Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mindspore-ai/akg --skill triton-ascend-memorygit clone --depth 1 https://github.com/mindspore-ai/akgWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-memory)<a href="https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-memory"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/triton-ascend-memory/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mindspore-ai/akg/triton-ascend-memory"><img src="https://agentmods.dev/badge/skills/mindspore-ai/akg/triton-ascend-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00068 | $0.00697 |
| Opus 5 | $0.00034 | $0.00349 |
| Sonnet 5 | $0.00014 | $0.00139 |
| Haiku 4.5 | $0.00007 | $0.00070 |
Grade A, and why
triton-ascend-memory scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
内存访问优化
块大小选择原理
块大小需要根据算子类型和硬件存储层级来平衡:
- VEC 类算子(element-wise、reduce、softmax 等):数据需放入 UB(192KB/VEC),
BLOCK_SIZE * sizeof(dtype)需小于 UB 可用容量,同时兼顾计算并行度。过小并行度不足,过大溢出 UB - CUBE 类算子(matmul、attention 等):左矩阵放 L0A(* KB),右矩阵放 L0B(* KB),结果放 L0C(* KB),具体参考硬件信息文档:
m0 * k0 * sizeof(A.dtype) ≤ * KB(L0A)k0 * n0 * sizeof(B.dtype) ≤ * KB(L0B)m0 * n0 * sizeof(C.dtype) ≤ * KB(L0C)
- 所有数据传输按 256 Bytes 对齐,BLOCK_SIZE 为 32 的倍数最优
2D 数据:优先 tl.make_block_ptr
A_block_ptr = tl.make_block_ptr(
base=A_ptr, shape=(M, K), strides=(stride_am, stride_ak),
offsets=(pid_m * BLOCK_M, 0), block_shape=(BLOCK_M, BLOCK_K), order=(1, 0),
)
a = tl.load(A_block_ptr, boundary_check=(0, 1))
# 移动指针
A_block_ptr = tl.advance(A_block_ptr, (0, BLOCK_K))
连续内存:一维访问
非连续张量先 .contiguous() 转换,再用一维 ptr + offsets 访问:
class ModelNew(torch.nn.Module):
def forward(self, x):
if not x.is_contiguous():
x = x.contiguous()
out = torch.empty_like(x)
n = x.numel()
grid = (triton.cdiv(n, BLOCK_SIZE),)
kernel[grid](x, out, n, BLOCK_SIZE=1024)
return out
一维访问比 stride 计算效率更高,推荐优先使用。
对齐要求
- Ascend 256B 对齐: element-wise / reduce 算子
- Ascend 512B 对齐: MatMul 切分
- 数据搬运带宽上限约 256*256B,据此设计搬运策略
要点
- 优先
.contiguous()+ 一维访问 - 连续内存访问效率远高于 stride 计算开销
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 63 lines · 68 tokens per session scan A 97696bd43dab
triton-ascend-memory is a skill published in the GitHub repository mindspore-ai/akg (259 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 68 tokens to every session and 697 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
spark-environment-setup
Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). Use when installing PyTorch/Unsloth/TRL/vLLM on DGX Spark, hitting libcudart or wheel-ABI errors on aarch64, or choosing between NGC containers and bare pip installs.
spark-memory-thermal-ops
Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.
spark-training-gotchas
Preflight and diagnose the ten known failure modes for ML training on NVIDIA DGX Spark. Use when a training run on DGX Spark fails to start, OOMs below the 128GB limit, slows down mid-run, or before any multi-hour training job on GB10.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
minicpm5-deploy-vllm-ascend
Deploy MiniCPM5-2B with vLLM on Huawei Ascend NPU using vLLM-Ascend. Use when the user mentions vLLM-Ascend, Ascend NPU, Huawei Ascend, CANN, torchnpu, davinci devices, or wants an OpenAI-compatible MiniCPM5 server on Ascend hardware.
minicpm5-deploy-litert
Run MiniCPM5-2B or MiniCPM5-1B on-device with Google's LiteRT-LM runtime — the litert-lm CLI or its OpenAI-compatible server on a desktop, the Kotlin API or the AI Edge Gallery app on Android, the same .litertlm bundle on CPU or GPU. Use when the user says "LiteRT", "LiteRT-LM", "litertlm", ".litertlm", "Android"…