Profiling

Quick profiling in your terminal

Note

This is only meant to be used for quick profiling or programmatically accessing the profiling results. For more detailed and GUI friendly profiling proceed to the next section.

Simply replace the use of Base.@time or Base.@timed with Reactant.Profiler.@time or Reactant.Profiler.@timed. We will automatically compile the function if it is not already a Reactant compiled function (with sync=true).

using Reactant

x = Reactant.to_rarray(randn(Float32, 100, 2))
W = Reactant.to_rarray(randn(Float32, 10, 100))
b = Reactant.to_rarray(randn(Float32, 10))

linear(x, W, b) = (W * x) .+ b

Reactant.@time linear(x, W, b)

WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
I0000 00:00:1771605373.022347    4299 profiler_session.cc:117] Profiler session initializing.
I0000 00:00:1771605373.022404    4299 profiler_session.cc:132] Profiler session started.
I0000 00:00:1771605373.022729    4299 profiler_session.cc:81] Profiler session collecting data.
I0000 00:00:1771605373.023359    4299 save_profile.cc:150] Collecting XSpace to repository: /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13/runnervmwffz4.xplane.pb
I0000 00:00:1771605373.023536    4299 save_profile.cc:123] Creating directory: /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13

I0000 00:00:1771605373.023653    4299 save_profile.cc:129] Dumped gzipped tool data for trace.json.gz to /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13/runnervmwffz4.trace.json.gz
I0000 00:00:1771605373.023669    4299 profiler_session.cc:150] Profiler session tear down.
I0000 00:00:1771605373.038097    4299 stub_factory.cc:163] Created gRPC channel for address: 0.0.0.0:33031
I0000 00:00:1771605373.038408    4299 grpc_server.cc:94] Server listening on 0.0.0.0:33031 with max_concurrent_requests 1
I0000 00:00:1771605373.038443    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: memory_profile with options: {} using ProfileProcessor
I0000 00:00:1771605373.038453    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: memory_profile
I0000 00:00:1771605373.038456    4299 memory_profile_processor.cc:47] Processing memory profile for host: runnervmwffz4
I0000 00:00:1771605373.038834    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool memory_profile: 389.387us
I0000 00:00:1771605373.051943    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: op_profile with options: {} using ProfileProcessor
I0000 00:00:1771605373.051958    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: op_profile
I0000 00:00:1771605373.052119    4299 xprof_thread_pool_executor.cc:22] Creating derived_timeline_trace_events XprofThreadPoolExecutor with 4 threads.
I0000 00:00:1771605373.053167    4299 xprof_thread_pool_executor.cc:22] Creating ProcessTensorCorePlanes XprofThreadPoolExecutor with 4 threads.
I0000 00:00:1771605373.056331    4299 xprof_thread_pool_executor.cc:22] Creating op_stats_threads XprofThreadPoolExecutor with 4 threads.
I0000 00:00:1771605373.057268    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool op_profile: 5.31613ms
I0000 00:00:1771605373.066763    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: overview_page with options: {} using ProfileProcessor
I0000 00:00:1771605373.066783    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: overview_page
I0000 00:00:1771605373.066786    4299 overview_page_processor.cc:64] OverviewPageProcessor::ProcessSession
I0000 00:00:1771605373.067116    4299 xprof_thread_pool_executor.cc:22] Creating ConvertMultiXSpaceToInferenceStats XprofThreadPoolExecutor with 1 threads.
I0000 00:00:1771605373.067925    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool overview_page: 1.147251ms
  runtime: 0.00026745s
  compile time: 4.58946207s

Reactant.@timed nrepeat=100 linear(x, W, b)

AggregateProfilingResult(
    runtime = 0.00026745s, 
    compile_time = 0.13289662s, )

Note that the information returned depends on the backend. Specifically CUDA and TPU backends provide more detailed information regarding memory usage and allocation (something like the following will be displayed on GPUs):

