DOCA Compress
Examples
- "Offload this gzip on the BlueField." — this is in scope if
the codec is DEFLATE (or LZ4 decompress). Do not invent LZ4
encode or zstd offload.
- "Compress task returns INVALID_VALUE on alloc_init." — size
the
doca_buf against the per-task cap query before changing
the algorithm.
Instructions
Follow Loading order below. Open the tasks companion for configure /
test / debug. Open the capabilities companion for codec support
and buffer-size policy. Do not open every companion file.
Where to start: This skill assumes DOCA is already installed and
the user is doing hands-on Compress work (bulk DEFLATE
compression or decompression) on a BlueField / ConnectX / host
with DOCA. 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 can
DOCA Compress express on this version. If the user has not
installed DOCA yet, route to
doca-setup/SKILL.md first. If the user is
asking "should I even offload this compression to the
accelerator?", the size-threshold path-selection rule in
Capabilities and modes
is the first stop — bulk compress is the canonical fit, tiny
one-shot is not.
Example questions this skill answers well
The CLASSES of DOCA Compress 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.
- "Should I offload this compression to DOCA Compress, or just
zlib/zstd on the CPU?" — worked example: "I am compressing a
4 MiB log buffer before writing it to storage; is doca-compress
worth the setup vs zlib on the CPU?". Answered by the
size-threshold path-selection rule in
Capabilities and modes
- the "when NOT to use doca-compress" bullets in
Safety policy.
- "Does my device support the compress (or decompress) task I
want, and how big a buffer can it move per submission?" —
worked example: "is
doca_compress_task_compress_deflate on
this BlueField, and what is the max source size per task?".
Answered by the per-task capability-query rule
(doca_compress_cap_task_compress_deflate_is_supported,
_decompress_deflate_is_supported, the matching
_get_max_buf_size queries) inCapabilities and modes
configure.
- "I just need to decompress incoming network data — is that a
valid standalone use of doca-compress?" — worked example:
"my client receives DEFLATE-compressed payloads and never
produces any compressed output of its own". Answered by the
"decompression-only" note in
Capabilities and modes
task-type table + the per-task configuration matrix inconfigure step 5.
- "What permissions does the source / destination mmap need?" —
worked example: "my
doca_compress_task_compress_deflate
returns DOCA_ERROR_NOT_PERMITTED". Answered by the permission
matrix inSafety policy
- the mmap-set-permissions checklist in
test.
- "Is this DOCA Compress API available on my installed DOCA
version?" — worked example: "is
doca_compress_task_decompress_deflate in the DOCA I have
installed?". Answered by the version-compatibility overlay inVersion compatibility,
which cross-links the canonical detection chain in
doca-version/SKILL.md and adds the
Compress-specific "discover per-task support, do not assume"
bullets.
- "What does this
DOCA_ERROR_* from a Compress call mean and
which layer caused it?" — worked example: "DOCA_ERROR_INVALID_VALUE
on doca_compress_task_compress_deflate_alloc_init". Answered
by the Compress overlay on the cross-library taxonomy inError taxonomy
debug that escalates to
doca-debug/SKILL.md.
Audience
This skill serves external developers building applications that
consume the DOCA Compress library — i.e., users whose code calls
doca_compress_* (directly in C/C++, or through FFI/bindings from
another language) to offload bulk DEFLATE compression or
decompression onto a BlueField DPU or ConnectX accelerator. It is
not for NVIDIA developers contributing to DOCA Compress itself.
Language scope. DOCA Compress ships as a C library with
pkg-config module name doca-compress. The shipped samples are
written in C. C and C++ consumers are the canonical case and the
worked examples in the tasks companion assume that path. Other-language
consumers (Rust, Go, Python, …) consume the same *.so through
FFI or language-specific bindings; the skill's contribution in
that case is to keep the lifecycle, capability-discovery,
permission, error-taxonomy, and compress-vs-decompress guidance
language-neutral, and to route the agent to the public C ABI as
the authoritative surface that any wrapper will eventually call.
