trace.mark GPU Virtual ISA NVIDIA
Miscellaneous Instructions
trace.mark{.payload_size} event-id {, payload} {, event-label};
trace.mark instruction creates and marks an event for performance monitoring.
Encoding
PTX is a virtual instruction set. It has no single, stable native
binary encoding - the compiler lowers this instruction to different native machine
code depending on the selected NVIDIA target architecture (compute capability).
This page intentionally shows no bit-diagram; see the target/version requirements
below for what governs how this instruction compiles.
Syntax Forms
One mnemonic covers many type / state-space / scope / modifier combinations - each row below is an independently valid form.
| Syntax | Data Types | State Space(s) | Modifiers | Min. Target | Description |
|---|---|---|---|---|---|
| trace.mark{.payload_size} event-id {, payload} {, event-label}; | sm_107f | trace.mark instruction creates and marks an event for performance monitoring. (see the official PTX ISA docs for the full description) |
Operands
At a Glance
Related
More in Miscellaneous Instructions
Reference
NVIDIA PTX ISA
Description
trace.mark instruction creates and marks an event for performance monitoring.
The resulting event record consists of two components: timestamp and event information;
where event information is formed by concatenation of event-id and payload of size
as listed in the table below:
Event information Size in bits event-id 12 payload 32 or 64
The size of the resulting event information can be 12-bits, 44-bits or otherwise 76 bits. Since
the payload operand is optional, its size is crucial in determining the overall size of the
resulting event information.
The mandatory operand event-id represents static event identifier corresponding to the
instruction instance. (see the official PTX ISA docs for the full description)
Examples
.metadata_section {
.metadata_domain 1
.metadata 1 { "kernel: event1", "init portion", .b32 1 }
.metadata 2 { "kernel: event2", "compute intense portion" }
.metadata 3 { "kernel: event3", "epilogue portion", .b32 3, .b32 8, .b32 64 }
.metadata 4 { "associate_label1_event1", .b32 0, .metadata_label "event_label1", .metadata_index 1 }
// (truncated - see the official PTX ISA docs for the full example)Reproduced from NVIDIA's official PTX ISA documentation for technical accuracy.
Sources
-
Parallel Thread Execution ISA ↗
- NVIDIA Corporation, Chapter 9 - Instruction Set
Deep-linked directly to this instruction's section.