external-gitcode-ascend-triton-operator-precision-eval

Triton Operator Precision Evaluation

Core Principles

Precision is the bottom line for operator correctness. Do not perform performance optimization before verification passes.

Workflow

1. Write the Torch reference implementation (baseline)
2. Write test cases (covering multiple shapes × dtypes × edge cases)
3. Execute the Triton operator and the Torch reference implementation on the NPU
4. Use test_common.validate_cmp() to compare and generate a precision report

MANDATORY - READ ENTIRE FILE: Before writing tests, read in full scripts/test_common.py。

Error Thresholds

Criteria:MERE < threshold and MARE < 10 × threshold

Data Type	Threshold	MERE Upper Limit	MARE Upper Limit
float16	2⁻¹⁰ ≈ 9.77e-4	9.77e-4	9.77e-3
float32	2⁻¹³ ≈ 1.22e-4	1.22e-4	1.22e-3
bfloat16	2⁻⁷ ≈ 7.81e-3	7.81e-3	7.81e-2
int8/uint8/int16/uint16/int32/uint32/int64/uint64	Exactly equal	—	—
bool	Exactly equal	—	—

Where:

• MERE = mean(|y_cal - y_ref| / |y_ref|), i.e., mean relative error
• MARE = max(|y_cal - y_ref| / |y_ref|), i.e., maximum relative error

The report must include

• Verification configuration: operator name, test shapes, dtype, number of cores
• Precision standards: thresholds and criteria for all dtypes (MERE < threshold and MARE < 10 × threshold)
• Verification results: total passed/failed
• Error metrics: MERE (mean relative error), MARE (maximum relative error)
• Criteria

Output Conventions

• The precision report is written to the operator directory under precision_report.md(Markdown format)
• The precision evaluation script is written to the operator directory under precision_eval.py
• Refer to examples/ the example scripts and report format in the directory

Anti-Pattern Checklist (NEVER)

• ❌ Perform precision verification without providing a Torch reference implementation
• ❌ Use the wrong threshold (e.g., using the FP32 threshold for FP16)
• ❌ Not use the MERE/MARE criteria (using only rtol/atol is not strict enough)
• ❌ Do not increase precision to FP32 for reduction operations
• ❌ Test only one dtype/shape and assert correctness
• ❌ Skip edge cases (non-aligned dimensions)
• ❌ Perform performance optimization before precision passes

Checklist

• The Torch reference implementation is correct and verified on the NPU
• Covers multiple shapes × dtypes
• Use FP32 for reduction operations
• Generate a standardized precision report