Fix save_fields usage in README and disable dataset viewer
Browse files
README.md
CHANGED
|
@@ -1,3 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# NKIBench
|
| 2 |
|
| 3 |
NKIBench is a benchmark of AWS [Neuron Kernel Interface (NKI)](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/general/nki/index.html) kernels paired with NumPy reference implementations. Each task provides a specification, a ground-truth NumPy forward pass, and an optimized NKI kernel targeting AWS Trainium / Inferentia devices, together with tooling to compile, check numerical correctness, and measure on-device latency.
|
|
@@ -10,7 +27,9 @@ NKIBench/
|
|
| 10 |
├── reference/ # NumPy reference implementations with concrete shapes
|
| 11 |
├── kernels/ # Optimized NKI kernels (one per case)
|
| 12 |
├── summary.json # Index mapping task → case → {seed, reference, kernel}
|
|
|
|
| 13 |
└── kernel_wrapper.py # Profiler: compile, correctness check, latency benchmark
|
|
|
|
| 14 |
```
|
| 15 |
|
| 16 |
### `summary.json`
|
|
@@ -101,7 +120,7 @@ save_fields = json.load(open("save_fields.json"))
|
|
| 101 |
case = summary["matmul"]["cases"]["3"]["impls"][0]
|
| 102 |
|
| 103 |
k = NKIKernel(program_path=case["kernel"], base_numpy_path=case["task"])
|
| 104 |
-
result = k.profile(save_fields=)
|
| 105 |
|
| 106 |
print("compiled:", result.compiled)
|
| 107 |
print("correct :", result.correct)
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- nki
|
| 7 |
+
- aws-neuron
|
| 8 |
+
- trainium
|
| 9 |
+
- kernel
|
| 10 |
+
- benchmark
|
| 11 |
+
- agent
|
| 12 |
+
pretty_name: NKIBench
|
| 13 |
+
size_categories:
|
| 14 |
+
- n<1K
|
| 15 |
+
viewer: false
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
# NKIBench
|
| 19 |
|
| 20 |
NKIBench is a benchmark of AWS [Neuron Kernel Interface (NKI)](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/general/nki/index.html) kernels paired with NumPy reference implementations. Each task provides a specification, a ground-truth NumPy forward pass, and an optimized NKI kernel targeting AWS Trainium / Inferentia devices, together with tooling to compile, check numerical correctness, and measure on-device latency.
|
|
|
|
| 27 |
├── reference/ # NumPy reference implementations with concrete shapes
|
| 28 |
├── kernels/ # Optimized NKI kernels (one per case)
|
| 29 |
├── summary.json # Index mapping task → case → {seed, reference, kernel}
|
| 30 |
+
|-- save_fields.json
|
| 31 |
└── kernel_wrapper.py # Profiler: compile, correctness check, latency benchmark
|
| 32 |
+
|
| 33 |
```
|
| 34 |
|
| 35 |
### `summary.json`
|
|
|
|
| 120 |
case = summary["matmul"]["cases"]["3"]["impls"][0]
|
| 121 |
|
| 122 |
k = NKIKernel(program_path=case["kernel"], base_numpy_path=case["task"])
|
| 123 |
+
result = k.profile(save_fields=save_fields)
|
| 124 |
|
| 125 |
print("compiled:", result.compiled)
|
| 126 |
print("correct :", result.correct)
|