Intel Graphics Performance Analyzers
Intel Graphics Performance Analyzers
Intel® Graphics Performance Analyzers Help
Intel Graphics Performance Analyzers
Intel® Graphics Performance Analyzers Help
Metrics Descriptions
Extended Metrics Description
GPU Metrics
RT Sub-Span Writes
RT Sub-Span Writes
Description
The metric represents the number of sub-span writes to the render targets.
RT Sub-Span Writes
Intel Graphics Performance Analyzers > RT Sub-Span Writes
Intel® Graphics Performance Analyzers Help
Overview
New in Release 2012
Intel® GPA Updates
Best Practice
Product Website and Support
Intel® GPA Monitor
Introduction to the Intel® GPA Monitor
Graphical User Interface
Monitor Context Menu
Analyze Application Dialog Box
Profiles... Dialog Box
Preferences... Dialog Box
Log... Window
About... Window
Media Performance Dialog Box
Using Intel® GPA Monitor
Launching an Application
Launching the Intel® GPA Tools
Starting Application Analysis Automatically
Creating a Profile
Assigning a Profile to the Applications
Using Profiles Created with Earlier Versions of the Tool
Modifying a Profile
Changing the Profile Name
Enabling Tracing in Your Application
Setting up Tracing Options
Setting up a Trigger
Selecting Domains to be Traced
Customizing the Keyboard-shortcut Assignments
Configuring Metrics to Be Collected
Configuring HUD Display Options
Analyzing Media Workloads
About Performance Analysis of GPU-accelerated Media Applications
Analyzing a GPU-accelerated Media Application Performance
Monitoring Media Performance Metrics in Real Time
Analyzing Real-time Media Performance Metrics
Creating a Media Performance Trace Capture File
Viewing the Intel® GPA Monitor System Information
Intel® GPA System Analyzer Heads-up Display
Introduction to the Intel® GPA System Analyzer HUD
Graphical User Interface
Summary Field
Metrics Charts
Messages Field
Default Keyboard-shortcut Assignments
Using Intel GPA System Analyzer HUD
Analyzing Metrics
Creating a Frame Capture File
Creating a Trace Capture File
Creating an Intel® Graphics Checker Data Capture
Capturing a Specific Frame/Trace File in the Application
Using the Pause/Play Keyboard Shortcuts
Optimizing Graphics-rich Web Content
Changing a D3D Device for the Instrumented Process
Finding Performance Bottlenecks Using Override Modes
Override Modes Overview
Using Override Modes
Disable all Overrides, Null Hardware and Disable Draw Calls
Wireframe
2x2 Texture
Simple Pixel Shader
1x1 Scissor Rect
Disable Texture Filtering
Disable Z-Test, Disable Z-Write
Cull None, Cull Clockwise, Cull Counter-Clockwise
Disable Alpha Blending
Disable Alpha Test
Overdraw Visualization
FPS Limit
Finding the Exact Frame in the Application
Exporting Metrics
Creating a Frame and Trace Capture Files
Intel® GPA Frame Analyzer
Introduction to the Intel® GPA Frame Analyzer
Graphical User Interface
About the Graphical User Interface
Menu
File Menu
Edit Menu
View Menu
Advisor Menu
Help Menu
Visualization Settings Tool Bar
Visualization Panel
Scene Overview Panel
Render Target Viewer
Render Target Viewer Overview
Render Target View
Render Target Thumbnail List
Visualization Configuration Options
Visualization Configuration Options Overview
View
Selected Ergs
Non-Selected Ergs
Zoom in and Zoom out Buttons
Image Information Bar
Histogram
Tabs
Frame Overview Tab
Details Tab
Texture Tab
State Tab
Shaders Tab
Experiments Tab
Pixel History Tab
Erg Advisor Tab (Beta)
API Log Tab
Ergs Information
Using the Intel® GPA Frame Analyzer
Loading a Frame Capture File
Importing and Exporting Frame Capture Files
Configuring the Bar Chart
Analyzing the Graphics Application Performance
Discovering How Much Time the Graphics Application Spends within Each Pass
Discovering the Most Expensive Ergs
Running Experiments
Modifying DirectX* State
Modifying the Shader Code
Determining whether Texture Bandwidth is a Performance Bottleneck
Minimizing Overdraw
Investigating Ergs-based Issues with the Erg Advisor (Beta)
Investigating Frame-based Issues with the Frame Advisor (Beta)
Saving the Render Target Image
Saving the Texture Image
Saving the Pixel History Image
Viewing API Details
D3D Events and Markers
Intel® GPA Platform Analyzer
Introduction to the Intel® GPA Platform Analyzer
Symbol Resolution for Function Tasks
Tracing Data
Defining Styles for Tasks and Markers
Graphical User Interface
Docking User Interface
GUI Layout
Task Timeline
About the Task Timeline
Basic Concepts
Time Ruler
Aggregate Track
DX CPU and GPU Threads
Logical Tracks
Queue Tasks
Custom Tasks
Software Thread View
Hardware Thread View
Stacked Task View
Layout of the Media Performance Trace Capture File
Menu
File Menu
View Menu
Tracks Menu
Panels Menu
Available Panels
Task Groups
Statistics
Summary
Task Tree
Markers
Relation
Metadata
Window Menu
Help Menu
Auxiliary Buttons
Search Toolbar
Using the Intel® GPA Platform Analyzer
Loading a Trace Capture File
Loading Multiple Trace Capture Files
Importing and Exporting Trace Capture Files
Navigating the Task Timeline
Selecting Tasks
Showing/Hiding Hardware Tracks
Viewing Stacked Tasks
Disabling Undesirable Task Tracks
Enabling Hidden Task Tracks
Sorting the Task Tracks Order
Viewing Queue Track Tasks Instrumented with States
Displaying Aggregated Tasks with the Same Name
Analyzing a Media Performance Trace Capture File
Viewing the Trace Capture File Information
Analysis Workflows
I See Lots of Traced Data – Where Is the Majority of Time Being Spent?
