knob

enable-user-sync

true | false. Default: false

Collect synchronization data via the User-Defined Synchronization API.

enable-user-tasks

true | false. Default: false

Analyze tasks, events and counters specified in your application via the Task API. This option causes higher overhead and increases result size.

analyze-openmp

true | false.

Default: true for the HPC Performance Characterization analysis; false for other analysis types.

hotspots, advanced-hotspots, concurrency, hpc-performance, memory-access, general-exploration, runsa

Instrument the OpenMP* Analysis in your application to group performance data by regions/work-sharing constructs and detect inefficiencies such as imbalance, lock contention, or overhead on performing scheduling, reduction, and atomic operations. Using this option may cause higher overhead and increase the result size.

enable-gpu-usage

true | false. Default: false

Analyze frame rate and usage of Intel HD Graphics and Intel Iris Graphics engines and identify whether your application is GPU or CPU bound.

enable-gpu-runtimes

true | false. Default for gpu-hotspots: true, for runss: false.

Analyze execution of OpenCL™ kernels and Intel® Media SDK programs on Intel HD Graphics and Intel Iris Graphics. This option may affect the performance of your application on the CPU side.

gpu-sampling-interval

A number (in milliseconds) between 0.1 and 1000. Default: 1.

Specify an interval between GPU samples.

gpu-counters-mode

none (default for runss), overview (default for gpu-hotspots), global-local-accesses, compute-extended, full-compute (preview)

Analyze performance data from Intel HD Graphics and Intel Iris Graphics based on the preset counter sets:

• overview - track general GPU memory accesses such as Memory Read/Write Bandwidth, GPU L3 Misses, Sampler Busy, Sampler Is Bottleneck, and GPU Memory Texture Read Bandwidth. These metrics can be useful for both graphics and compute-intensive applications.

• global-local-accesses - include metrics that distinguish accessing different types of data on a GPU: Untyped Memory Read/Write Bandwidth, Typed Memory Read/Write Transactions, SLM Read/Write Bandwidth, Render/GPGPU Command Streamer Loaded, and GPU EU Array Usage. This metrics are useful for compute-intensive workloads on the GPU.

• compute-extended - analyze GPU activity on the Intel processor code name Broadwell. This metrics set is disabled for other systems.

• full-compute (preview) - collect both overview and compute-basic metrics with the allow-multiple-runs option enabled to analyze all types of EUs array stalled/idle issues in the same view.

sampling-interval

For user-mode sampling and tracing types: a number (in milliseconds) between 1 and 1000. Default: 10

For hardware event-based sampling types: a number (in milliseconds) between 0.01 and 1000. Default: 1.

Specify a sampling interval (in milliseconds) between CPU samples.

collection-detail

hotspots-sampling (default)

advanced-hotspots, system-overview

Identify application hotspots based on such basic hardware events as Clockticks and Instructions Retired.

stack-sampling

Identify hardware hotspots, explore statistically reconstructed call flow of your program and analyze thread scheduling.

stack-and-callcount

Identify hardware hotspots, analyze thread scheduling, explore call stacks and statistically approximated number of calls to sampled functions. This value is used for advanced-hotspots only.

stack-call-and-tripcount

Extend the stack-and-callcount collection with an analysis of loop trip count statistically estimated using the hardware events. This value is used for advanced-hotspots only.

enable-stack-collection

true | false. Default: false

tsx-hotspots, hpc-performance, gpu-hotspots, runsa

Enable Hardware Event-based Sampling Collection with Stacks.

dram-bandwidth-limits

true | false.

Default: true for the HPC Performance Characterization and General Exploration analysis with collect-memory-bandwidth knob enabled; true for the Memory Access analysis.

memory-access, general exploration, hpc-performance, runsa

Evaluate maximum achievable local DRAM bandwidth before the collection starts. This data is used to scale bandwidth metrics on the timeline and calculate thresholds.

collect-memory-bandwidth

true | false. Default: false

Collect data to identify where your application is generating significant bandwidth to DRAM. To view collected data in GUI, enable the Interpreting Memory Usage Data.

analyze-mem-objects

true | false. Default: false

Enable the instrumentation of memory allocation/de-allocation and map hardware events to memory objects. This option is supported for Linux targets only running on the Intel microarchitecture code name Sandy Bridge (or later).

mem-object-size-min-thres

Default: 1024 bytes

Specify a minimal size of memory allocations to analyze. This option helps reduce runtime overhead of the instrumentation.

