Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


  • CPU Usage: Determine where BrightScript code execution is happening
  • Wall-Clock time: Determine where the most time (execution and waiting) is being spent in when a channel channel is running 
  • Function call counts
  • Memory usage, including memory leak detection – Available since firmware version 9 


  1. Add at least the required manifest entries to the channel to run the profiler
  2. Run and then Exit the channel to generate data and metrics
  3. Save the profiling data to the device or stream it to the machine (PC) you are using for development
  4. Analyze the profiling data as necessary 

Manifest entries

Below is the list of manifest keys used by the profiler: 

Manifest Key
bsprof_data_destenumlocal , networklocalYes

If this entry value is local, profiling data is collected on the device and can be downloaded from the Application Installer after the channel terminates. This is the default.

If this entry value is network, the profiling data is sent over the network rather than being stored on the device. See the section on Retrieving Profiling Data for details.

bsprof_enableboolean0 , 10YesTurns on BrightScript Profiling when the channel is runrunning. This is a the master flag and must be set to 1 for any other profiling options to take effect.
bsprof_enable_linesboolean0 , 10


Collects memory and CPU data for each line of BrightScript source code. This makes it easier to pinpoint memory and CPU usage issues. This value is set to 0 by default, which means that data is collected for each BrightScript function as a whole. Enabling this feature can have a significant impact on device performance.

Requires BrightScript profiling to be enabled (bsprof_enable=1), and the RSG version to be set to 1.2 (rsg_version=1.2).

Available since firmware version 9.1

bsprof_enable_memboolean0 , 10If Yes, if using memory profiling

Turns on memory profiling. Only has effect if bsprof

Requires BrightScript profiling to be enabled ( bsprof_enable=1).

If  If this is enabled,  the bsprof_sample_ratio is forced automatically set to 1.0.

bsprof_pause_on_startboolean0 , 10No

Immediately upon starting after launching the channel, profiling is paused until manually resumed with the bsprof-resume command on the port 8080 debug console. Only has effect if bsprof_enable=1 This is useful for profiling isolated parts of a channel's UI or operations, rather than profiling the entire startup sequence of the channel.

Requires BrightScript profiling to be enabled ( bsprof_enable=1).

bsprof_sample_ratiofloat0.001 to 1.01.0Yes

Sets how often profiling samples are taken, while the channel is running. Only has effect if bsprof_enable=1

Asampleratio of 1.0 causes every BrightScript statement to be measured and integrated into the profiling data for the channel. Unfortunately, a sample ratio of 1.0 may cause some complex channels to run very slowly when profiling is enabled, making them difficult to test. Choosing a lower sample ratio for such channels can make them more usable, while profiling is enabled. A higher ratio will yield more accurate data, so the ratio should be set at 1.0 whenever possible.

If .If memory profiling is enabled (bsprof_enable_mem=1), this value is forced automatically set to 1.0.

The bs_prof_sample_ratio can be adjusted from 0.001 to 1.0. A sample ratio of 1.0 is the default and


provides the most accurate data because every BrightScript statement is measured. A sample ratio of 1.0


, however, may slow down device performance, but does not typically affect the usability of


the channel. If slower deice performance is observed, reduce the ratio to


lessen the profiler’s overhead.







To use line-level profiling, this must be set to 1.2 (rsg_version=1.2). If it is not set to 1.2, profiling will still work correctly; however, line-level data will not be generated.

rsg_version 1.2 provides significant performance improvements; therefore, you should set it to 1.2 regardless whether your channel is using line-level profiling.

See Roku Channel Manifest for more information.

Running the profiler on a channel

To initiate the memory profiler, sideload, run, and then exit the channel. The profiling data is complete only after the channel exits. Note that memory data can be streamed using over a network. The advantage to streaming the data over the network is that it consumes significantly less memory on the device while the channel is running.

Pausing and resuming profiling

Channel profiling can be paused and resumed at any time. Use the following two commands on port 8080 to either pause or resume the memory profiler:


If the profiler is paused, very little data is written regardless of the data destination. This allows the profiling data (generally, the data relevant to specific parts of a channel's UI or other operation) to be collected and analyzed later analyzed. These two commands are particularly useful when combined with the bsprof_pause_on_start manifest entry.


For example, if starting video playback is slow or seems to cause memory leaks, the bsprof_pause_on_start=1 entry can be set in the channel's manifest. After the channel is launched, but prior to video playback, execute the bsprof-resume command on port 8080 , to begin collecting profiling data. After performing the UI operations to be profiled, execute the bsprof-pause command to suspend the storing operation of the profiling data. Then, exit the channel to make the profiling data available for analysis. In this scenario, the profiling data will pertain specifically to the operations performed between bsprof-resume and bsprof-pause.