When to load this skill
Load this skill when the user is doing hands-on DOCA Compress
work, in any language. Concretely:
- Initializing a
doca_compress context on a doca_dev and
configuring at least one task type
(doca_compress_task_compress_deflate and/or
doca_compress_task_decompress_deflate) before
doca_ctx_start().
- Choosing between the compress task and the decompress
task for the user's data flow. The two task types are
independent: a consumer that only decompresses inbound DEFLATE
payloads is a valid shape and does not need to enable the
compress task.
- Setting permissions on
doca_mmap correctly for the source
buffer (DOCA_ACCESS_FLAG_LOCAL_READ_ONLY at minimum) and the
destination buffer (DOCA_ACCESS_FLAG_LOCAL_READ_WRITE).
- Sizing the source buffer against the per-task
doca_compress_cap_task_*_get_max_buf_size(devinfo) ceiling
and the destination buffer against the worst-case output size
the algorithm can produce on this input.
- Checking which task types this device's accelerator advertises
via
doca_compress_cap_task_compress_deflate_is_supported and
doca_compress_cap_task_decompress_deflate_is_supported
against the active doca_devinfo.
- Validating against a round-trip (compress → decompress, or
decompress → compare against a known DEFLATE-encoded fixture)
before pushing bulk input through the accelerator.
- Deciding whether to offload at all — the size-threshold rule
in ## Capabilities and modes
says doca-compress is the right answer only for bulk inputs
(rule of thumb: ≥ a few KiB); below that, CPU compression beats
the DMA-to-accelerator round-trip.
- Debugging a
DOCA_ERROR_* returned from a Compress call
(lifecycle vs. buffer-sizing vs. permission vs.
unsupported-task) and the task-completion event on the progress
engine.
- Designing or extending non-C bindings (Rust, Go, Python, …)
that wrap the Compress C ABI — for the lifecycle, permission,
capability, and compress-vs-decompress rules the wrapper must
honor.
Do not load this skill for general DOCA orientation, install
of DOCA itself, non-DEFLATE compression libraries on CPU (use
zlib / zstd / similar), or other DOCA libraries. For those, use
doca-public-knowledge-map/SKILL.md.
What this skill provides
This is a thin loader. The body keeps only the orientation
needed to pick the right next file. The substantive
Compress-specific material lives in two companion files:
- the capabilities companion — what DOCA Compress can express on this
version: the four task types (compress-deflate,
decompress-deflate, decompress-lz4-stream, and
decompress-lz4-block, each independently capability-gated), the
per-task capability-query surface (
doca_compress_cap_* for
task support and per-submission max buffer size), the Compress
error taxonomy (mapped onto the cross-library DOCA_ERROR_*
set), the observability surface (per-task completion events on
the progress engine), the safety policy that gates source /
destination mmap permission decisions, and the size-threshold
path-selection rule (when to use doca-compress versus CPU
compression or doca-dma for the no-compression copy case).
- the tasks companion — step-by-step workflows for the six in-scope
Compress verbs:
configure, build, modify, run, test,
debug. Plus a ## rollback overlay (Compress-specific
five-step teardown that drains in-flight tasks before
doca_ctx_stop, unregisters mmap regions in reverse-register
order, and re-verifies the device path with the round-trip
smoke) and the 5-phase universal debug-loop instantiation
appended to ## debug. Plus a Deferred task verbs block
that points out-of-scope questions at the right next skill.
The skill assumes a host or BlueField where DOCA is already
installed at the standard location and the user has the
privileges their public install profile expects. It does not
cover installing DOCA — that path goes through
doca-setup/SKILL.md.
What this skill deliberately does not ship
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:
- Pre-written DOCA Compress application source code, in any
language. The verified Compress source code is the shipped C
samples at
/opt/mellanox/doca/samples/doca_compress/, plus
the File Compression reference application linked from the
public DOCA Compress guide. 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/SKILL.md,
layered with the Compress-specific overrides inmodify.
- Pre-encoded DEFLATE fixtures for arbitrary inputs. The
skill tells the agent to use a round-trip validation
(compress → decompress and compare to the original, or
decompress a known DEFLATE blob produced by a published CPU
encoder like
zlib) as the known-good smoke; it does not ship
a fixture bank of its own.
