CPU vs GPU: Performance & Power Analysis
March 2025Other
CPU vs GPU: Performance & Power Analysis is a research-based project that evaluates how CPUs and GPUs perform in matrix multiplication tasks. It focuses on execution time, power consumption, and scalability using parallel processing techniques like OpenMP (CPU) and CUDA (GPU). The project provides real experimental results, showing how GPUs outperform CPUs in large-scale computations while maintaining better energy efficiency.
A performance study comparing CPU and GPU efficiency in matrix multiplication, highlighting speed and power trade-offs in parallel computing.
Tech Stack
CC++OpenMPCUDAHWiNFO
Features
- Parallel matrix multiplication using OpenMP and CUDA
- Performance comparison across multiple matrix sizes
- Power consumption analysis using real metrics
- Visual results with graphs and benchmarks
Challenges
- System crashes due to heavy computational load
- Setting up CUDA and GPU environment correctly
- Managing threads and parallel execution on CPU
- Handling large data and memory efficiently
Learnings
- Learned parallel programming and multithreading concepts
- Gained hands-on experience with CUDA and GPU computing
- Understood real differences between CPU and GPU architectures
- Improved performance optimization and benchmarking skills