AI/rocm-systems
2
0
Форкнуть 0
Код Задачи Запросы на слияние Действия Пакеты Проекты Релизы Вики Активность

rocr: GFX9, GFX10, GFX11: Use view3dAs2dArray flag, for thick/3D swizzle modes. (#58)

Просмотр исходного кода 
A HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG environment flag exists already to
enable/disable this. Default value is false (view3dAs2dArray = 1)
Enabling this flag will enable support for swizzles that do 3D
interleaving on GFX9, GF10 and GFX11. By default support for swizzles that
do 3D interleaving is disabled.
Этот коммит содержится в:
Khatri, Shweta
2025-02-26 09:38:17 -05:00
коммит произвёл GitHub
родитель d3a4dc9687
Коммит 0984a1f0fd
4 изменённых файлов: 114 добавлений и 8 удалений
Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы
runtime/hsa-runtime/image/image_manager_ai.cpp
+29 -2
Просмотреть файл
@@ -48,6 +48,7 @@
#include <algorithm>
#include <climits>
#include "core/inc/runtime.h"
#include "hsakmt/hsakmt.h"
#include "inc/hsa_ext_amd.h"
#include "core/inc/hsa_internal.h"
@@ -535,8 +536,34 @@ uint32_t ImageManagerAi::GetAddrlibSurfaceInfoAi(
in.resourceType = ADDR_RSRC_TEX_2D;
break;
case HSA_EXT_IMAGE_GEOMETRY_3D:
in.resourceType = ADDR_RSRC_TEX_3D;
break;
{
in.resourceType = ADDR_RSRC_TEX_3D;
/*
* 3D swizzle modes enforce alignment
* of the number of slices to the block depth.
* If numSlices = 3 then the 3 slices are
* interleaved for 3D locality among the 8 slices
* that make up each block. This causes the memory
* footprint to jump to a 3x size of the ideal size
*
* 'enable3DSwizzleMode' flag tests for env variable
* HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG to enable or disable
* 3D swizzle:
* true: Keep view3dAs2dArray = 0 for real 3D interleaving.
* false: Use view3dAs2dArray = 1 to avoid the alignment
* expansion.
* 2D swizzle modes can lower size overhead but may yield
* suboptimal cache behavior for fully 3D volumetric
* operations.
*/
bool enable3DSwizzleMode = core::Runtime::runtime_singleton_->flag().enable_3d_swizzle();
if (enable3DSwizzleMode)
in.flags.view3dAs2dArray = 0;
else
in.flags.view3dAs2dArray = 1;
break;
}
}
in.flags.texture = 1;
runtime/hsa-runtime/image/image_manager_gfx11.cpp
+29 -2
Просмотреть файл
@@ -48,6 +48,7 @@
#include <algorithm>
#include <climits>
#include "core/inc/runtime.h"
#include "inc/hsa_ext_amd.h"
#include "core/inc/hsa_internal.h"
#include "addrlib/src/core/addrlib.h"
@@ -658,8 +659,34 @@ uint32_t ImageManagerGfx11::GetAddrlibSurfaceInfoNv(
break;
case HSA_EXT_IMAGE_GEOMETRY_3D:
in.resourceType = ADDR_RSRC_TEX_3D;
break;
{
in.resourceType = ADDR_RSRC_TEX_3D;
/*
* 3D swizzle modes enforce alignment
* of the number of slices to the block depth.
* If numSlices = 3 then the 3 slices are
* interleaved for 3D locality among the 8 slices
* that make up each block. This causes the memory
* footprint to jump to a 3x size of the ideal size
*
* 'enable3DSwizzleMode' flag tests for env variable
* HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG to enable or disable
* 3D swizzle:
* true: Keep view3dAs2dArray = 0 for real 3D interleaving.
* false: Use view3dAs2dArray = 1 to avoid the alignment
* expansion.
* 2D swizzle modes can lower size overhead but may yield
* suboptimal cache behavior for fully 3D volumetric
* operations.
