Puzzle
Across
- 2. mechanism to data ensure reliability through redundancy and protection in storage unit
- 5. benchmark suite designed to measure the memory bandwidth performance
- 6. __________ partitioning is used to minimize contention and maximize resource utilization in software lockout problem
- 9. Intermediate memories that are small and fast
- 11. computing node
- 12. _______ scaling evaluates the efficiency of parallelization by scaling problem size proportionally as the number of computing node increases
- 15. ______ parallelism that reduces runtime by parallel performance tuning
- 16. ________ simulation is an example of embarrassingly parallel supercomputing problem
- 18. instructions that reduces runtime by serial performance tuning
- 19. programming model that offloads computationally intensive tasks for GPU
- 21. ________ helps in providing high bandwidth and low latency in supercomputers
- 22. main memory
- 23. A situation that occurs where two or more processes or threads are unable to proceed because each is waiting for a resource held by another process
- 25. performance metric that assess the raw computational power of a supercomputer
- 27. Highly efficient Parallel computing model
Down
- 1. __________ balancing redistributes the computational tasks
- 3. _______ management is essential for uniform cooling across all computing nodes
- 4. ______ replication is used to achieve deterministic behaviour in parallel algorithms
- 5. _____scaling is concerned with speedup of a fixed problem size with respect to the number of processors and is governed by Amdahl’s law
- 7. highly customisable open source OS for various hardware architectures
- 8. ______ algorithm produces the same output for a given input and execution environment, regardless of how many times it is executed.
- 10. common network tompolgy for supercomputers
- 13. Hardware Architecture for Supercomputing
- 14. ____________ synchronization mechanism is used to prevent race condition
- 17. fastest and the smallest memory unit directly connected to the CPU
- 20. ___________ address space allows computing nodes to access remote memory directly.
- 24. programming model that facilitates parallelization within a single computing node
- 26. optimization that reduces runtime by serial performance tuning