Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV/Parquet I/O, nullable semantics, and multi-GPU DataFrame workloads.
中文简介正在准备中。当前显示原文说明。
Use this skill when the user has an existing DPDK application and is adding DOCA capabilities in-place — most commonly DOCA Flow hardware steering — without rewriting the data-plane in DOCA-native form: binding a DPDK port id to a `doca_dev` (`doca_dpdk_port_probe` / `doca_dpdk_port_as_dev`), converting `rte_mbuf` ↔ `doca_buf`, querying `doca_dpdk_cap_is_rep_port_supported`, or debugging `DOCA_ERROR_*` from a bridge call. Trigger even without "DOCA DPDK Bridge": "how do I add DOCA Flow to my DPDK app", "make a DPDK port visible to DOCA", "the bridge loads but every operation returns errors", "pkg-config --exists doca-dpdk-bridge fails", or "DOCA_ERROR_NOT_FOUND on port registration". Route elsewhere for fresh DOCA-native packet I/O (doca-eth), flow-rule programming (doca-flow), DOCA or DPDK install (doca-setup), or RDMA data movement (doca-rdma).
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
doca_dpdk_port_probe; do not
rewrite the data-plane as native doca-eth.Follow Loading order below. Open the tasks companion for configure / probe / debug. Open the capabilities companion for port binding and mbuf conversion. Do not open every companion file.
Where to start: This skill assumes DOCA is already installed,
DPDK is already installed, the user has an existing DPDK
application, and they want to add DOCA capabilities to it
in-place (most commonly DOCA Flow for hardware steering)
without migrating the data-plane to DOCA-native APIs. Open
the tasks companion if the user wants to do something
(configure / build / modify / run / test / debug); open
the capabilities companion when the question is what
the bridge can express on this version. If the user has not
installed DOCA yet, route to
doca-setup first. If the user is
starting fresh (no DPDK code yet) and just wants line-rate
packet I/O against DOCA, route to
doca-eth instead — the bridge exists
for the interop case, not the start-fresh case.
How to use this skill. Read this file, then search the companion file
for the ## <name> heading the question needs and read that whole section.
Answer completely from it, quoting only its symbols and commands. Do not
browse docs; use the entry points in doca-public-knowledge-map. Give
environment checks as commands for the user's DOCA host. For out-of-scope
requests, name the skill that owns them and stop.
The CLASSES of DOCA DPDK Bridge questions this skill is built to answer, each with one worked example. The agent should treat the class as the load-bearing piece — the worked example is a single instance.
rte_eth_*
ports; I want to install DOCA Flow steering rules on those
ports for HW offload". Answered by the bridge-vs-native
selection rule in
rte_eth_dev_*; how does
DOCA see it". Answered by the DPDK-port-id ↔ doca_dev
mapping (doca_dpdk_port_probe / doca_dpdk_port_as_dev) in
rte_mbuf to my fastpath; I
want a DOCA library to operate on the payload". Answered by
the mbuf ↔ DOCA-buf conversion shape in
pkg-config --exists doca-dpdk-bridge returns failure on a host that has DPDK
separately installed". Answered by the version-coupling rule
in
doca-dpdk-bridge or doca-eth?" —
worked example: "new project; I have not committed to DPDK
yet". Answered by the path-selection table in
DOCA_ERROR_* from a bridge call mean and
which layer caused it?" — worked example: "DOCA_ERROR_NOT_FOUND
on a port-registration call after the DPDK port came up
cleanly". Answered by the bridge overlay on the cross-library
taxonomy in
doca-debug.This skill serves external developers who already maintain a
DPDK-based packet-processing application — i.e., users whose
data-plane already drives rte_eth_* ports and rte_mbuf
buffers — and who want to add DOCA capabilities (most
commonly hardware steering via DOCA Flow) by linking the bridge
into the same process. It is not for users starting a fresh
DOCA-native project (route to doca-eth)
and not for users running pure DPDK with no interest in DOCA
(no skill in this bundle applies). It is also not for NVIDIA
developers contributing to DOCA DPDK Bridge itself.