The channel's manifest entry bsprof_data_dest determines how the profiling data is retrieved from the device. The data can be stored locally on the device and downloaded after the channel has exitedfinishes running and exits, or it can be streamed over a network connection while the channel is running.  Streaming consumes significantly less memory on the device while the channel is running.  In addition, if the channel fails, the memory data will have been accurately collected up until the time of the crash (CPU data, however, is typically lost if a failure occurs).

Data Destination: Local

Local data storage is the default mode of operationstorage for profiling, though it can be explicitly selected by adding bsprof_data_dest=local to the channel's manifest. When operating in using this mode destination, the data becomes available in on the device's Application Installer after the channel has exitedexits:

  1. Launch the channel as you normally would and run through your the test cases. Once you exit the channel , open the page to your finishes running and exits, open Roku device's Developer Settings and click on Utilities.

  2. After you have run through your channel's test cases, click Click Profiling Data to generate a .bsprof file and a link to download the data from your Roku device.

    The .bsprof format is unique to Roku to ensure the format is as efficient and small as possible and easy to generate even on low-end Roku devices.


In order to stream a channel's profiling data to the a network while the channel is running, add bsprof_data_dest=network to the channel's manifest. Streaming data over the network is especially useful when profiling a channel's memory usage because all memory operations are included in the profiling data, and the amount of space necessary to store the data can be very large. By streaming the data to the to a network, the data size is limited primarily by the host computer receiving the data, and not by the available memory on the device itself. Even while streaming the profiling data to the network, there are still additional demands placed on the device's resource resources while profiling as compared to running a channel without profiling. However, the use of resources on the device is significantly reduced.

When this feature is enabled, the start of the channel is delayed until a network connection is received by the device, which will be is the destination for the data. When the channel is launched, a message similar to this will appear the following appears on the port 8085 developer console:

08-31 23:15:29.542 [] Waiting for connection to

The URL should be is used with wget, curl, or a web browser. Once a connection is received from one of those programs, this the following message will appear appears on the developer console:

08-31 23:15:38.939 [] profiler connected

When the channel is exitedexits, this the following message will appear appears on the developer console:


Once that message is seen, the profiling connection is closed by the device and the remote file is completepopulated with profiling data.


Viewing profiling data

After downloading the .bsprof file, the data can be viewed using the BrightScript Profiler Visualization Tool.  

Image Removed

BrightScript Profiler Visualization Tool - CPU output view


Image Removed

BrightScript Profiler Visualization Tool - Memory output view

Understanding the data

The profiling data is divided into 5 main sections:

  • The function (and associated call path which can be expanded), 
  • CPU time, 
  • wall-clock time, 
  • function call counts,
  • memory profiler output

The CPU time and wall-clock time sections are further divided into separate sections for selfcallees, and total:

  • self refers to the CPU/wall-clock time the function consumes itself
  • callees is the amount of time consumed by any functions called by the original function.
  • total is the amount of time consumed by the original function (self) and any callee functions.

Function call paths

This section of the profiling data contains the function The tool lists the function calls in each thread. For SceneGraph applications, each thread corresponds to either the main BrightScript thread or a single instance of a <component>. For example, if you have a Task node that is instantiated multiple times, each instance will appear as a separate thread. The results will be are the same for any custom <component> in your  in the channel that is instantiated multiple times. The main BrightScript thread (Thread main) is also represented as a single thread even though it has no <component>.

CPU time

The first 3 columns of the visualization tool lists:


For each function listed in the tool, you can expand it to drill down further into its call path and get more detailed data on the function. If you enabled line-level profiling for the channel, you can also click on the ellipses to the right of a function to view profiling data for the individual lines of code within a function.

Image Added


The CPU tab lists CPU time, wall-clock time, and call count statistics for each function observed and for the function's associated call path (expand a function to view its call path). CPU time refers to the number of operations each function takes to complete


; this number should be equal on the low end and high-end Roku devices.


Wall-clock time


The next 3 columns lists:


refers to the real world time that a function takes to complete. This value can vary across different Roku devices. For example, a function may take an equal number of operations to complete across different Roku devices but low-end Roku devices can take more real-world time to complete one operation than a high-end Roku device.

Function call counts

As the name implies, this column lists the number of times they were called when the channel ran with profiling enabled.

Values from memory profiling 




The CPU and wall-clock time columns are further divided into separate sections for selfcallees, and total:

  • self. The CPU/wall-clock time the function consumes itself.
  • callees. The amount of CPU/wall-clock time consumed by any functions called by the original function.
  • total. The amount of CPU/wall-clock time consumed by the original function (self) and any callee functions.

