Introduction¶

1. Which of the following programming models are covered in this course for GPU computing?

• A. CUDA, OpenCL, HIP
• B. OpenACC, OpenMP Offloading, HIP
• C. OpenACC, CUDA, OpenCL
• D. OpenCL, HIP, OpenMP Offloading

2. True or False: OpenACC and OpenMP Offloading are directive-based programming models that make it easier to develop applications for both CPUs and GPUs.

3. Which of the following statements best describes HIP?

• A. A programming model exclusively for AMD GPUs
• B. A CUDA-compatible framework for multi-GPU architectures
• C. A programming framework designed for CPU-only computations
• D. A directive-based language only for NVIDIA GPUs

4. True or False: GPGPU programming is well-suited for tasks that require high memory bandwidth and massive parallelism.

5. GPUs are particularly effective for which types of applications?

• A. Word processing and spreadsheet applications
• B. Scientific computing, machine learning, and video rendering
• C. Simple data entry tasks
• D. Basic image editing

Accelerators in Exascale and Pre-Exascale Supercomputers¶

1. Which supercomputer currently holds the top position in the TOP500 list as of June 2024?

• A. Eagle
• B. Fugaku
• C. Frontier
• D. Aurora

2. True or False: All top four supercomputers in the TOP500 list as of June 2024 include GPU accelerators.

3. What type of GPUs does the Frontier supercomputer use?

• A. Intel Data Center GPU Max Series
• B. NVIDIA H100
• C. AMD Instinct MI250X
• D. Fujitsu A64FX

4. Which of the following is NOT true about heterogeneous computing in high-performance computing (HPC)?

• A. It involves using both CPUs and GPUs for computation.
• B. It is necessary for fully utilizing modern supercomputer hardware.
• C. It relies solely on CPU resources.
• D. It enables efficient use of GPUs in scientific applications.

5. True or False: Aurora, one of the top supercomputers, utilizes Intel CPUs and Data Center GPUs.

The Role of HPC in Science, Engineering, and AI¶

1. Which types of equations are commonly solved in scientific computing and engineering?

• A. Algebraic equations
• B. Partial Differential Equations (PDEs)
• C. Logical equations
• D. Trigonometric equations

2. True or False: GPUs are generally less effective than CPUs for tasks involving a large number of arithmetic operations.

3. In the context of scientific computing, which numerical methods are often used to solve systems of equations?

• A. Fourier Transform and Laplace Transform
• B. Finite Difference and Finite Element methods
• C. Machine Learning and Deep Learning
• D. Matrix Inversion and Determinant Calculation

4. What is the main benefit of GPUs in artificial intelligence and scientific computing?

• A. They are cost-effective compared to CPUs.
• B. They are faster for simple arithmetic calculations.
• C. They provide high-throughput and parallel processing power.
• D. They consume less energy.

5. True or False: Computational workloads in AI and scientific computing can be managed easily without the use of HPC resources.