Language scope. DOCA DPDK Bridge ships as a C library with
pkg-config module name doca-dpdk-bridge (the agent must
confirm the spelling against the user's install via
pkg-config --exists doca-dpdk-bridge; some DOCA releases use a
slightly different module name and the agent must not guess).
The shipped samples are written in C. C and C++ consumers are
the canonical case; the worked examples in the tasks companion assume
that path. Other-language consumers (Rust, Go, Python, …)
typically do not use this bridge — DPDK itself is C-shaped, and
a non-C DPDK app is rare; if the user is in that minority, the
skill's contribution is to keep the port-handover, capability,
permission, and error-taxonomy guidance language-neutral and
route them to the public C ABI as the authoritative surface.
Load this skill when the user is doing hands-on DOCA DPDK Bridge work, in any language. Concretely:
rte_eth_dev_* ports and rte_mbuf buffers),
and they want to layer DOCA on top — adding DOCA Flow rules,
feeding packets into a DOCA accelerator (Compress, AES-GCM, …),
or wiring a DOCA service into the same process.doca_dev (via
doca_dpdk_port_probe / doca_dpdk_port_as_dev) so DOCA Core
/ DOCA Flow can operate on the same physical port.DOCA_ERROR_* returned from a bridge
call (lifecycle vs. permission vs. capability vs. DPDK-port-not-
registered vs. mbuf-conversion-failed) and the layered DPDK ↔
DOCA stack underneath.doca-eth or doca-dpdk-bridge?"
and the agent needs the path-selection rule.Do not load this skill for general DOCA orientation, install
of DOCA itself, fresh DOCA-native packet I/O (use
doca-eth), flow-rule programming on
its own (use doca-flow), host ↔ DPU
control messaging (use doca-comch),
or RDMA data movement (use doca-rdma).
For DOCA documentation orientation, use
doca-public-knowledge-map.
This is a thin loader. The body keeps only the orientation needed to pick the right next file. The substantive bridge- specific material lives in two companion files:
doca-dpdk-bridge vs native doca-eth), the
bridge-object surface (DPDK-port-id ↔ doca_dev mapping, the
doca_dpdk_mempool mbuf ↔ DOCA-buf conversion), the
capability-query surface (the single
doca_dpdk_cap_is_rep_port_supported), the bridge error taxonomy
(mapped onto the cross-library DOCA_ERROR_* set), the
observability surface (DPDK-side counters + DOCA-side PE
events), and the safety policy that gates the matched-pair
DPDK ↔ DOCA version coupling, the EAL-must-be-up
precondition, and DPDK-side privileges.configure, build, modify, run, test,
debug. Plus a Deferred task verbs block that points
out-of-scope questions at the right next skill, and a
Command appendix of the recurring commands the agent
reaches for.The skill assumes a host or BlueField where DOCA AND DPDK are
already installed at the standard locations and the user has the
privileges their public install profile expects (typically sudo
or mlnx-group membership for DOCA, plus the DPDK-side
privileges to mount hugepages, bind PCIe ports, and open the
target rte_eth device). It does not cover installing DOCA —
that path goes through doca-setup;
it does not cover installing or learning DPDK itself — that
belongs in upstream DPDK documentation reachable via
doca-public-knowledge-map.
This skill is agent guidance, not a samples or templates bundle. To keep the boundary clean, it deliberately does not contain — and pull requests should not add:
/opt/mellanox/doca/samples/doca_flow/ (e.g. flow_common.c,
which calls doca_dpdk_port_probe() / doca_dpdk_port_as_dev()),
plus the canonical reference applications that pair DPDK +
DOCA Flow (see
doca-public-knowledge-map
for the reference-applications index). The agent's job is to
route the user to those files and prescribe a minimum-diff
modification on them via the universal modify-a-sample
workflow in
doca-programming-guide,
layered with the bridge-specific overrides in
meson.build,
CMakeLists.txt, Cargo.toml, …) parked inside the skill.
