...
 
Commits (2)
# Artifact Evaluation for *Novel methodologies for predictable CPU-to-GPU command offloading*
# API comparison of CPU-to-GPU command offloading latency on embedded platforms
Artifact companion for the paper:
Roberto Cavicchioli, Nicola Capodieci, Marco Solieri and Marko Bertogna,
*Novel methodologies for predictable CPU-to-GPU command offloading*,
in Proceedings of the 31st Conference on Real-Time Systems (ECRTS'19),
to appear.
## Requirements
......
......@@ -3,6 +3,8 @@ Copyright © 2019
Roberto Cavicchioli, Nicola Capodieci
See LICENSE.txt
The CUDA C implementation to compare with is provided in comments.
*/
#version 450
......@@ -68,4 +70,5 @@ __global__ void activationGPU(TYPE *d_Result, TYPE *d_Data, int dataW, int dataH
else
d_Result[gLoc] = 1.0 / (1.0 + __expf((float)d_Data[gLoc]));
}
*/
\ No newline at end of file
*/
......@@ -3,6 +3,8 @@ Copyright © 2019
Roberto Cavicchioli, Nicola Capodieci
See LICENSE.txt
The CUDA C implementation to compare with is provided in comments.
*/
#version 450
......@@ -136,4 +138,5 @@ for (int i = -KERNEL_RADIUS; i <= KERNEL_RADIUS; i++) // row wise
d_Result[gLoc] = sum;
}
*/
\ No newline at end of file
*/
......@@ -3,6 +3,8 @@ Copyright © 2019
Roberto Cavicchioli, Nicola Capodieci
See LICENSE.txt
The CUDA C implementation to compare with is provided in comments.
*/
#version 450
......@@ -67,4 +69,5 @@ __global__ void matrixMulGPU(TYPE *d_Result, TYPE *d_Data, TYPE *d_Weights, int
}
}
*/
\ No newline at end of file
*/