- Standalone build manifests (
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, where pkg-config --modversion doca-compress is the source of truth.
- A
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.
Loading order
- Read this skill first to confirm the user's question is in
scope.
- For the Compress capability matrix, the compress / decompress
task split (including decompression-only as a standalone
shape), per-task capability-query rules, permission matrix,
error taxonomy, observability, and safety / path-selection
policy, see the capabilities companion.
- For step-by-step workflows — configure, build, modify, run,
test, debug — see the tasks companion.
Both companion files cross-link to each other,
doca-version/SKILL.md for the canonical
version-handling rules, and
doca-public-knowledge-map/SKILL.md
whenever the right answer is "look it up in the public docs or
the installed package layout" rather than "Compress-specific
guidance".
Related skills
doca-public-knowledge-map/SKILL.md —
the routing table for every public DOCA documentation source
and the on-disk layout of an installed DOCA package. The DOCA
Compress page lives at
docs.nvidia.com/doca/sdk/DOCA-Compress/; the File Compression
reference application is the canonical worked example.
doca-setup/SKILL.md — env preparation,
install verification, and the I have no install yet path
with the public NGC DOCA container. This skill assumes its
preconditions are satisfied.
doca-version/SKILL.md — canonical DOCA
version-handling rules. This skill's
## Version compatibility
cross-links the four-way match rule and adds only the
Compress-specific "discover per-task support + per-task max
buffer size via cap query" overlay.
doca-structured-tools-contract/SKILL.md —
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/SKILL.md —
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 Compress specifics
on top.
doca-dma/SKILL.md — the right library when
the data flow is a pure mmap-to-mmap copy with no compression
required. This skill's path-selection rule routes to DMA when
Compress is not the answer (e.g. the user only wants to move
bytes, not encode them).
doca-debug/SKILL.md — the cross-cutting
debug ladder (install / version / build / link / runtime /
program / driver). Compress-specific debug (task-not-supported,
source-buffer-too-large, destination-buffer-too-small) overlays
on top of that ladder.
1---2license: Apache-2.03name: doca-compress4description: >5 Hardware compression offload with DOCA Compress on a BlueField DPU,6 ConnectX NIC, or host: DEFLATE compress and decompress, plus LZ47 decompress (stream and block; LZ4 encode is NOT supported), instead of8 CPU zlib / zstd. Trigger on "offload this gzip", "decompress incoming9 network data on the NIC", "compress task returns INVALID_VALUE on10 alloc_init", "submitted a task but no completion arrives", or11 "decompress LZ4 on the BlueField", even without naming DOCA Compress.12 Covers enabling compress-deflate, decompress-deflate,13 decompress-lz4-stream and decompress-lz4-block tasks on a doca_compress14 context, sizing source / destination doca_buf against the per-task cap15 query, mmap permissions, offload vs CPU decisions, a round-trip smoke,16 and DOCA_ERROR_* from Compress calls. Refuse and route elsewhere for17 zstd / Snappy / brotli, LZ4 encode (a CPU LZ4 library), pure18 mmap-to-mmap copies (doca-dma), or DOCA Core lifecycle internals.19metadata:20 source: https://skillmd.com/skills/nvidia/doca-compress21 author: NVIDIA <doca-feedback@nvidia.com>22 kind: library23 doca_version: "3.5"24 tags:25 - doca26 - compress27 - deflate28compatibility: >29 Requires DOCA SDK installed at /opt/mellanox/doca on Linux30 (Ubuntu 22.04/24.04 or RHEL/SLES) with a BlueField DPU or31 ConnectX NIC attached. Reads the user's local install via32 `pkg-config doca-compress` and inspects33 /opt/mellanox/doca/{lib,include,samples,applications}.34 The validation steps require a live DOCA install.35---3637# DOCA Compress3839## Examples4041- **"Offload this gzip on the BlueField."** — this is in scope if42 the codec is DEFLATE (or LZ4 decompress). Do not invent LZ443 encode or zstd offload.44- **"Compress task returns INVALID_VALUE on alloc_init."