Learn to say popular GPU programming and compute ecosystem names correctly.
0 / 5 completed
1 / 5
How is CUDA (NVIDIA's parallel computing platform and programming model for GPUs) correctly pronounced?
CUDA is pronounced 'KOO-duh' — stress on KOO, two syllables. In a technical interview: "CUDA let us move the entire matrix multiplication onto the GPU and cut the training step time dramatically."
2 / 5
How is cuDNN (NVIDIA's GPU-accelerated library of primitives for deep neural networks) correctly pronounced?
cuDNN is pronounced 'KOO-DEE-EN-EN' — 'cu' (from CUDA) plus 'D-N-N' spoken as letters. In a technical interview: "cuDNN gave us a hand-tuned convolution kernel instead of the naive one our framework shipped by default."
3 / 5
How is Triton (open-source language and compiler for writing efficient GPU kernels) correctly pronounced?
Triton is pronounced 'TRY-tun' — stress on TRY, echoing the Greek sea god's name. In a technical interview: "Triton let us write a custom fused kernel in a few dozen lines of Python instead of hand-rolled CUDA C."
4 / 5
How is ROCm (AMD's open-source software stack for GPU computing) correctly pronounced?
ROCm is pronounced 'RAHK-em' — short for Radeon Open Compute, stress on RAHK. In a technical interview: "ROCm let us run the same PyTorch training script on an AMD card with almost no code changes."
5 / 5
How is oneAPI (Intel's open, cross-architecture programming model for CPUs, GPUs, and accelerators) correctly pronounced?
oneAPI is pronounced 'WUN-AY-PEE-EYE' — 'one' exactly like the number, plus 'A-P-I' spoken as letters. In a technical interview: "oneAPI let the same kernel source target both the CPU and the integrated GPU without a rewrite."