Intel® Advisor User Guide

ID 766448
Date 11/07/2023

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents


Run the specified type of analysis and collect data.

GUI Equivalent

Analysis Workflow

File > New > Start [Name] Analysis


-c=<string> [--action-options] [--global-options] [[--] <target> [<target options>]]

--collect=<string> [--action-options] [--global-options] [[--] <target> [<target options>]]


<string> is the type of analysis:




Survey the target (your executable application) and collect data about code that may benefit from (more) parallelism.


Collect dependencies data to predict and eliminate data sharing problems.


Collect memory access patterns data.


Run the Offload Modeling perspective analyses with a single command.


Project performance on a target device.


Run the Survey analysis immediately followed by the Trip Counts & FLOP analysis to visualize actual performance against hardware-imposed performance ceilings.


Collect suitability data by executing annotated code to analyze the proposed threading parallelism opportunities and estimate where performance gains are most likely.


Collect the following data and add it to the Survey report: loop iteration, floating-point and integer operation, and memory traffic statistics, and more.


No default argument


accuracy, app-working-dir, assume-dependencies, assume-hide-taxes, assume-ndim-dependency, assume-single-data-transfer, auto-finalize, batching, benchmarks-sync, cache-config, cache-simulation, cache-sources, cachesim, cachesim-associativity, cachesim-cacheline-size, cachesim-mode, cachesim-sampling-factor, cachesim-sets, check-profitability, config, count-logical-instructions, count-memory-instructions, count-memory-objects-accesses, count-mov-instructions, count-send-latency, cpu-scale-factor,custom-config, data-limit, data-reuse-analysis, data-transfer, data-transfer-histogram, data-transfer-page-size, delete-tripcounts, disable-fp64-math-optimization, dry-run, duration, enable-cache-simulation, enable-data-transfer-analysis, enforce-baseline-decomposition, enforce-fallback, enforce-offloads. estimate-max-speedup, evaluate-min-speedup, exclude-files, executable-of-interest, exp-dir, filter-by-scope, filter-reductions, flop, force-32bit-arithmetics, force-64bit-arithmetics, gpu, gpu-carm, gpu-sampling-interval, hide-data-transfer-tax, ignore, ignore-app-mismatch, ignore-checksums, instance-of-interest, integrated, interval, loop-call-count-limit, loop-filter-threshold, loops, mark-up, mark-up-list, mkl-user-mode, model-baseline-gpu, model-children, model-extended-math, model-system-calls, module-filter, module-filter-mode, mpi-rank, mrte-mode, ndim-depth-limit, option-file, overlap-taxes, profile-gpu, profile-intel-perf-libs, profile-jit, profile-python, profile-stripped-binariesproject-dir, quiet, record-mem-allocations, record-stack-frame, refinalize-survey, resume-after, return-app-exitcode, search-dir, search-n-dim, select, set-dependency, set-parallel, set-parameter, show-report, small-node-filter, spill-analysis, stack-access-granularity, stack-stitching, stack-unwind-limit, stacks, stackwalk-mode, start-paused, static-instruction-mix, strategy, support-multi-isa-binaries, target-device, target-gpu, target-pid, target-process, threads, trace-mode, trace-mpi, track-memory-objects, track-stack-accesses, track-stack-variables, trip-counts, verbose


Survey the application to find candidates for code that may benefit from (more) parallelism.

advisor --collect=survey --search-dir src:r=./src --project-dir=./advi_results -- ./bin/myApplication

Collect memory access patterns data on the specified loops.

advisor --collect=map --mark-up-list=5,10,12 --search-dir src:r=./src --project-dir=./advi_results -- ./bin/myApplication

Collect survey data on four nodes of an MPI cluster into the shared ./advi project directory.

mpirun -n 4 "advisor --collect=survey --project-dir=./advi_results" -- <PATH>/mpi-sample/1_mpi_sample_serial

Collect dependencies data for all innermost loops that account for over 2% of the total CPU time.

advisor --collect=dependencies --loops="loop-height=0,total-time>2 --project-dir=./advi_results" -- ./bin/myApplication

Run the Offload Modeling perspective with low accuracy.

advisor --collect=offload --accuracy=low --config=xehpg_512xve --search-dir src:r=./src --project-dir=./advi_results -- ./bin/myApplication

See Also