** — size45 the `doca_buf` against the per-task cap query before changing46 the algorithm.4748## Instructions4950Follow Loading order below. Open the tasks companion for configure /51test / debug. Open the capabilities companion for codec support52and buffer-size policy. Do not open every companion file.5354**Where to start:** This skill assumes DOCA is already installed and55the user is doing **hands-on Compress work** (bulk DEFLATE56compression or decompression) on a BlueField / ConnectX / host57with DOCA. Open the tasks companion if the user wants to *do*58something (configure / build / modify / run / test / debug); open59the capabilities companion when the question is *what can60DOCA Compress express* on this version. If the user has not61installed DOCA yet, route to62`doca-setup/SKILL.md` first. If the user is63asking *"should I even offload this compression to the64accelerator?"*, the size-threshold path-selection rule in65## Capabilities and modes66is the first stop — bulk compress is the canonical fit, tiny67one-shot is not.6869## Example questions this skill answers well7071The CLASSES of DOCA Compress questions this skill is built to72answer, each with one worked example. The agent should treat the73*class* as the load-bearing piece — the worked example is a single74instance.7576- **"Should I offload this compression to DOCA Compress, or just77 zlib/zstd on the CPU?"** — worked example: *"I am compressing a78 4 MiB log buffer before writing it to storage; is doca-compress79 worth the setup vs zlib on the CPU?"*. Answered by the80 size-threshold path-selection rule in81 ## Capabilities and modes82 + the *"when NOT to use doca-compress"* bullets in83 ## Safety policy.84- **"Does my device support the compress (or decompress) task I85 want, and how big a buffer can it move per submission?"** —86 worked example: *"is `doca_compress_task_compress_deflate` on87 this BlueField, and what is the max source size per task?"*.88 Answered by the per-task capability-query rule89 (`doca_compress_cap_task_compress_deflate_is_supported`,90 `_decompress_deflate_is_supported`, the matching91 `_get_max_buf_size` queries) in92 ## Capabilities and modes93 + the discovery step in94 ## configure.95- **"I just need to decompress incoming network data — is that a96 valid standalone use of doca-compress?"** — worked example:97 *"my client receives DEFLATE-compressed payloads and never98 produces any compressed output of its own"*. Answered by the99 *"decompression-only"* note in100 ## Capabilities and modes101 task-type table + the per-task configuration matrix in102 ## configure step 5.103- **"What permissions does the source / destination mmap need?"** —104 worked example: *"my `doca_compress_task_compress_deflate`105 returns `DOCA_ERROR_NOT_PERMITTED`"*. Answered by the permission106 matrix in107 ## Safety policy108 + the mmap-set-permissions checklist in109 ## test.110- **"Is this DOCA Compress API available on my installed DOCA111 version?"** — worked example: *"is112 `doca_compress_task_decompress_deflate` in the DOCA I have113 installed?"*. Answered by the version-compatibility overlay in114 ## Version compatibility,115 which cross-links the canonical detection chain in116 `doca-version/SKILL.md` and adds the117 Compress-specific *"discover per-task support, do not assume"*118 bullets.119- **"What does this `DOCA_ERROR_*` from a Compress call mean and120 which layer caused it?"** — worked example: *"`DOCA_ERROR_INVALID_VALUE`121 on `doca_compress_task_compress_deflate_alloc_init`"*. Answered122 by the Compress overlay on the cross-library taxonomy in123 ## Error taxonomy124 + the layered ladder in125 ## debug that escalates to126 `doca-debug/SKILL.md`.127128## Audience129130This skill serves **external developers building applications that131consume the DOCA Compress library** — i.e., users whose code calls132`doca_compress_*` (directly in C/C++, or through FFI/bindings from133another language) to offload bulk DEFLATE compression or134decompression onto a BlueField DPU or ConnectX accelerator. It is135*not* for NVIDIA developers contributing to DOCA Compress itself.136137**Language scope.