I use my Threadripper for massive electromagnetic simulations for printed circuit boards and packages - I love it!
It's ideal for building big things, like Chromium or AOSP.
> not a GPU
Well, 3970x is able to soft-render DotA in 4k with 8-11 fps...
These things can chew through compiles so fast I/O to disk will be the bottleneck, rather than actual compilation time. Great for Rust hackers.
This is the most practical answer I've read thus far. However, I actually code via SSH into AWS EC2, so my local CPU is irrelevant for compiles.
Shoving somewhere in a datacenter and virtualizing so that 96 different clients can use it at once.
But ThreadRipper is a workstation-targeted product, not server, isn't it?
I like to use my lots-of-cores Xeons for media conversion and 3d rendering, it's nice to see that many cores spin up for SMP workloads.
I’m looking at one for research purposes. Agent based simulation, non-GPU-implementable Ml modeling, large scale data processing. There are a lot of tasks in my work that are greatly sped up by having tons of threads.
Developers will love it. With limited resources nobody tests their code for compatibility: good testing requires an order of magnitude more cores. Test against multiple competing libraries? Compiled with multiple competitive alternative dependencies? Maybe it is also important to test dependencies with various flags? Maybe test in multiple Linux distributions? And then cross-product multiply all possible combinations and get the best result.
Production 3D rendering is still done on CPU at VFX studios.
Exactly, no guarantees that your scene will even fit in GPU VRAM given the complexity of some VFX shots.
Allowing you to plug more GPUs into your motherboard.
After Effects and Ableton will love this one
I’m using it to still feel productive while working on our massive, TMP-heavy C++ codebase.