Repository navigation
ROCm CLI v0.3.0 (2026-11-18)
ROCm CLI v0.3.0 builds on v0.2.1. It adds an AIM serving engine, an Anthropic-compatible /v1/messages endpoint, safer serving and runtime changes, and repo-native driver installs, along with a large round of WSL, Windows, dashboard, install and diagnose fixes.
Highlights
- AIM serving engine.
--engine aimserves AMD Enterprise AI AIM Docker containers alongside vLLM and Lemonade. It is opt-in - Anthropic-compatible endpoint.
rocm serveexposes/v1/messages, including tool use and streaming - Follow a starting server.
rocm logs --service <id> --followstreams a server's output, including download progress, until it is up - Safer serving and runtimes.
rocm servepicks a free port without collisions, and runtime activation and rollback are transactional - Newest stable ROCm by default.
rocm install sdknow installs the newest stable release (ROCm 10.2 today) and takes--python-versionfor ROCm 10 installs - Repo-native driver installs.
rocm install driveruses the same flow on apt, dnf and zypper hosts, adds Ubuntu 26.04, and no longer usesamdgpu-install - Better diagnosis.
rocm diagnosetakes--symptom-file, estimates memory for a GGUF model, and flags an outdated Visual C++ runtime. vLLM startup failures now name the failing line - Runtime visibility.
rocm runtimes listandrocm examineshow where each runtime came from, androcm runtimes list --jsonprints JSON - Lemonade on Linux. Lemonade is pinned to 2026.44.1 so a ROCm 10.1 SDK gets a ROCm backend, and installs check glibc and OpenSSL up front
- Security. Endpoint API key files are owner-only on every platform, and vLLM requires the endpoint key on every server route
Breaking Changes
amdgpu-installis removed as an install method, with no deprecation period.rocm install driveron dnf and zypper hosts now uses the same repo-native flow as apt: it imports the AMD key, writes a package-manager repo file and installsamdgpu-dkms. The default driver source moves from a pinned7.2.4tolatest(#181)- The
ROCM_CLI_AMDGPU_VERSIONandROCM_CLI_AMDGPU_PACKAGE_RELEASEenvironment variables are no longer read. SetROCM_CLI_AMDGPU_DRIVER_VERSIONto pin a specific amdgpu release.rocm install driverrefuses to plan while either old variable is still set, so an old pin is never silently turned intolatest
Changes in Behaviour
rocm serve <model>on a host with no usable AMD GPU now succeeds when it matches an already-running managed service for the same engine and model, and prints the reuse summary. Every other case still refuses under the defaultgpu_requiredpolicy (#703)- The Dash serve wizard is now model selection plus inline Advanced settings, validated the same way as the CLI before approval (#243)
- The dashboard assistant no longer advertises the
list_skillsandskill_plantools, whose install recipes nothing could run (#882) rocm install sdkwithout--versionnow installs the newest stable ROCm release instead of ROCm 7.14. Pin--version 7.14.1for the earlier release (#1033)
Notable Fixes
rocm serve --engine vllmon native Windows no longer reports a launch that cannot work. It refuses up front with the reason (#837)rocm uninstallremoves thePATHchange the installer made, and no longer empties a symlinked folder or leavesrocm.exebehind (#636, #562)- On WSL,
rocm examineandrocm serve --gpu autodetect the GPU andlibrocdxgcorrectly (#915, #870, #1091) rocm serveno longer hangs on a badamd-smidriver state, and stopping a managed server escalates toSIGKILLwhen an engine ignores the polite stop (#447, #842, #1088)- Managed services stop reliably, and an unconfirmed stop is no longer reported as stopped (#799, #960)
- vLLM installs on ROCm 10.x no longer fail on an unsatisfiable
amd-aiterpin (#755) - Automation stores no longer lose records to concurrent writers or grow without bound (#738, #744)
- The dashboard runs exactly the tool call its approval card shows (#980)
- An unreadable
config.jsonis no longer replaced with defaults (#541)
Known Issues
- Instinct serving on RHEL 9.6 can still need manual vLLM, OpenMPI and amdsmi workarounds (#254)
- ROCm SDK 10.2.0 with Lemonade 2026.44.1 (observed on Strix Halo in WSL2) copies between ordinary host memory and the GPU run at roughly 20 MiB/s, over 100× slower than on 10.1.0, so loading a large model takes many minutes. rocm serve with Lemonade stops the server at its 2-minute readiness timeout and marks the service failed. Workaround: install ROCm SDK 10.1.0 with
rocm install sdk --version 10.1.0, orrocm runtimes rollbackif a 10.1 runtime was the previous one, and start serving withrocm serve <model>again. (#1143)
All new features
--engine aim: serve AMD Enterprise AI AIM pre-built Docker containers as a third serving engine, alongside vLLM and Lemonade. It is opt-in, and the catalog gains the AIM images, which resolve throughrocm serveand show inrocm model --verbose(#708, #794)rocm logs --service <id> --follow(alsorocm services logs <id> --follow) streams a starting server's output, including in-place download progress bars, until the server is no longer starting or running. Ctrl-C stops following and leaves the server running (#857)rocm servewithout--portnow picks the first free loopback port in11435..=11535. Port claims are serialized between the CLI androcmd, so two launches can no longer take the same port; an explicit--portbehaves as before (#243)rocm runtimes activateandrollback, and the--activatetail ofrocm updateandrocm install sdk, are now transactional. They name the running servers that keep the old runtime, and a failed second write no longer leaves the config and the active-runtime marker disagreeing (#424)rocm install driverplans the Ubuntu 26.04 (resolute) DKMS flow and warns aboutamdgpu-dkmsleft by a previous install, without removing anything (#778)rocm examineshows the AMD display driver version, date and age on Windows, and flags drivers older than the first HIP SDK for Windows (#680)rocm diagnoseflags a Visual C++ runtime older than 14.40, which can crash programs such as Lemonade's llama.cpp server.rocm examine --jsongainsmsvc_redist_version, and a new print-only fix covers it (#679)rocm examinedetects gfx90c (Renoir/Cezanne Vega iGPUs) instead of mapping it togfx90X-dcgpu(#648)- vLLM serve summaries give out-of-memory diagnostics and guidance (#284)