This option is supported for Linux targets only running on the Intel microarchitecture code name Sandy Bridge (or later).

event-mode

all | user | os. Default: all

Limit event-based sampling collection to OS or USER mode.

enable-cache-metrics

true | false. Default: true

Analyze the data locality. Good data accessibility makes your code efficient and helps benefit from vectorization. This analysis includes only L1 metrics since L2 and FILL events on the Intel® Xeon Phi™ coprocessor code named Knights Corner are counting both demand loads and stores as well as multiple types of prefetches. Since other events do not count all of the prefetches accurately, the formulas cannot be adjusted to calculate real demand L2 hits or misses.

enable-l2-events

true | false. Default: true

Extend cache analysis with adding the L2_DATA_READ/WRITE_MISS_CACHE_FILL and L2_DATA_READ/WRITE_MISS_MEM_FILL events for some insight into L2 data locality. The general ratio of these events may indicate remote cache accesses that have as high a latency as memory accesses and should be avoided if possible.

enable-tlb-metrics

true | false. Default: false

Translation Lookaside Buffer (TLB) is a cache used for mapping virtual address to physical ones. Analyze the rate of TLB misses to identify performance problems.

enable-vpu-metrics

true | false. Default: false

Identify whether the level of vectorization usage (instruction-level parallelism) in your program is sufficient.

analysis-step

cycles | aborts. Default: cycles

Specify a step for analyzing Intel Transactional Synchronization Extensions behavior. Typically, you start with measuring transactional success (cycles) and then, if the aborts rate is high, you run the TSX Exploration to analyze for aborts.

analyze-loops

true | false. Default: false

Extend loop analysis to collect advanced loops information such as instruction set usage and display analysis results by loops and functions.

mrte-type

java java,python | python . Default: java

Specify a type of managed runtime to analyze. Available values: Java* analysis, combined Java and Python* analysis, and Python only.

io-mode

off | stack | nostack. Default: off

Enable to identify where threads are waiting or compute thread concurrency. The collector instruments APIs, which causes higher overhead and increases result size.

ftrace-config

Available events are freq, idle, sched, disk, filesystem, irq, kvm, workq, softirq, sync.

Default for Linux targets: sched,freq,idle,workq,irq,softirq

Default for Android targets: sched,freq,idle,workq,filesystem, irq,softirq,sync,disk

Collect Linux Ftrace* framework events.

• This option is supported for Linux target systems only.

• On some systems, Linux Ftrace events collection is possible only for the root user.

stackwalk-mode

online | offline. Default: offline

Choose between online (during collection) and offline (after collection) modes to analyze stacks. Offline mode reduces analysis overhead and is typically recommended.

stack-stitching

true | false. Default: true

For Intel TBB-based applications, restructure the call flow to attach stacks to a point introducing a parallel workload.

cpu-samples-mode

off | stack | nostack. Default: off

Enable to periodically sample the application. Samples can be collected with or without stacks.

waits-mode

off | stack | nostack. Default: off

Enable to identify where threads are waiting or compute thread concurrency. The collector instruments APIs, which causes higher overhead and increases result size.

signals-mode

off | objects | stack | nostack. Default: off

Enable to view synchronization transitions in the timeline and signalling call stacks for associated waits. The collector instruments signalling APIs, which causes higher overhead and increases result size.

no-altstack

true | false. Default: false

Disable using alternative stacks for signal handlers. Consider this option for profiling standard Python 3 code on Linux.

collect-io-waits

true | false. Default: false

Analyze the percentage of time each thread and CPU spends in I/O wait state.

stack-size

A number between 0 and 2147483647. Default is 0 (unlimited stack size).

Reduce the collection overhead and limit the stack size (in bytes) processed by the VTune Amplifier.

stack-type

software | lbr. Default: software

Choose between software stack and hardware LBR-based stack types. Software stacks have no depth limitations and provide more data while hardware stacks introduce less overhead. Typically, software stack type is recommended unless the collection overhead becomes significant. Note that hardware LBR stack type may not be available on all platforms.

enable-call-counts

true | false. Default: false

Obtain statistical estimation of call counts based on hardware events.

enable-trip-counts

true | false. Default: false

Obtain statistical estimation of loop trip counts based on hardware events.

event-config

<event_name1>,<event_name2>,...

Configure PMU events to collect with the hardware event-based sampling collector. Multiple events can be specified as a comma-separated list (no spaces).

chipset-event-config

"event1,event2,..."

Specify a comma-separated list of Android chipset events (up to 5 events) to monitor with the hardware event-based sampling collector.

enable-system-cswitch

true | false. Default: false

Analyze detailed scheduling layout for all threads on the system and identify the nature of context switches for a thread (preemption or synchronization).

atrace-config

Available events are gfx, input, view, webview, wm, am, audio, video, camera, hal, res, dalvik.

Collect Android framework events from Systrace*.

collect-tsx-cycles

true | false. Default: false

Collect the events required to analyze transactional success.

enable-context-switches

true | false. Default: false

Analyze detailed scheduling layout for all threads in your application, explore time spent on a context switch and identify the nature of context switches for a thread (preemption or synchronization).