Why Is this Task Longer in Duration than I Expect?
Intel® Graphics Checker Data Capture
Analyzing the Intel® Graphics Checker Data Capture
Troubleshooting
Basic Troubleshooting Tips
Troubleshooting the Intel® GPA Monitor
Troubleshooting the Intel® GPA System Analyzer HUD
Troubleshooting the Intel® GPA Platform Analyzer
Troubleshooting the Intel® GPA Frame Analyzer
Metrics Descriptions
Metrics List for Intel® Graphics Performance Analyzers
Extended Metrics Description
Compute Shader Metrics
CS Invocations
UAV Typed Reads
UAV Typed Writes
UAV Typed Atomic Ops
UAV Untyped Reads
UAV Untyped Writes
SLM Reads
SLM Writes
CPU Metrics
Aggregated CPU Load
Application Time
CPU n Load
Target App CPU Load
DirextX* Metrics
Buffer Creations
Buffer Maps
Clears
Color Fills
Draw Calls
Frames per Second
Frame Number
Frame Time
IB Creations
IB Lock Time
IB Locks
Locks
Lock Time
Maps
Resource Copy
Resource Creations
RT Changes
RT Clears
RT Data Gets
Shader Creations
State Block Applies
State Block Captures
State Changes
Stretch Rects
Surface Creations
Surface Lock Time
Surface Locks
Surface Updates
Subresource Copy
Subresource Update
Texture1D Maps
Texture2D Maps
Texture3D Maps
Texture Creations
VB Creations
VB Lock Time
VB Locks
Volume Lock Time
Volume Locks
Z/Stencil Clears
Memory Read Rate
Memory Write Rate
Memory R/W Rate
Execution Units Metrics
EUs Active in CS
EUs Active in DS
EUs Active in GS
EUs Active in HS
EUs Active in PS
EUs Active in VS
EUs Stalled in CS
EUs Stalled in DS
EUs Stalled in GS
EUs Stalled in HS
EUs Stalled in PS
EUs Stalled in VS
GPU EUs Active
GPU EUs Busy
GPU EUs Idle
GPU EUs Stalled
EUs Stalled on DAP
EUs Stalled on Math
EUs Stalled on Samp.
Geometry Shader Metrics
GS Invocations
Post-GS Primitives
GPU Metrics
GPU Busy
HUD Overhead Time
Non-Culled Polygons
PS Threads
RT Sub-Span Writes
Sampler Active
Texels Sampled
Texture Memory Reads
Z-Buffer Throughput
Input-Assembler Metrics
Primitive Count
Vertex Count
Main Metrics
GPU Duration
GPU Frequency
CS Duration
DS Duration
GS Duration
HS Duration
PS Duration
VS Duration
GS Is Bottleneck
VS Is Bottleneck
Memory Metrics
Texture Reads
GPU Memory Writes
GPU Memory Reads
Output-Merger Metrics
Alpha Tests Failed
Blended Pixels
Pixels Rendered
Sampler Is Bottleneck
Pixel Shader Metrics
PS Invocations
PS Killed Pixels
Power Metrics
CPU Power
GPU Power
Socket Power
Rasterizer Metrics
Clipper Active
Clipper Invocations
Post-Clip Primitives
Prim. Setup Active
Tesselation Metrics
Samples Written
DS Invocations
HS Invocations
Texture Sampler Metrics
Post-Filter Texels
Sampler Busy
Adding Basic Intel® ITT Instrumentation to Your Application
Minimizing Intel® ITT API Overhead
Unicode Support
Intel ® ITT API Programmer’s Guide
Marking up Your Application for Viewing in the Intel(R) GPA Platform Analyzer
Compiling and Linking with Intel® ITT
Configuring Your Build System
Including the Intel® ITT API Header in Your Application
Linking the Intel® GPA SDK Static Library to Your Application
API Reference
Custom Tracks and Track Groups
Custom_Tracks and Track Groups Overview
Functions
__itt_track_create
__itt_track_group_create
__itt_set_track
Enumerations
__itt_track_group_type
__itt_track_type
Domains
Domains Overview
Functions
__itt_domain_create
Frames
Frames Overview
Functions
__itt_frame_begin_v3
Enumerations
__itt_metadata_type
__itt_scope
Functions
__itt_metadata_add
__itt_metadata_str_add
__itt_metadata_add_with_scope
__itt_metadata_str_add_with_scope
Miscellaneous
Functions
__itt_thread_set_name
Named Entities and Instances
Named Entities and Instances Overview
IDs
IDs Overview
Functions
__itt_id_create
__itt_id_make
Task Groups
Task Groups Overview
Functions
__itt_task_group
Task States
Task States Overview
Tasks
Tasks Overview
Functions
__itt_task_begin
__itt_task_begin_fn
__itt_task_end
Markers
Markers Overview
Functions
__itt_marker
Relations
Relations Overview
Enumerations
__itt_relation
Functions
__itt_relation_add
__itt_relation_add_to_current
String Handles
String Handles Overview
Functions
__itt_string_handle_create
Functions List
Enumerations List
Disclaimer and Legal Information
Intel Graphics Performance Analyzers
en