Optical computing startup Lightelligence has developed a new GPU that outstrips the very best GPUs on the planet -- at least in the most challenging mathematical problems -- with the introduction of the Photonic Arithmetic Computing Engine (PACE).
Lightelligence's new Photonic Arithmetic Computing Engine (PACE) thrashes NVIDIA's third-fastest Turing-based GPU in the GeForce RTX 3080 by nearly 100x in the NP-complete class of problems. The new PACE accelerator under the Ising model -- an example of a thermodynamic system that's used for understanding phase transitions -- with record-breaking results.
The new PACE accelerator when compared to the GeForce RTX 3080 is around 100x faster, with Lightelligence using 12,000 optical devices integrated onto a circuit, clocked at 1GHz. Up against the purpose-built Toshiba's simulated bifurcation machine based on FPGAs, Lightelligence's new PACE accelerator still beats that system by 25x -- just as impressive.
Lightelligence's new Photonic Arithmetic Computing Engine (PACE) system uses standard silicon photonics integration of Mach-Zehnder Interferometer (MZI) for computing and MEMS to change the waveguide shape in the MZI.
TechPowerUp makes a good point here, in that Lightelligence's new PACE accelerator shows the world that these super-advanced chips aren't all just about AI, HPC, ray tracing, and gaming. Massively complicated mathematical problems being super-sped up by 25-100x shouldn't be ignored, at all.
These types of accelerators can be used in material science, thermodynamics, bioinformatics, cryptography, circuit design, power grid optimization, and much more.