AggregateProfilingResult(
    runtime = 0.00003829s, 
    compile_time = 2.18053260s,  # time spent compiling by Reactant
    GPU_0_bfc = MemoryProfileSummary(
        peak_bytes_usage_lifetime = 64.010 MiB,  # peak memory usage over the entire program (lifetime of memory allocator)
        peak_stats = MemoryAggregationStats(
            stack_reserved_bytes = 0 bytes,  # memory usage by stack reservation
            heap_allocated_bytes = 30.750 KiB,  # memory usage by heap allocation
            free_memory_bytes = 23.518 GiB,  # free memory available for allocation or reservation
            fragmentation = 0.514931,  # fragmentation of memory within [0, 1]
            peak_bytes_in_use = 30.750 KiB # The peak memory usage over the entire program
        )
        peak_stats_time = 0.04975365s, 
        memory_capacity = 23.518 GiB # memory capacity of the allocator
    )
    flops = FlopsSummary(
        Flops = 2.8369974648038653e-9,  # [flops / (peak flops * program time)], capped at 1.0
        UncappedFlops = 2.8369974648038653e-9, 
        RawFlops = 4060.0,  # Total FLOPs performed
        BF16Flops = 4060.0,  # Total FLOPs Normalized to the bf16 (default) devices peak bandwidth
        RawTime = 0.00040298422s,  # Raw time in seconds
        RawFlopsRate = 1.0074836180930361e7,  # Raw FLOPs rate in FLOPs/seconds
        BF16FlopsRate = 1.0074836180930361e7,  # BF16 FLOPs rate in FLOPs/seconds
    )
)

Additionally for GPUs and TPUs, we can use the Reactant.@profile macro to profile the function and get information regarding each of the kernels executed.

Reactant.@profile linear(x, W, b)

I0000 00:00:1771605373.758584    4299 profiler_session.cc:117] Profiler session initializing.
I0000 00:00:1771605373.759396    4299 profiler_session.cc:132] Profiler session started.
I0000 00:00:1771605373.759558    4299 profiler_session.cc:81] Profiler session collecting data.
I0000 00:00:1771605373.759965    4299 save_profile.cc:150] Collecting XSpace to repository: /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13/runnervmwffz4.xplane.pb
I0000 00:00:1771605373.760196    4299 save_profile.cc:123] Creating directory: /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13

I0000 00:00:1771605373.760351    4299 save_profile.cc:129] Dumped gzipped tool data for trace.json.gz to /tmp/reactant_profile/plugins/profile/2026_02_20_16_36_13/runnervmwffz4.trace.json.gz
I0000 00:00:1771605373.760396    4299 profiler_session.cc:150] Profiler session tear down.
I0000 00:00:1771605373.760458    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: memory_profile with options: {} using ProfileProcessor
I0000 00:00:1771605373.760467    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: memory_profile
I0000 00:00:1771605373.760471    4299 memory_profile_processor.cc:47] Processing memory profile for host: runnervmwffz4
I0000 00:00:1771605373.760609    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool memory_profile: 147.516us
I0000 00:00:1771605373.760631    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: op_profile with options: {} using ProfileProcessor
I0000 00:00:1771605373.760635    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: op_profile
I0000 00:00:1771605373.760716    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool op_profile: 82.955us
I0000 00:00:1771605373.760738    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: overview_page with options: {} using ProfileProcessor
I0000 00:00:1771605373.760741    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: overview_page
I0000 00:00:1771605373.760743    4299 overview_page_processor.cc:64] OverviewPageProcessor::ProcessSession
I0000 00:00:1771605373.760831    4299 xprof_thread_pool_executor.cc:22] Creating ConvertMultiXSpaceToInferenceStats XprofThreadPoolExecutor with 1 threads.
I0000 00:00:1771605373.761308    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool overview_page: 571.326us
I0000 00:00:1771605373.856038    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: kernel_stats with options: {} using ProfileProcessor
I0000 00:00:1771605373.856069    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: kernel_stats
I0000 00:00:1771605373.856226    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool kernel_stats: 162.553us
I0000 00:00:1771605373.969326    4299 xplane_to_tools_data_with_profile_processor.cc:141] serving tool: framework_op_stats with options: {} using ProfileProcessor
I0000 00:00:1771605373.969357    4299 xplane_to_tools_data_with_profile_processor.cc:163] Using local processing for tool: framework_op_stats
I0000 00:00:1771605373.969604    4299 xplane_to_tools_data_with_profile_processor.cc:168] Total time for tool framework_op_stats: 251.419us

