| name | triton-ascend-grid-config |
| description | Grid/Block 配置策略,包括核数选择、并行度调优、二次切分和大 shape 算子处理方案。适用于需要确定 kernel 启动参数、优化多核并行效率、或处理超大规模数据的内核代码生成场景 |
| category | fundamental |
| version | 1.0.0 |
| metadata | {"backend":"ascend","dsl":"triton_ascend","hardware":"Atlas A2, Atlas A3"} |
Grid 配置策略
Grid 限制
- Grid 必须是 tuple,最多 3 维:
(x,), (x, y), (x, y, z)
- 各维度乘积不超过 65535
- BLOCK_SIZE 必须小于 65536
推荐方案:交错循环(固定 Grid 为核心数)
适用于按行/按块独立处理的算子(Element-wise、Reduce、Normalization 等)。
@triton.jit
def kernel(
input_ptr, output_ptr, M, N,
stride_m, stride_n,
BLOCK_N: tl.constexpr,
CORE_NUM: tl.constexpr,
):
pid = tl.program_id(0)
for row_idx in range(pid, M, CORE_NUM):
row_ptr = input_ptr + row_idx * stride_m
out_ptr = output_ptr + row_idx * stride_m
for col_start in range(0, N, BLOCK_N):
offs = col_start + tl.arange(0, BLOCK_N)
mask = offs < N
data = tl.load(row_ptr + offs * stride_n, mask=mask)
result = compute(data)
tl.store(out_ptr + offs * stride_n, result, mask=mask)
动态获取核心数
必须在 __init__ 中获取,禁止在 forward 中调用(触发设备同步)。
import torch
import triton
class ModelNew(torch.nn.Module):
def __init__(self):
super().__init__()
try:
device = torch.npu.current_device()
properties = triton.runtime.driver.active.utils.get_device_properties(device)
self.VEC_CORE_NUM = properties.get("num_vectorcore", 40)
self.CUBE_CORE_NUM = properties.get("num_aicore", 20)
except:
self.VEC_CORE_NUM = 40
self.CUBE_CORE_NUM = 20
def forward(self, x):
M, N = x.shape
out = torch.empty_like(x)
grid = (self.VEC_CORE_NUM,)
kernel[grid](x, out, M, N, x.stride(0), x.stride(1),
BLOCK_N=256, CORE_NUM=self.VEC_CORE_NUM)
return out
核心数选择
- 向量算子(element-wise、softmax、归一化):使用
VEC_CORE_NUM
- 矩阵算子(matmul、attention):使用
CUBE_CORE_NUM
多次切分策略
若 BLOCK_SIZE 超限或单次切分超硬件缓存,可嵌套循环做多层切分:
for m_start in range(pid_m * BLOCK_M, min((pid_m + 1) * BLOCK_M, M), SUB_BLOCK_M):
for n_start in range(pid_n * BLOCK_N, min((pid_n + 1) * BLOCK_N, N), SUB_BLOCK_N):