Scale Lang

Part 4 of a series on 'why Spectral exists', 10 minute read. Part 3 argued that the CUDA programming model is more general than the hardware it grew up on, and that the gap between the hardware it can target and the hardware it was...

Part 3 of a series on 'why Spectral exists', 10 minute read. Part 2 argued that the compiler stack matters more in the agent era, not less, and that the substrate that agents write against should stay stable while the hardware varies....

This month's SCALE release is a meaty one. It pushes on three fronts at once: fleet deployability for cloud providers, raw HPC performance, and the PyTorch coverage that AI shops need to take AMD seriously. Head over to our install page to...

_Part 2 of a series on why Spectral and SCALE exists._ In Part 1, I argued that cross-vendor portability in accelerated computing must be delivered by a company, rather than a committee, because the implementation is the standard. A...

_Part 1 of a series on why Spectral and SCALE exists, see part 2._ The way we standardize accelerated computing is broken, and it has been broken for long enough that the workaround: single-vendor capture, is now the default. It is worth...

For nearly two decades, NVIDIA's proprietary programming model, CUDA, has been the "gold standard" for unlocking the capabilities of the GPU. In June 2024, Jensen Huang announced that the ecosystem has reached 5 million developers, growing...

Compilation of CUDA shuffles to use AMD's Data-Parallel Primitives (DPP) is a novel and unique compiler optimization in SCALE, first added in version 1.4.2. It's important to be able to execute shuffles efficiently as they are used in...

After our first GROMACS post back in May, we kept digging into the parts of the CUDA backend that GROMACS relies on, especially after getting a very helpful email from the GROMACS Team. Their feedback pointed out a few places where our...

A Big Update from Spectral Compute, the team behind SCALE. Hey World Yeah, we know, it's been a while. We've been heads-down, buried in code, and frankly, just incredibly busy making CUDA work everywhere. So, grab a coffee. We've got a lot...