looterz

First thing that came to my mind as well, I honestly forgot it was a feature.

What do I have to do to convince one of you masochists with a 3900x to compile UE4 in Visual Studio 2019 setup to use 11 cores? Nobody ever posts large codebase benchmarks.

Looking at cooling solutions for a new 3900x build. Think the Noctua NH-U12A is still one of the champs? I went over to AIOs awhile back and while the temps have been good overall I had a pump die on one and the noise was a bit high, came across LTT's "Why you shouldn't water cool your PC" video again and I think it's time to switch back.

That's definitely a thing, I was trying to address the common misconception that devs go out of their way to favor one vendor over another, in general, developers want their products to run the best it can on any hardware available. One of the best examples I have worked with related to what you mentioned is the AMD Bulldozer bugs related to audio, specifically with unreals audio subsystem killing the CPU performance. Still an issue to this day, despite several attempted fixes from Epic and AMD.

The only real time this might occur is when games use middleware created for and tested on NVidia hardware primarily (Gameworks), or integrate intel's "fast-math" libraries for code (I've never seen this personally). Games favoring certain GPUs or CPUs past that are likely just the hardware and not intentional.

Matrixes, Vectors, Angles and hexadecimal are fairly common, as well as basic algebra and curves, interpolation, etc. Nothing too crazy really. Most universities will teach you different algorithms for traversing trees of data or building them as well, as they come up on programming tests at big companies often.

The blade stealth is pricey but not bad for the specs, and works for programming pretty well (can even get linux running without too much trouble nowdays). The most important specs for programming will be CPU & Ram, I would also recommend using Visual Studio Code as it's very light on resources in comparison to other IDEs and can provide the functionality you need with plugins.

If some pre-processing time is an option for getting both files into a workable format, something like a bloom-filter may work, if that's usable I'd recommend it over a plain hashmap but if not something like Redis may work. Parsing out the files into data in memory for fast lookups is your main challenge here. I would also recommend Rust or C++ for this.

deluge is the best client I have used so far, and even has a very useful command-line client you can use via ssh on a vps to avoid any trouble from your ISP if you have the need to be careful. uTorrent is trash, they've shipped malware themselves and accidentally through ads before, have been bought by sketchy companies a few times the past decade and have caved in to MPAA pressure, and more than likely sell your usage data to any buyer with enough money.

Do yourself a favor, if you're learning C++ become familiar with the entire process of object file -> linker -> compiler using gcc or clang (works for windows now too) before you start using Visual Studio to hold your hand, or you'll just become confused and frustrated from not understanding VS's error output when you run into problems working with external libraries. If you hate VS just grab clang and work with makefiles or batch/terminal for building and Atom with the clang plugin, the workflow is surprisingly smooth. If you want a posix like work environment on windows grab babun and TDM for now, supposedly Windows 10 will be getting a big update this summer to support a posix environment within windows.