Aegis-GPU7.3 CPU versus GPU for temporal analytics
Why forecasting gets expensive and how cuML drops into skforecast.
Key points
Each extra model and each extra step multiplies the compute. GPUs run these many similar computations in parallel.
What NVIDIA says (1)
“With growing datasets and techniques like direct multi-step forecasting that require you to run several models at once, forecasts can quickly become computationally expensive when running on CPU-based infrastructure.”
cuML has a scikit-learn compatible API. Tools built on that API can use cuML models as drop-in replacements. API means application programming interface.
What NVIDIA says (2)
“Bringing accelerated computing to direct multistep forecasting cuML can be dropped into existing skforecast workflows.”
“NVIDIA cuML is a GPU-accelerated machine learning library for Python with a scikit-learn compatible API.”
Key terms
- Central processing unit: A computer's general-purpose processor, with a few fast cores that suit step-by-step work and small data.
Sample question
Why can time series forecasting become slow on CPUs?
Show the answer
Answer: Large datasets and methods like direct multi-step forecasting run many models at once
Each extra model and each extra step multiplies the compute. GPUs run these many similar computations in parallel.
What NVIDIA says (1)
“With growing datasets and techniques like direct multi-step forecasting that require you to run several models at once, forecasts can quickly become computationally expensive when running on CPU-based infrastructure.”
Practice 7.3 (2 questions) Full Advance Data Structures guide
← 7.2 Missing and irregular timestamps in cuDF · 7.4 Representing data as graphs →