Home > Servers > Rack and Tower Servers > Intel > White Papers > Performance comparison of Dell EMC PowerEdge T350 and T340 server models > Comparison results and technical highlights
We tested and compared the performance regarding inventory management in handling a different amount of store keeping units (SKUs). See the performance comparison below.
On average, Dell EMC PowerEdge T350 uses 5.5% less CPU than T340.
On average, Dell EMC PowerEdge T350 writes to SSD 8.2% faster than T340.
On average, Dell EMC PowerEdge T350 uses 17.7% less time to train the ML model than T340.
For the inherent performance testing, we implemented 3 models:
During the testing, NBeats model has shown the best results and is used by default.
NBeats is a block-based deep neural architecture for univariate time series point forecasting that is similar in its philosophy to very deep models (for example, ResNet) used in more common deep learning applications such as image recognition.
This model takes into account long-term trends and seasonality. NBeats was applied to M3 and M4 datasets from Kaggle competitions. In each case, it beats the accuracy of existing models that combine ML and statistical approaches on common datasets. For the current generated dataset of 25 SKUs, this algorithm shows SMAPE equals 62.4%.
Learn more about NBeats model at N-BEATS: NEURAL BASIS EXPANSION ANALYSIS FOR INTERPRETABLE TIME SERIES FORECASTING.