The agent constructs the build manifest in the user's project
directory against the user's installed DOCA + DPDK, where
pkg-config --modversion doca-dpdk-bridge and
pkg-config --modversion libdpdk are the source of truth.samples/, bindings/, or reference/ subtree of any
kind. A mock or incomplete artifact in this skill's tree, even
one labeled "reference", is misleading: users will read it as
buildable.Both companion files cross-link to each other,
doca-version for the canonical
DOCA version-handling rules,
doca-eth for the start-fresh path that
is the bridge's deliberate alternative,
doca-flow for the steering library
that is the most common reason to use this bridge in the first
place, and
doca-public-knowledge-map
whenever the right answer is "look it up in the public docs or
the installed package layout" rather than "bridge-specific
guidance".
doca-public-knowledge-map —
the routing table for every public DOCA documentation source
and the on-disk layout of an installed DOCA package. The
bridge URL slug is DOCA-DPDK-Bridge. Reference applications
that pair DPDK + DOCA Flow (the canonical adopters of this
bridge) are reachable from the same map's reference-
applications index.doca-setup — env preparation,
install verification, hugepage mounts (DPDK requirement),
port-state checks (devlink dev show, ip link), permission
and group-membership requirements for opening a doca_dev
AND for binding PCIe ports under DPDK. This skill assumes its
preconditions are satisfied.doca-version — canonical
DOCA version-handling rules. This skill's ## Version compatibility cross-links the four-way match rule and adds
only the bridge-specific overlay (DPDK ↔ DOCA matched-pair
coupling).doca-structured-tools-contract —
the bundle's structured-tools precedence rule (detect / prefer
/ fall back / report). The Command appendix in
the tasks companion honors this contract.doca-programming-guide —
general DOCA programming patterns shared by every library:
the canonical pkg-config + meson build pattern, the
universal modify-a-shipped-sample first-app workflow, the
universal lifecycle, the cross-library DOCA_ERROR_*
taxonomy, and the program-side debug order. This skill layers
bridge specifics on top.doca-eth — the alternative path
for users starting fresh: pure DOCA-native queue I/O via
doca_eth_rxq / doca_eth_txq with no DPDK in the picture.
The bridge exists because migrating an existing DPDK app to
doca-eth is often the wrong tradeoff; for a new project
with no DPDK lock-in, doca-eth is the right choice.doca-flow — the steering library
that is the most common reason a DPDK app reaches for this
bridge. The bridge binds a DPDK port id to a doca_dev;
DOCA Flow programs steering rules onto that doca_dev. The two
libraries are designed to compose.doca-debug — the cross-cutting
debug ladder (install / version / build / link / runtime /
program / driver). Bridge-specific debug (DPDK ↔ DOCA version
drift, port-not-registered symptoms, mbuf-conversion failures)
overlays on top of that ladder.还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV/Parquet I/O, nullable semantics, and multi-GPU DataFrame workloads.
中文简介正在准备中。当前显示原文说明。
Customize NVIDIA Nemotron Voice Agent's Generic Pipecat example for healthcare appointment, five-field patient intake, or custom tool-calling workflows without a separate backend.
中文简介正在准备中。当前显示原文说明。
Calibrate a new dataset from live RTSP camera streams via the AutoMagicCalib REST API. Use when the user provides RTSP URLs or asks to calibrate live cameras; VIOS records clips, AMC ingests them, then runs calibration.
中文简介正在准备中。当前显示原文说明。
Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.
中文简介正在准备中。当前显示原文说明。
Calibrates pre-recorded `cam_*.mp4` datasets through the AutoMagicCalib REST API. Use for user-supplied local MP4s; route live RTSP streams to `amc-run-rtsp-calibration`.
中文简介正在准备中。当前显示原文说明。
Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.
中文简介正在准备中。当前显示原文说明。