** DOCA Compress ships as a C library with138`pkg-config` module name `doca-compress`. The shipped samples are139written in C. C and C++ consumers are the canonical case and the140worked examples in the tasks companion assume that path. Other-language141consumers (Rust, Go, Python, …) consume the same `*.so` through142FFI or language-specific bindings; the skill's contribution in143that case is to keep the lifecycle, capability-discovery,144permission, error-taxonomy, and compress-vs-decompress guidance145language-neutral, and to route the agent to the public C ABI as146the authoritative surface that any wrapper will eventually call.147148## When to load this skill149150Load this skill when the user is doing hands-on DOCA Compress151work, in any language. Concretely:152153- Initializing a `doca_compress` context on a `doca_dev` and154 configuring at least one task type155 (`doca_compress_task_compress_deflate` and/or156 `doca_compress_task_decompress_deflate`) before157 `doca_ctx_start()`.158- Choosing between the **compress** task and the **decompress**159 task for the user's data flow. The two task types are160 independent: a consumer that only decompresses inbound DEFLATE161 payloads is a valid shape and does not need to enable the162 compress task.163- Setting permissions on `doca_mmap` correctly for the source164 buffer (`DOCA_ACCESS_FLAG_LOCAL_READ_ONLY` at minimum) and the165 destination buffer (`DOCA_ACCESS_FLAG_LOCAL_READ_WRITE`).166- Sizing the source buffer against the per-task167 `doca_compress_cap_task_*_get_max_buf_size(devinfo)` ceiling168 and the destination buffer against the worst-case output size169 the algorithm can produce on this input.170- Checking which task types this device's accelerator advertises171 via `doca_compress_cap_task_compress_deflate_is_supported` and172 `doca_compress_cap_task_decompress_deflate_is_supported`173 against the active `doca_devinfo`.174- Validating against a round-trip (compress → decompress, or175 decompress → compare against a known DEFLATE-encoded fixture)176 before pushing bulk input through the accelerator.177- Deciding *whether to offload at all* — the size-threshold rule178 in ## Capabilities and modes179 says doca-compress is the right answer only for bulk inputs180 (rule of thumb: ≥ a few KiB); below that, CPU compression beats181 the DMA-to-accelerator round-trip.182- Debugging a `DOCA_ERROR_*` returned from a Compress call183 (lifecycle vs. buffer-sizing vs. permission vs.184 unsupported-task) and the task-completion event on the progress185 engine.186- Designing or extending non-C bindings (Rust, Go, Python, …)187 that wrap the Compress C ABI — for the lifecycle, permission,188 capability, and compress-vs-decompress rules the wrapper must189 honor.190191Do **not** load this skill for general DOCA orientation, install192of DOCA itself, non-DEFLATE compression libraries on CPU (use193zlib / zstd / similar), or other DOCA libraries. For those, use194`doca-public-knowledge-map/SKILL.md`.195196## What this skill provides197198This is a **thin loader**. The body keeps only the orientation199needed to pick the right next file. The substantive200Compress-specific material lives in two companion files:201202- the capabilities companion — what DOCA Compress can express on this203 version: the four task types (compress-deflate,204 decompress-deflate, decompress-lz4-stream, and205 decompress-lz4-block, each independently capability-gated), the206 per-task capability-query surface (`doca_compress_cap_*` for207 task support and per-submission max buffer size), the Compress208 error taxonomy (mapped onto the cross-library `DOCA_ERROR_*`209 set), the observability surface (per-task completion events on210 the progress engine), the safety policy that gates source /211 destination mmap permission decisions, and the size-threshold212 path-selection rule (when to use doca-compress versus CPU213 compression or `doca-dma` for the no-compression copy case).214- the tasks companion — step-by-step workflows for the six in-scope215 Compress verbs: `configure`, `build`, `modify`, `run`, `test`,216 `debug`. Plus a `## rollback` overlay (Compress-specific217 five-step teardown that drains in-flight tasks before218 `doca_ctx_stop`, unregisters mmap regions in reverse-register219 order, and re-verifies the device path with the round-trip220 smoke) and the 5-phase universal debug-loop instantiation221 appended to `## debug`. Plus a `Deferred task verbs` block222 that points out-of-scope questions at the right next skill.223224The skill assumes a host or BlueField where DOCA is already225installed at the standard location and the user has the226privileges their public install profile expects. It does not227cover installing DOCA — that path goes through228`doca-setup/SKILL.md`.229230## What this skill deliberately does not ship231232This skill is **agent guidance**, not a samples or templates233bundle. To keep the boundary clean, it deliberately does not234contain — and pull requests should not add:235236- **Pre-written DOCA Compress application source code, in any237 language.