Release v13.3 of the CUDA samples with CUDA 13.3 Toolkit (#435)

This is the release of the CUDA 13.3 samples, which include additions for CUDA Tile C++, and updated CCCL and Python samples.
This commit is contained in:
Dheemanth
2026-05-27 14:50:59 -07:00
committed by GitHub
parent aeab82ff30
commit b7c5481c55
150 changed files with 7977 additions and 246 deletions

View File

@@ -90,6 +90,15 @@ elapsed_ms = (end_event - start_event) / n_iterations
- Python 3.10 or newer
- See `requirements.txt` for Python packages
### Platform Support:
The benchmark loops in this sample read kernel results back from
`ManagedMemoryResource` allocations between launches, which requires the
device property `concurrent_managed_access=True`. This is only supported on
Linux with HMM (Pascal and newer). On Windows (WDDM/MCDM/TCC) the property
is `False`, so the sample exits early with a waive message and exit code
`2`.
## Installation
```bash
@@ -115,8 +124,8 @@ Compute Capability: X.X
Compiling CUDA kernels with cuda.core.Program...
Target architecture: sm_XX
✓ vector_add kernel compiled
✓ reduce_sum kernel compiled
[OK] vector_add kernel compiled
[OK] reduce_sum kernel compiled
============================================================
VECTOR ADDITION - Launch Configuration Tuning
@@ -132,11 +141,11 @@ Block Size: 64 | Blocks: 156250 | Time: X.XXXX ± X.XXXX ms
...
------------------------------------------------------------
✓ OPTIMAL: block_size=XXX (X.XXXX ms)
✗ WORST: block_size=XXX (X.XXXX ms)
[OK] OPTIMAL: block_size=XXX (X.XXXX ms)
[FAIL] WORST: block_size=XXX (X.XXXX ms)
Speedup: X.XXx
✓ Results verified correct!
[OK] Results verified correct!
...