Memory Graph

The Memory Graph tab visualizes the memory consumption for each thread in the channel application. Move the mouse pointer over a column to view the allocated, free, and unfreed memory for the thread. 

Image Added


The Memory tab lists allocated, freed, and unfreed memory statistics for each function observed (self) and for the function's associated call path (total).  See Profiling Values for more information on the memory statistics displayed in this tab.

Image Added

Line-Level profiling

You can collect profile data for each line of BrightScript source code to better pinpoint high CPU and memory usage. To do this, enable profiling (bsprof_enable=1), line-level profiling (bsprof_enable_lines=1), and RSG version 1.2  (rsg_version=1.2) in the channel manifest. In the CPU or Memory tab, click on the ellipses to the right of the function to drill down into its individual lines of code. 

 Image Added 

The BrightScript Profiler lists the CPU or memory usage at the time the individual lines of code in the function were executed.

Image Added

See Profiling Values for more information on the CPU and memory statistics displayed in the tool.

Profiling Values 

The following table describes the profile statistics displayed in the BrightScript profiler tool. The descriptions are applicable for the functions as a whole or for individuals lines of code in the functions (if using line-level profiling). 


Number of times a function was called


CPU* used in this by the function , itself


CPU* used in functions called by this a function

Cpu.self + cpu.callees


Memory allocated within this a function itself, but not freed (leaks)


Memory allocated by functions called by this a function, but not freed (leaks)

Mem.self + mem.callees


Real (wall-clock) time ** spent in this on the function , itself


Real (wall-clock) time ** spent in on functions called by this a function

Tm.self + tm.callees









Average of the metric, over the number of calls (e.g., if cpu.self=100 and calls=2, avg_cpu_self will be 50).


A “memory leak” is simply any memory that is allocated, but not freed while the profiler was running. If memory is freed while profiling is paused, the free


memory is not


tracked and the memory may show up as “leaked.”

Time is measured as if a stopwatch were used to time the action. For example, if a function makes a network call, there may be very little CPU time used, but a significant amount of time waiting for the network response.


If any of these metrics appear in a call path, there they are specific to that call path. For example, in this call path:


<root>: cpu.self=0,cpu.callees=14700,tm.self=0.000,tm.callees=1.989,mem.self=0,mem.callees=324452,calls=0
+- func1(): pkg:/components/file1.brs:83,cpu.self=200,cpu.callees=14500,tm.self=0.728,tm.callees=1.261,mem.self=5840,mem.callees=318612,calls=1
|  +- func2(): pkg:/components/file2.brs:22,cpu.self=14500,cpu.callees=0,tm.self=1.261,tm.callees=0.000,mem.self=31800,mem.callees=612,calls=1

The metrics for func2() are specific to when it is called from func1().


However, in this the table below:


------------- BEGIN: TOP CONSUMERS: CPU.SELF -----------------
  1: func1(): pkg:/components/file1.brs:83,cpu.self=300,,tm.self=0.001,,mem.self=0,,calls=5
  2: func2(): pkg:/components/file2.brs:22,cpu.self=55430,,tm.self=0.126,,mem.self=0,,calls=3
-------------- END: TOP CONSUMERS: CPU.SELF -----------------


The metrics displayed are totals for all calls to each function, on any call path

Top Offenders

The Top Offenders tab displays the functions or function call paths that used the greatest amount of CPU, memory, or time. This tab is typically the best place to start diagnosing potential performance issues. The top 10 functions with the most unfreed memory are listed by default. You can change the report to a top 20, 100, 500 list or enter a custom depth, and you can select which metric to view. For example, the top 5 methods with the highest total CPU time are displayed in the following image:  

Image Added

Click Show call paths to open the CPU or Memory tab with more detailed profiling data for the selected function and metric.

Using this data

Here are a few key points on how to use this data to improve channel performance:

Data TypeDefinition and Best Use
High wall-clock time but low CPU timeThis pattern shows a function is consistently waiting, whether it be for input or a response from an external source. These functions are best suited for Task nodes so that it doesn't block the main thread.
Complex functionsTry to simplify the functions as much as possible. If a function handles multiple tasks, consider breaking it out into several functions to further isolate how much CPU or wall-clock time is consumed by each task.
Functions that consume a large amount of CPU or wall-clock timeTry to reduce the number of calls to these functions as much as possible. Move functions to Task nodes, if they are consistently waiting. A function can be determined to be waiting if it's wall-clock time is high, but its CPU cost is low.