Presentation is loading. Please wait.

Presentation is loading. Please wait.

Matthew Royle Supervisor: Prof Shaun Bangay.  How do we implement OpenCL for CPUs  Differences in parallel architectures  Is our CPU implementation.

Similar presentations


Presentation on theme: "Matthew Royle Supervisor: Prof Shaun Bangay.  How do we implement OpenCL for CPUs  Differences in parallel architectures  Is our CPU implementation."— Presentation transcript:

1 Matthew Royle Supervisor: Prof Shaun Bangay

2

3

4  How do we implement OpenCL for CPUs  Differences in parallel architectures  Is our CPU implementation efficient  Is parallel faster than sequential  Is GPU execution faster than CPU

5

6  Devices and Context  Buffers  Kernel  Execution Model

7

8

9

10

11

12  Devices and Context  Buffers  Kernel  Execution Model

13

14

15

16

17

18

19

20 int (a[1024]) 2 int a[1024] int (a[1024]) 2

21

22  We can implement OpenCL for CPUs  OpenCL design doesn’t take advantage of CPU architecture  Executing kernels on CPUs not efficient  Parallel execution faster than sequential  GPUs more suited to parallel processing

23  Heterogeneous Parallel Processing Language  Design tailored for GPUs  Implement for CPUs  Variance in parallel execution times  Parallel execution faster than sequential  NVIDIA’s Implementation for GPUs faster than our implementation for CPUs


Download ppt "Matthew Royle Supervisor: Prof Shaun Bangay.  How do we implement OpenCL for CPUs  Differences in parallel architectures  Is our CPU implementation."

Similar presentations


Ads by Google