Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
Employee
109 Views

vtune result broken on devcloud

dear dev cloud team:

       Did you notice that GPU vtune result broken on devcloud BETA 07, compute task information has gone.

BRs

0 Kudos
4 Replies
Highlighted
Moderator
98 Views

Re:vtune result broken on devcloud

Hi,

Thanks for reaching out.

Kindly provide the below details so that we can try from our side.

1)Your application details.

2)Commands used.

3)How the results are being viewed(CLI or GUI)



0 Kudos
Highlighted
Employee
74 Views

Re: Re:vtune result broken on devcloud

1)Your application details.

  code fragment such as below.

 

void __attribute__((noinline)) MatrixMulOpenMpGpuOffloading ()
{
int i, j, k;

// a is identity matrix
for (i=0; i<M; i++)
for (j=0; j<N; j++)
a[i][j] = 1.0;

// each column of b is the sequence 1,2,...,N
for (i=0; i<N; i++)
for (j=0; j<P; j++)
b[i][j] = i+1.;

// c is initialized to zero
for (i=0; i<M; i++)
for (j=0; j<P; j++)
c[i][j] = 0;

auto t1 = chrono::steady_clock::now();

for (int i = 0 ; i < 1000; i++)
{

// Parallelize on target device
#pragma omp target teams distribute parallel for simd map(to: a, b) map(tofrom: c) //thread_limit(128)
{
for (i=0; i<M; i++) {
// Each element of the product is just the sum 1+2+...+n
for (j=0; j<P; j++) {
for (k=0; k<N; k++) {
c[i][j] += a[i][k] * b[k][j];
}
}
}
}
}
auto t2 = chrono::steady_clock::now();

cout << "matmul default takes :" << chrono::duration_cast<chrono::microseconds>(t2 - t1).count() / 1000 << "us" << std::endl;


}

 

2)Commands used.

 

vtune -collect gpu-hotspot 

 

3)How the results are being viewed(CLI or GUI)

 

GUI VTUNE 

0 Kudos
Highlighted
Moderator
24 Views

Re:vtune result broken on devcloud

Hi Leevan,

 


We are working on this internally, will get back to you soon. Thanks for your patience.


0 Kudos
Highlighted
Moderator
11 Views

Re:vtune result broken on devcloud

Hi,


We are forwarding the case to SME.



0 Kudos