Reference no: EM133105280
Assignment: Intro to HPC: implement a memory bandwidth benchmark
For this assignment, you are required to implement using x86-64 assembly, a simplified (i.e., single- core) version of the STREAM benchmark. This benchmark measures the memory bandwidth of the computer using simple array operations (kernels).
Considering A, B, C are arrays of length N , and q is a scalar, the four STREAM kernels are the following:
• COPY: A[i] = B[i], for i ≤ N
• SCALE: A[i] = q · B[i], for i ≤ N
• ADD: A[i] = B[i] + C[i], for i ≤ N
• TRIAD: A[i] = B[i] + q · C[i], for i ≤ N
Exercise: For this assignment, your goal is to implement the four STREAM array operations in x86-64 assembly and compare its performance with the default (single-core) STREAM im- plementation. The input arrays should consist of 64 bit unsigned integers. Is the performance better/worse? Why?
Basic requirements:
• Implement the four array operations.
• Find N large enough such that the data does not fit into the CPU cache (i.e., if the array is too small, the bandwidth you measure may be the cache bandwidth).
• Measure the running time of each operation (for this you may use a system call).
• Run each operation 20 times and report average running time.
• Based on the runtime, report the best achieved memory bandwidth.
Extra requirements:
• Vectorize the previously designed code using Intel AVX, or your SIMD instruction set of choice.
• Compare its achieved performance with the basic version. Does it perform better? Why?
What is blue ocean strategy
: What is blue ocean strategy and what is its role in digital economy? explain with example and citation.
|
Upcoming ai projects and initiatives of a company
: Provide a general framework design that ensures that the upcoming AI projects and initiatives of a company that you are familiar with, will remain within an eth
|
What is the taxable equivalent yield
: What is the taxable equivalent yield on a 20-year par value municipal bond that generates a tax-free bond yield of 5.75% if the marginal income tax rate
|
Hotel management-rooms division discussion
: AirBnB's (bed and breakfast) are new competition for hotel chains. They offer a place to stay, at a more affordable price, and provide laundry, cooking, and fri
|
Implement a memory bandwidth benchmark
: Implement the four STREAM array operations in x86-64 assembly and compare its performance with the default (single-core) STREAM
|
Define sustainability in context of social responsibility
: Define sustainability in the context of social responsibility. How does adopting this concept affect the way businesses operate?
|
Construct a time series plot
: For the Hawkins Company, the monthly percentages of all shipments received on time over the past 12 months are 80, 82, 84, 83, 83, 84, 85, 84, 82, 83, 84, and 8
|
Determine how many of pound of direct materials in inventory
: Determine how many of pounds of direct materials are currently in Howard's January 1st inventory. Beginning materials inventory
|
Importance of gathering or collecting and interpreting data
: What is the importance of gathering or collecting and interpreting data and information about competitors? What practices should a firm use to gather competitor
|