** The verified Compress source code is the shipped C238 samples at `/opt/mellanox/doca/samples/doca_compress/`, plus239 the File Compression reference application linked from the240 public DOCA Compress guide. The agent's job is to route the241 user to those files and prescribe a minimum-diff modification242 on them via the universal modify-a-sample workflow in243 `doca-programming-guide/SKILL.md`,244 layered with the Compress-specific overrides in245 ## modify.246- **Pre-encoded DEFLATE fixtures for arbitrary inputs.** The247 skill tells the agent to use a *round-trip* validation248 (compress → decompress and compare to the original, or249 decompress a known DEFLATE blob produced by a published CPU250 encoder like `zlib`) as the known-good smoke; it does not ship251 a fixture bank of its own.252- **Standalone build manifests** (`meson.build`, `CMakeLists.txt`,253 `Cargo.toml`, …) parked inside the skill. The agent constructs254 the build manifest *in the user's project directory* against255 the user's installed DOCA, where `pkg-config --modversion256 doca-compress` is the source of truth.257- **A `samples/`, `bindings/`, or `reference/` subtree** of any258 kind. A mock or incomplete artifact in this skill's tree, even259 one labeled "reference", is misleading: users will read it as260 buildable.261262## Loading order2632641. Read this skill first to confirm the user's question is in265 scope.2662. **For the Compress capability matrix, the compress / decompress267 task split (including decompression-only as a standalone268 shape), per-task capability-query rules, permission matrix,269 error taxonomy, observability, and safety / path-selection270 policy, see the capabilities companion.**2713. **For step-by-step workflows — configure, build, modify, run,272 test, debug — see the tasks companion.**273274Both companion files cross-link to each other,275`doca-version/SKILL.md` for the canonical276version-handling rules, and277`doca-public-knowledge-map/SKILL.md`278whenever the right answer is "look it up in the public docs or279the installed package layout" rather than "Compress-specific280guidance".281282## Related skills283284- `doca-public-knowledge-map/SKILL.md` —285 the routing table for every public DOCA documentation source286 and the on-disk layout of an installed DOCA package. The DOCA287 Compress page lives at288 `docs.nvidia.com/doca/sdk/DOCA-Compress/`; the File Compression289 reference application is the canonical worked example.290- `doca-setup/SKILL.md` — env preparation,291 install verification, and the *I have no install yet* path292 with the public NGC DOCA container. This skill assumes its293 preconditions are satisfied.294- `doca-version/SKILL.md` — canonical DOCA295 version-handling rules. This skill's296 `## Version compatibility`297 cross-links the four-way match rule and adds only the298 Compress-specific *"discover per-task support + per-task max299 buffer size via cap query"* overlay.300- `doca-structured-tools-contract/SKILL.md` —301 the bundle's structured-tools precedence rule (detect / prefer302 / fall back / report). The Command appendix in303 the tasks companion honors this contract.304- `doca-programming-guide/SKILL.md` —305 general DOCA programming patterns shared by every library: the306 canonical `pkg-config` + meson build pattern, the universal307 modify-a-shipped-sample first-app workflow, the universal308 lifecycle, the cross-library `DOCA_ERROR_*` taxonomy, and the309 program-side debug order. This skill layers Compress specifics310 on top.311- `doca-dma/SKILL.md` — the right library when312 the data flow is a pure mmap-to-mmap copy with no compression313 required. This skill's path-selection rule routes to DMA when314 Compress is *not* the answer (e.g. the user only wants to move315 bytes, not encode them).316- `doca-debug/SKILL.md` — the cross-cutting317 debug ladder (install / version / build / link / runtime /318 program / driver). Compress-specific debug (task-not-supported,319 source-buffer-too-large, destination-buffer-too-small) overlays320 on top of that ladder.