-
OpenMP program that executes on four threads, identifies the threads executing and the total number of threads, and displays results on screen.
-
OpenMP program that computes the sum of the
p
elements of a vector of doubles. The maximum number of threads to be expanded isp / 2
. Suggestion: Takep
as power of two. -
OpenMP program that computes the sum of the
n
elements of a vector of doubles, spanning a maximum ofp
threads, both with and without using reduction instructions. Suggestion: Takep
as power of 2 andn
as a multiple ofp
. -
OpenMP program that computes the element holding the greatest absolute value in a vector of doubles, using a maximum of
p
threads. -
OpenMP program computing the scalar product of two double vectors of
n
elements each, spanning a maximum ofp
threads. -
OpenMP program that computes the euclidean norm of a float vector, spanning a maximum of
p
threads. Suggestion: It must not fail if the euclidean norm is smaller than the greatest positive number that can be represented using floating simple precision. -
OpenMP program that computes the sum of two float vectors, one of them multiplied by an
alpha
constant (saxpy operation), spanning a maximum ofp
threads. -
OpenMP program that computes the integral, between
a
andb
, of a polynomical function of gradem
, splitting the interval[a, b]
inn
subintervals and spanning a maximum ofp
threads. -
OpenMP program that multiplies two long integers stored as a couple of vectors, spanning a maximum of
p
threads. -
OpenMP program that computes the matrix of minimum distances in a directed graph using Floyd's algorithm.
-
MPI program that executes on four processes, sends a greetings message to the others and prints the greetings message from each process identifying its range.
-
MPI program that computes the sum of two vectors of
n
real elements expressed in simple precision, one of them multiplied by a constantalpha
(operation saxpy). -
MPI program that computes the sum of the
n
elements of a double vector, with and without using reduction instructions. The root process must generate data, distribute them to other processes, gather partial results and aggregate them into final results to be displayed. Suggestion: Take the number of processorsp
as a power of 2 andn
as a multiple ofp
. -
MPI program that computes the scalar product of two
n
-length double vectors. Process 0 must generate data, distribute them to other processes, and retrieve and display final results. -
MPI program that computes the machine network's values for
T
andB
. A ping-pong algorithm must be implemented for it, which sends a message from a processor to another one. Next, the second processor sends back the message to the first, and the roundtrip time is measured. If the message size is 0 bytes (or a small number), time will basically denote theB
term. If the message size is large enough, time will basically denote theT
term. Don't forget to repeat the sending a high number of times to avoid errors in the measurement of small times. -
MPI program that computes the machine network's values for
T
andB
, calculating a regression by least squares. A ping-pong algorithm must be implemented for it, which sends a message from a processor to another one. Next, the second processor sends back the message to the first. The roundtrip time will be measured for several message sizes. Empirical data must be adjusted to a line. The ordinate in the origin point will basically match theb
term. The line's slope will basically match thet
term. -
MPI program that computes the euclidean norm of a float vector containing
n
elements. Suggestion: It must not fail if the euclidean norm is smaller than the maximum positive number that can be represented using floating simple precision. The root process must generate data, distribute them to other processes, and retrieve, gather and display the final results. -
MPI program that computes the integral, between
a
andb
, of a polynomical function of gradem
, splitting the interval[a, b]
inn
subintervals. -
MPI program that multiplies two long integers stored as a couple of vectors.
-
MPI program that computes the matrix of minimum distances in a directed graph using Floyd's algorithm.
-
CUDA program that computes the sum of two vectors with
n
elements each. -
CUDA program that computes the sum of two matrices of
m * n
size. -
CUDA program that computes the sum of two vectors with
n
real elements expressed in simple precision, one of them multiplied by analpha
constant (operation saxpy). -
CUDA program that computes the product of a matrix of dimensions
m * n
and a vector withn
real elements expressed in simple precision. -
CUDA program that computes the product of two matrices of
m * n
unsigned long integer elements.