nice on receive, but transmits the same as others.

[Marek Kolodziej @marekinfo]

@relistan That’s really impressive. Mine is maybe 12 ft off the ground throughout. 33ft would have been ideal. Haven’t found a way to do that yet. Verticals are likely harder (radials or counterpoises in a tiny backyard). I tried one and the EFHW is still much better.

[Marek Kolodziej @marekinfo]

@relistan Also if 20W, I gather you upgraded from the (tr)USDX. What’s your new rig?

[Marek Kolodziej @marekinfo]

@olafurw Undefined behavior?

[Marek Kolodziej @marekinfo]

@relistan @SDRadioClub Congrats! 73 de N1KOL

[Marek Kolodziej @marekinfo]

@olafurw Use int in C(++), target many architectures.

[Marek Kolodziej @marekinfo]

@olafurw gcc / clang -O3

[Marek Kolodziej @marekinfo]

@jakobo I’m so sorry for your loss! Here’s mine.

[Horace He @cHHillee]

Here's a recent cool effort that's helped to reduce PyTorch's "entropy" - cleaning up convolutions :P

Over the years, PyTorch has accrued ... many convolution implementations. In theory, that's not a problem. The issue is that it's leaked out into the actual operator set. (1/3)

[Bartosz Ciechanowski @BCiechanowski]

GPS – https://t.co/JkfJLVjRQs

[Sibin @sibinmohan]

I’ve been using #AnnotatedEquations in my recent papers. I think it really adds to the readability and understanding of the math.

Here are some examples. It uses #tikz in #latex.

Let me know if you like it. Happy for any feedback. [GitHub link: next tweet]

#AcademicChatter

+

[NASA Webb Telescope @NASAWebb]

Sorry for the shade, @NASASun, @NASAMoon & @NASAEarth! But you're just not cool enough 😎

Now fully deployed, our sunshield blocks your heat & light so that our mirrors will be able to detect faint heat signals in the universe: https://t.co/mggJkh6swO #UnfoldTheUniverse

[Nadav Rotem @nadavrot]

I spent some time optimizing memset and memcpy in x86 assembly. Here are a few interesting things about memset and memcpy. 1/

[PyTorch @PyTorch]

PyTorch 1.10 is here!

Highlights include updates for:
- CUDA Graphs APIs updates
- Several frontend APIs moved to Stable
- Automatic fusion in JIT Compiler support for CPU/GPUs
- Android NNAPI now in beta

Blog: https://t.co/jMRDpA6T22
Release: https://t.co/QuaxJA2xfh

[Marek Kolodziej @marekinfo]

In addition to full-time engineers, my team is also looking for 2022 GPU algorithm summer interns! Please check out the opportunity! https://t.co/lRGTEM22ip