blender
/
blender
137
404
Fork 586
Code Issues 6.2k Pull Requests 757 Projects 19 Wiki Activity

Cyles: switch primitive.h inline hints to forceinline 

This change helps decrease Intel GPU binaries compile time by 5-10
minutes without impacting other backends.

Reviewed By: sergey, brecht

Differential Revision: http://developer.blender.org/D15273
This commit is contained in:
Xavier Hallade 2022-06-23 18:30:49 +02:00
parent 2eba15d3e8
commit 633c2f07da
1 changed files with 29 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
intern/cycles/kernel/geom/primitive.h
View File

@ -18,11 +18,11 @@ CCL_NAMESPACE_BEGIN
* attributes for performance, mainly for GPU performance to avoid bringing in
* heavy volume interpolation code. */
ccl_device_inline float primitive_surface_attribute_float(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float *dx,
ccl_private float *dy)
ccl_device_forceinline float primitive_surface_attribute_float(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float *dx,
ccl_private float *dy)
{
if (sd->type & PRIMITIVE_TRIANGLE) {
if (subd_triangle_patch(kg, sd) == ~0)
@ -49,11 +49,11 @@ ccl_device_inline float primitive_surface_attribute_float(KernelGlobals kg,
}
}
ccl_device_inline float2 primitive_surface_attribute_float2(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float2 *dx,
ccl_private float2 *dy)
ccl_device_forceinline float2 primitive_surface_attribute_float2(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float2 *dx,
ccl_private float2 *dy)
{
if (sd->type & PRIMITIVE_TRIANGLE) {
if (subd_triangle_patch(kg, sd) == ~0)
@ -80,11 +80,11 @@ ccl_device_inline float2 primitive_surface_attribute_float2(KernelGlobals kg,
}
}
ccl_device_inline float3 primitive_surface_attribute_float3(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float3 *dx,
ccl_private float3 *dy)
ccl_device_forceinline float3 primitive_surface_attribute_float3(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc,
ccl_private float3 *dx,
ccl_private float3 *dy)
{
if (sd->type & PRIMITIVE_TRIANGLE) {
if (subd_triangle_patch(kg, sd) == ~0)
@ -149,15 +149,15 @@ ccl_device_forceinline float4 primitive_surface_attribute_float4(KernelGlobals k
* attributes for performance, mainly for GPU performance to avoid bringing in
* heavy volume interpolation code. */
ccl_device_inline bool primitive_is_volume_attribute(ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
ccl_device_forceinline bool primitive_is_volume_attribute(ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
{
return sd->type == PRIMITIVE_VOLUME;
}
ccl_device_inline float primitive_volume_attribute_float(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
ccl_device_forceinline float primitive_volume_attribute_float(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
{
if (primitive_is_volume_attribute(sd, desc)) {
return volume_attribute_value_to_float(volume_attribute_float4(kg, sd, desc));
@ -167,9 +167,9 @@ ccl_device_inline float primitive_volume_attribute_float(KernelGlobals kg,
}
}
ccl_device_inline float3 primitive_volume_attribute_float3(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
ccl_device_forceinline float3 primitive_volume_attribute_float3(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
{
if (primitive_is_volume_attribute(sd, desc)) {
return volume_attribute_value_to_float3(volume_attribute_float4(kg, sd, desc));
@ -179,9 +179,9 @@ ccl_device_inline float3 primitive_volume_attribute_float3(KernelGlobals kg,
}
}
ccl_device_inline float4 primitive_volume_attribute_float4(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
ccl_device_forceinline float4 primitive_volume_attribute_float4(KernelGlobals kg,
ccl_private const ShaderData *sd,
const AttributeDescriptor desc)
{
if (primitive_is_volume_attribute(sd, desc)) {
return volume_attribute_float4(kg, sd, desc);
@ -194,7 +194,7 @@ ccl_device_inline float4 primitive_volume_attribute_float4(KernelGlobals kg,
/* Default UV coordinate */
ccl_device_inline float3 primitive_uv(KernelGlobals kg, ccl_private const ShaderData *sd)
ccl_device_forceinline float3 primitive_uv(KernelGlobals kg, ccl_private const ShaderData *sd)
{
const AttributeDescriptor desc = find_attribute(kg, sd, ATTR_STD_UV);
@ -262,8 +262,8 @@ ccl_device float3 primitive_tangent(KernelGlobals kg, ccl_private ShaderData *sd
/* Motion vector for motion pass */
ccl_device_inline float4 primitive_motion_vector(KernelGlobals kg,
ccl_private const ShaderData *sd)
ccl_device_forceinline float4 primitive_motion_vector(KernelGlobals kg,
ccl_private const ShaderData *sd)
{
/* center position */
float3 center;