Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions projects/hipblaslt/docs/index.rst
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,7 @@ The hipBLASLt public repository is located at `<https://github.com/ROCm/hipBLASL

* :doc:`Data type support <./reference/data-type-support>`
* :doc:`hipBLASLt datatypes <./reference/datatypes>`
* :doc:`hipBLASLt environment variables <./reference/env-variables>`
* :doc:`hipBLASLt API <./reference/api-reference>`
* :doc:`hipBLASLtExt API <./reference/ext-reference>`
* :doc:`hipBLASLtExt operation API <./reference/ext-ops>`
Expand Down
119 changes: 119 additions & 0 deletions projects/hipblaslt/docs/reference/env-variables.rst
Original file line number Diff line number Diff line change
@@ -0,0 +1,119 @@
.. meta::
:description: hipBLASLt environment variables reference
:keywords: hipBLASLt, ROCm, API, environment variables, environment, reference

.. _environment-variables:

********************************************************************
hipBLASLt environment variables
********************************************************************

This section describes the important hipBLASLt environment variables,
which are grouped by functionality.

Logging and debugging
=====================

The logging and debugging environment variables for hipBLASLt are collected in the following table.
For more information, see :doc:`Use logging and heuristics <../how-to/use-logging-heuristics>`.

.. list-table::
:header-rows: 1
:widths: 70,30

* - **Environment variable**
- **Value**

* - | ``HIPBLASLT_LOG_LEVEL``
| Controls the verbosity level of hipBLASLt logging output.
- | 0: Off (logging disabled, default)
| 1: Error (only errors are logged)
| 2: Trace (API calls with kernel launches log parameters)
| 3: Hints (performance improvement suggestions)
| 4: Info (general library execution information)
| 5: API trace (detailed API call parameters)

* - | ``HIPBLASLT_LOG_MASK``
| Controls logging output using bit mask flags (can be combined).
- | 0: Off
| 1: Error
| 2: Trace
| 4: Hints
| 8: Info
| 16: API trace
| 32: Bench
| 64: Profile
| 128: Extended profile

* - | ``HIPBLASLT_LOG_FILE``
| Specifies path to logging file. Can contain ``%i`` for process ID replacement.
- | Path to log file (for example, ``logfile_%i.log``)
| If not defined: log messages printed to stdout

* - | ``HIPBLASLT_ENABLE_MARKER``
| Enables marker trace for ROCProfiler profiling.
- | 0 or unset: Disable marker trace
| 1: Enable marker trace


Offline tuning
===============

The offline tuning environment variables for hipBLASLt are collected in the following table.
For more information, see :doc:`Use hipBLASLt offline tuning <../how-to/how-to-use-hipblaslt-offline-tuning>`.

.. list-table::
:header-rows: 1
:widths: 70,30

* - **Environment variable**
- **Value**

* - | ``HIPBLASLT_TUNING_FILE``
| Specifies file to store tuning results with best solution indices for GEMM problems.
- | Path to tuning file (for example, ``tuning.txt``)
| File stores optimal kernel indices for reuse

* - | ``HIPBLASLT_TUNING_OVERRIDE_FILE``
| Specifies file to load tuning results and override default kernel selection.
- | Path to tuning file (for example, ``tuning.txt``)
| Loads previously saved optimal kernel choices

* - | ``HIPBLASLT_TUNING_USER_MAX_WORKSPACE``
| Sets maximum workspace size constraint during tuning stage.
- | Integer value in bytes (default: 128 * 1024 * 1024)
| Limits workspace size for solution selection

Stream-K configuration
======================

The Stream-K configuration environment variables for hipBLASLt are collected in the following table.
For more information, see :doc:`Use Stream-K with hipBLASLt <../how-to/how-to-use-streamk>`.

.. list-table::
:header-rows: 1
:widths: 70,30

* - **Environment variable**
- **Value**

* - | ``TENSILE_SOLUTION_SELECTION_METHOD``
| Controls hipBLASLt kernel selection strategy for GEMM operations.
- | 0: Default (standard tuned libraries, no Stream-K)
| 2: Stream-K (enables Stream-K library for consistent performance)

* - | ``TENSILE_STREAMK_DYNAMIC_GRID``
| Controls Stream-K dynamic grid size selection behavior.
- | 0: Disable dynamic grid (use all available compute units)
| 3: Default (automatically pick optimal workgroup count)

* - | ``TENSILE_STREAMK_FIXED_GRID``
| Overrides default grid size with specified number of workgroups for Stream-K kernels.
- | Integer value specifying number of workgroups
| Example: 64 (limits GEMM kernels to 64 workgroups)

* - | ``TENSILE_STREAMK_MAX_CUS``
| Sets maximum number of compute units for Stream-K kernels.
- | Integer value specifying maximum compute units
| Example: 32 (limits GEMM kernels to 32 compute units)
| Default: All available compute units
2 changes: 2 additions & 0 deletions projects/hipblaslt/docs/sphinx/_toc.yml.in
Original file line number Diff line number Diff line change
Expand Up @@ -47,6 +47,8 @@ subtrees:
- file: reference/data-type-support
- file: reference/datatypes
title: hipBLASLt datatypes
- file: reference/env-variables
title: hipBLASLt environment variables
- file: reference/api-reference
title: hipBLASLt API
- file: reference/ext-reference
Expand Down
Loading