https://github.com/vegaluisjose/neptune

https://github.com/vegaluisjose/neptune

Last synced: 4 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/vegaluisjose/neptune
• Owner: vegaluisjose
• Created: 2024-09-17T14:26:18.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2024-09-21T20:22:07.000Z (4 months ago)
• Last Synced: 2024-10-04T16:12:03.237Z (4 months ago)
• Language: Python
• Size: 11.7 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Neptune

## Getting started

* Install Triton, currently using `3.0.0`

```bash

python3 -m pip install triton

```

## Examples

### Vector addition

* Run

```bash

python3 vadd/test.py

```

* Output

```bash

Elapsed time (seconds): 0.45724202299606986

```

### Autotuning Vector addition

* Run

```bash

TRITON_PRINT_AUTOTUNING=1 python3 vadd_autotune/test.py

```

* Output

```bash

Triton autotuning for function add_kernel finished after 0.77s; best config selected: BLOCK_SIZE: 256, num_warps: 8, num_ctas: 1, num_stages: 2, maxnreg: None;

Elapsed time (seconds): 0.774298002012074

```

### Autotuning Matrix Multiplication

* Run

```bash

TRITON_PRINT_AUTOTUNING=1 python3 matmul_autotune/test.py

```

* Output

```bash

CUDA Compute Capability: (8, 9)

Number of SMs: 128

CUDA Capability Major/Minor version number: 8.9

Triton autotuning for function _kernel finished after 3.21s; best config selected: BLOCK_M: 32, BLOCK_N: 32, BLOCK_K: 32, SPLIT_K: 1, num_warps: 2, num_ctas: 1, num_stages: 6, maxnreg: None;

Elapsed time (seconds): 3.4771489950071555

```