Jeff Bolz
a20d810d79
This fixes some failures on Turing where "round to zero" rounds to the max f16 value but the CPU reference value is infinite.
22 lines
544 B
Plaintext
22 lines
544 B
Plaintext
#version 450
|
|
|
|
#include "rte.comp"
|
|
#include "generic_head.comp"
|
|
#include "types.comp"
|
|
|
|
#extension GL_EXT_control_flow_attributes : enable
|
|
|
|
layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
|
|
|
|
layout (binding = 0) readonly buffer X {A_TYPE data_a[];};
|
|
layout (binding = 1) writeonly buffer D {D_TYPE data_d[];};
|
|
|
|
void main() {
|
|
const uint i = gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x;
|
|
|
|
if (i >= p.KX) {
|
|
return;
|
|
}
|
|
data_d[i] = D_TYPE(exp(float(data_a[i])));
|
|
}
|