*/
bool enable3DSwizzleMode = core::Runtime::runtime_singleton_->flag().enable_3d_swizzle();
if (enable3DSwizzleMode)
in.flags.view3dAs2dArray = 0;
else
in.flags.view3dAs2dArray = 1;
break;
}
}
in.flags.texture = 1;
runtime/hsa-runtime/image/image_manager_kv.cpp
+28 -2
Просмотреть файл
@@ -48,6 +48,7 @@
#include <algorithm>
#include <climits>
#include "core/inc/runtime.h"
#include "hsakmt/hsakmt.h"
#include "inc/hsa_ext_amd.h"
#include "core/inc/hsa_internal.h"
@@ -858,8 +859,33 @@ bool ImageManagerKv::GetAddrlibSurfaceInfo(
case HSA_EXT_IMAGE_GEOMETRY_3D:
case HSA_EXT_IMAGE_GEOMETRY_2DA:
case HSA_EXT_IMAGE_GEOMETRY_2DADEPTH:
in.resourceType = ADDR_RSRC_TEX_3D;
break;
{
in.resourceType = ADDR_RSRC_TEX_3D;
/*
* 3D swizzle modes enforce alignment
* of the number of slices to the block depth.
* If numSlices = 3 then the 3 slices are
* interleaved for 3D locality among the 8 slices
* that make up each block. This causes the memory
* footprint to jump to a 3x size of the ideal size
* 'enable3DSwizzleMode' flag tests for env variable
* HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG to enable or disable
* 3D swizzle:
* true: Keep view3dAs2dArray = 0 for real 3D interleaving.
* false: Use view3dAs2dArray = 1 to avoid the alignment
* expansion.
* 2D swizzle modes can lower size overhead but may yield
* suboptimal cache behavior for fully 3D volumetric
* operations.
*/
bool enable3DSwizzleMode = core::Runtime::runtime_singleton_->flag().enable_3d_swizzle();
if (enable3DSwizzleMode)
in.flags.view3dAs2dArray = 0;
else
in.flags.view3dAs2dArray = 1;
break;
}
}
in.flags.texture = 1;
runtime/hsa-runtime/image/image_manager_nv.cpp
+28 -2
Просмотреть файл
@@ -48,6 +48,7 @@
#include <algorithm>
#include <climits>
#include "core/inc/runtime.h"
#include "inc/hsa_ext_amd.h"
#include "core/inc/hsa_internal.h"
#include "addrlib/src/core/addrlib.h"
@@ -653,8 +654,33 @@ uint32_t ImageManagerNv::GetAddrlibSurfaceInfoNv(
break;
case HSA_EXT_IMAGE_GEOMETRY_3D:
in.resourceType = ADDR_RSRC_TEX_3D;
break;
{
in.resourceType = ADDR_RSRC_TEX_3D;
/*
* 3D swizzle modes enforce alignment
* of the number of slices to the block depth.
* If numSlices = 3 then the 3 slices are
* interleaved for 3D locality among the 8 slices
* that make up each block. This causes the memory
* footprint to jump from an ideal size to 3x the size.
* 'enable3DSwizzleMode' flag tests for env variable
* HSA_IMAGE_ENABLE_3D_SWIZZLE_DEBUG to enable or disable
* 3D swizzle:
* true: Keep view3dAs2dArray = 0 for real 3D interleaving.
* false: Use view3dAs2dArray = 1 to avoid the alignment
* expansion.
* 2D swizzle modes can lower size overhead but may yield
* suboptimal cache behavior for fully 3D volumetric
* operations.
*/
bool enable3DSwizzleMode = core::Runtime::runtime_singleton_->flag().enable_3d_swizzle();
if (enable3DSwizzleMode)
in.flags.view3dAs2dArray = 0;
else
in.flags.view3dAs2dArray = 1;
break;
}
}
in.flags.texture = 1;