vkaku

17 ms·

271.
by vkaku 3y ago
A report of my experience after trying this: https://news.ycombinator.com/item?id=34662929
272.
by vkaku 3y ago
This theoretical expectation of such engineers is deeply flawed. - Every level balances between execution and communication. - Many companies fail due to not scoping and not collaborating enough. - Any engineer should follow good practices
273.
by vkaku 3y ago
I've dealt with this multiple times, so let me give my perspective. - It is hard for manufacturers to do this with small teams. Mostly because they do not always have good CI/CD or platforms available to keep being on top of vulne
274.
NopVer Specification(nopver.github.io)
2 points by vkaku 3y ago|1 comments
275.
by vkaku 3y ago
Software specifications are important. Or not, depending on how good your practice is. In case you do not want to use SemVer, or you hate it, here is a proposed alternative to it: https://nopver.github.io It is so easy to adopt,
276.
by vkaku 3y ago
Nice.
277.
by vkaku 3y ago
Moving the needle. But still as bloated as it can get. Never was invested in anything beyond Java SE, now I'm never going to invest in Java SE again.
278.
by vkaku 3y ago
Look at Microsoft's Antares. They use a wierd combination of LoadLibrary and WSL to run ROCM workloads. TL;DR not a good use of your time if AMD doesn't want to ship ROCM on Windows for all customers. Read this if you want to do s
279.
by vkaku 3y ago
What about /favicon.ico - or those pre-flight requests? The only real way to refuse requests is to just close the connection, no 404s or whatever should be sent. The deal with the way most people have designed the web from a client poi
280.
by vkaku 3y ago
Might as well release the Tupperware/Twine part and complete the picture :)
281.
by vkaku 3y ago
The client CPUs released until present do not have AVX-512 sadly. I'm not planning any hardware updates for the next 3 years at least. Not relevant for me at least.
282.
by vkaku 3y ago
Awesome. I just wrote about it above. Thanks for sharing!
283.
by vkaku 3y ago
Cloudflare could build a product that not only caches, but transfers S3 to R2 on usage. - Since non transactional data is likely to be on GitHub or somewhere, it should be easy to re-hydrate that on R2. - Once enough time passes, there coul
284.
by vkaku 3y ago
Answer is unconventional. Run CI/CD ... It's way easier to see if things will break live when you have it being run-tested on these stacks.
285.
by vkaku 3y ago
Agreed. What they've also been doing is stalling/removing support on gfx803 and that literally could have allowed people to still use the GPU for doing many decent small nets.
286.
by vkaku 3y ago
Beautiful! Waiting for someone to use this and get benchmarks with PyTorch now.
287.
by vkaku 4y ago
The issue I see is some companies' reluctance to even shift from Java 8 or move away from legacy. I guess that will be the biggest impediment to adoption of new features.
288.
by vkaku 4y ago
I write my private blog so I can reflect on the day and what matters. It helps me gather my thoughts well. This is just a private blog, hosted on my local PC ... I use this: https://github.com/guilt/mblog I do write ab
289.
by vkaku 4y ago
I was waiting for things to stabilize a bit on the software and kernel side, for the last 3 years. First was to ensure the Kernel driver would work, then ROCM would work, then something like Tensorflow would work. If ROCM wasn't going
290.
by vkaku 4y ago
I'm just beginning to think this is a terrible idea. A few years ago, I was like, Spring yeah, wow, complex is better but then Guice looked even better with less boilerplate. They just moved things from Java to XML and then back. That
291.
by vkaku 4y ago
This is great. Chimera Linux uses the FreeBSD userland version of the utilities. It's refreshing to see new distros taking charge of their development.
292.
by vkaku 4y ago
I like this idea. My developer machine until 2008 had only 512MB of RAM and for bloat I used to run Eclipse (by running JDT and some tweaks). I started off with a RIVA TNT2 (32MB) and ended up with a GeForce FX (256MB) at the end of 2008. A
293.
by vkaku 4y ago
Well, they are bad at advertising HIP support like the way NVIDIA does. But I'm on a 6900XT and I can run Pytorch nightly easy with it. No regrets. pip3 install --pre torch torchvision torchaudio --extra-index-url https://do
294.
by vkaku 4y ago
That they can't exhume the remains - random stuff like that on the Wikipedia article. If there is an actual burial of an alien, a government agency will likely cover it up.
295.
by vkaku 4y ago
Just when you think there's a real possibility of exhumation and further analysis there's known disinformation circulating. What I would really like to see is X Ray photography.
296.
by vkaku 4y ago
That doesn't account for physical disk characteristics at all. You could have disks from different generations all be part of a pool, CMR vs SMR(!) (needs zone resets and large chunk writes), disks could degrade etc. Plus, when you upd
297.
by vkaku 4y ago
I understand the first one may be a genuine pptimization needed. But for the second one: when a file is striped across different physical disks with possibly different characteristics, how would you defrag it without knowing the read access
298.
by vkaku 4y ago
I do want to train to become a plumber. It seems to pay well despite the amount of terrible things involved.
299.
by vkaku 4y ago
US copyright laws are a joke.
300.
by vkaku 4y ago
I-P-B. Some video formats only go I. Then there's not a lot different between images and video, as far as editing goes. Decoding for end user transportation has a lot more going on, but one has to start somewhere. Anyway - I think that

More ›