╔================================================================================╗
║ SUMMARY                                                                        ║
╚================================================================================╝

AggregateProfilingResult(
    runtime = 0.00026745s,
    compile_time = 0.12981680s,  # time spent compiling by Reactant
)

On GPUs this would look something like the following:

╔================================================================================╗
║ KERNEL STATISTICS                                                              ║
╚================================================================================╝

┌───────────────────┬─────────────┬────────────────┬──────────────┬──────────────┬──────────────┬──────────────┬───────────┬──────────┬────────────┬─────────────┐
│       Kernel Name │ Occurrences │ Total Duration │ Avg Duration │ Min Duration │ Max Duration │ Static Shmem │ Block Dim │ Grid Dim │ TensorCore │ Occupancy % │
├───────────────────┼─────────────┼────────────────┼──────────────┼──────────────┼──────────────┼──────────────┼───────────┼──────────┼────────────┼─────────────┤
│ gemm_fusion_dot_1 │           1 │    0.00000250s │  0.00000250s │  0.00000250s │  0.00000250s │    2.000 KiB │    64,1,1 │    1,1,1 │          ✗ │      100.0% │
│   loop_add_fusion │           1 │    0.00000131s │  0.00000131s │  0.00000131s │  0.00000131s │      0 bytes │    20,1,1 │    1,1,1 │          ✗ │       31.2% │
└───────────────────┴─────────────┴────────────────┴──────────────┴──────────────┴──────────────┴──────────────┴───────────┴──────────┴────────────┴─────────────┘

╔================================================================================╗
║ FRAMEWORK OP STATISTICS                                                        ║
╚================================================================================╝

┌───────────────────┬─────────┬─────────────┬─────────────┬─────────────────┬───────────────┬──────────┬───────────┬──────────────┬──────────┐
│         Operation │    Type │ Host/Device │ Occurrences │ Total Self-Time │ Avg Self-Time │ Device % │ Memory BW │    FLOP Rate │ Bound By │
├───────────────────┼─────────┼─────────────┼─────────────┼─────────────────┼───────────────┼──────────┼───────────┼──────────────┼──────────┤
│ gemm_fusion_dot.1 │ Unknown │      Device │           1 │     0.00000250s │   0.00000250s │   65.55% │ 1.82 GB/s │  1.6 GFLOP/s │      HBM │
│             +/add │     add │      Device │           1 │     0.00000131s │   0.00000131s │   34.45% │ 0.14 GB/s │ 0.05 GFLOP/s │      HBM │
└───────────────────┴─────────┴─────────────┴─────────────┴─────────────────┴───────────────┴──────────┴───────────┴──────────────┴──────────┘

╔================================================================================╗
║ SUMMARY                                                                        ║
╚================================================================================╝

AggregateProfilingResult(
    runtime = 0.00005622s, 
    compile_time = 2.32802137s,  # time spent compiling by Reactant
    GPU_0_bfc = MemoryProfileSummary(
        peak_bytes_usage_lifetime = 64.010 MiB,  # peak memory usage over the entire program (lifetime of memory allocator)
        peak_stats = MemoryAggregationStats(
            stack_reserved_bytes = 0 bytes,  # memory usage by stack reservation
            heap_allocated_bytes = 81.750 KiB,  # memory usage by heap allocation
            free_memory_bytes = 23.518 GiB,  # free memory available for allocation or reservation
            fragmentation = 0.514564,  # fragmentation of memory within [0, 1]
            peak_bytes_in_use = 81.750 KiB # The peak memory usage over the entire program
        )
        peak_stats_time = 0.00608052s, 
        memory_capacity = 23.518 GiB # memory capacity of the allocator
    )
    flops = FlopsSummary(
        Flops = 2.033375207640664e-8,  # [flops / (peak flops * program time)], capped at 1.0
        UncappedFlops = 2.033375207640664e-8, 
        RawFlops = 4060.0,  # Total FLOPs performed
        BF16Flops = 4060.0,  # Total FLOPs Normalized to the bf16 (default) devices peak bandwidth
        RawTime = 0.00005622s,  # Raw time in seconds
        RawFlopsRate = 7.220987105380169e7,  # Raw FLOPs rate in FLOPs/seconds
        BF16FlopsRate = 7.220987105380169e7,  # BF16 FLOPs rate in FLOPs/seconds
    )
)

Capturing traces

When running Reactant, it is possible to capture traces using the XLA profiler. These traces can provide information about where the XLA specific parts of program spend time during compilation or execution. Note that tracing and compilation happen on the CPU even though the final execution is aimed to run on another device such as GPU or TPU. Therefore, including tracing and compilation in a trace will create annotations on the CPU.

Let's setup a simple function which we can then profile

using Reactant

x = Reactant.to_rarray(randn(Float32, 100, 2))
W = Reactant.to_rarray(randn(Float32, 10, 100))
b = Reactant.to_rarray(randn(Float32, 10))

linear(x, W, b) = (W * x) .+ b

linear (generic function with 1 method)

The profiler can be accessed using the Reactant.with_profiler function.

Reactant.with_profiler("./") do
    mylinear = Reactant.@compile linear(x, W, b)
    mylinear(x, W, b)
end

10×2 ConcretePJRTArray{Float32,2}:
   2.76209    -7.57427
 -13.682     -10.1649
  -7.32817     6.09939
  -8.16748   -11.9265
  -1.87262     1.74517
   9.57113     7.02018
   3.58037    11.2912
   0.380071    6.90217
 -18.4023      8.64366
 -24.0053     -1.38619

Running this function should create a folder called plugins in the folder provided to Reactant.with_profiler which will contain the trace files. The traces can then be visualized in different ways.

Note

For more insights about the current state of Reactant, it is possible to fetch device information about allocations using the Reactant.XLA.allocatorstats function.

Perfetto UI

The first and easiest way to visualize a captured trace is to use the online perfetto.dev tool. Reactant.with_profiler has a keyword parameter called create_perfetto_link which will create a usable perfetto URL for the generated trace. The function will block execution until the URL has been clicked and the trace is visualized. The URL only works once.

Reactant.with_profiler("./"; create_perfetto_link=true) do
    mylinear = Reactant.@compile linear(x, W, b)
    mylinear(x, W, b)
end

Note

It is recommended to use the Chrome browser to open the perfetto URL.

Tensorboard

Another option to visualize the generated trace files is to use the tensorboard profiler plugin. The tensorboard viewer can offer more details than the timeline view such as visualization for compute graphs.

First install tensorboard and its profiler plugin:

pip install tensorboard tensorboard-plugin-profile

And then run the following in the folder where the plugins folder was generated:

tensorboard --logdir ./

Adding Custom Annotations

By default, the traces contain only information captured from within XLA. The Reactant.Profiler.annotate function can be used to annotate traces for Julia code evaluated during tracing.

Reactant.Profiler.annotate("my_annotation") do
    # Do things...
end

The added annotations will be captured in the traces and can be seen in the different viewers along with the default XLA annotations. When the profiler is not activated, then the custom annotations have no effect and can therefore always be activated.