- When vLLM aborts at startup because its engine cannot import (for example a missing
libtorch_cuda.so), the failure now names the failing line and therocm diagnose --symptom '<line>'command that diagnoses it.fix-17-torch-dlpacknow covers Linux and WSL (#646) - The dashboard shows recent serving throughput in the instance detail view, shows starting services on the Home and Serving tabs, and collects GPU and memory metrics in-process on Windows (#721, #858, #715)
- The dashboard Diagnose screen can prepare a Doctor report:
ppreviews what it would carry (nothing is sent), andssends only after approval, with Deny as the default (#647) - The Doctor skill stops when the CLI predates the failure catalog it describes, and covers the two vLLM startup failures (out of memory, and a
libtorch_cuda.soimport error) (#641, #644) rocm diagnose --reportnow says, before offering the mail, that the mail goes from your own account and carries your address, and that the report itself names no user, host or file path (#758)- The dashboard asks for confirmation before quitting while a model is still being served (#595)
rocm installprints cleaner progress and a completion banner (#207)rocm serveexposes an Anthropic-compatible/v1/messagesendpoint, including tool use and streaming requests (#883, #986, #987)rocm install sdk --python-version X.Ychooses the Python for a ROCm 10 SDK install. An install that does not need vLLM can use 3.10 to 3.14; the default stays 3.14 (#989)rocm install sdkreads the current stable and nightly TheRock hosts by default (stable.repo.amd.com,nightly.repo.amd.com) and keeps earlier pins on the old ones. Without--versionit now installs the newest stable release (ROCm 10.1 today) (#1033)rocm diagnose --symptom-file <path>takes the symptom text from a file or stdin, and each match says whether the symptom text contributed to it (#1048, #1094, #1051)rocm diagnoseestimates the context size and memory needed for a GGUF model (#826)rocm fix --jsondescribes each published exit code (#1067)- Lemonade installs announce themselves, show their progress, and are reported by
rocm storage(#853) - Lemonade managed-serve failures capture the backend's identity (#912)
- The dashboard shows one node-efficiency figure everywhere it appears, with a held flag (#1017), attributes VRAM and power to managed
rocm serveservices (#1026), and lists running managed services on Windows (#1002) - The dashboard's chat wait for a reply is two minutes by default and configurable (#1021)
- An engine caller can withhold an inherited endpoint key file (#1034)
rocm runtimes listandrocm examineshow where each runtime came from, androcm runtimes list --jsonprints the list as JSON (#1090, #1098)- vLLM discovery recognizes ROCm 10.2 builds and local version tags of the form
.rocmNNN(#1109) - The engine stdio protocol is now versioned, and an engine built for an incompatible protocol is refused with a clear message instead of failing later (#1073)
- On Linux,
rocm engines install lemonadeand starting Lemonade check the host's glibc and OpenSSL first and say what falls short instead of failing at launch.ROCM_CLI_DISABLE_LEMONADE_HOST_LIBRARY_CHECK=1turns the check into a warning for hosts that provide the libraries another way (#855)
All bug fixes
- Fixed:
rocm install sdk --dry-rundownloadinguvand installing a managed Python on a host without one. The preview now says a real install would prepare Python (python_note:) and changes nothing (#906) - Fixed:
rocm uninstall --dry-runcreating the folders it then listed for removal (#907) - Fixed:
rocm uninstallleaving behind thePATHchange the installer made, and on Windows leaving the runningrocm.exein the install folder. The installers now record theirPATHchange so uninstall can undo exactly it (#636) - Fixed: on WSL,
rocm examinereportingwsl_rocdxg_missingandrocm serve --gpu autorefusing withno usable AMD GPU detectedwhenlibrocdxg.sois reachable only throughLD_LIBRARY_PATH(#915) - Fixed:
rocm examine --jsonnot reporting an AMD GPU on WSL when the Windows host has one (#870) - Fixed:
rocm serve --engine vllmon native Windows printing "managed service launched" and exiting 0 while the engine died with no output. It, androcm engines install vllm, now refuse up front and give the platform reason (#837) - Fixed:
rocm serveon Windows reporting a managed server as started when it exited immediately. It now fails fast and shows the startup log (#243) - Fixed:
rocm servicesreporting an unconfirmed stop as stopped (#799) - Fixed: the dashboard daemon dropping managed services when the registry could not be read, and keeping a stale Lemonade instance after its host or port changed (#927)
- Fixed:
rocm servehanging, and surviving Ctrl-C, whenamd-smiran against a kernel driver in a bad state. The/dev/kfdpre-flight now runs at everyamd-smilaunch site, and anamd-smichild can no longer outlive its timeout (#447, #842) - Fixed:
--gpuaccepting an index Windows does not have (#836) - Fixed:
rocm diagnoserankingfix-13-hip-sdk-missingfirst on hosts with no HIP-based setup (#676) - Fixed: lines printed for Windows by
rocm diagnosenot running in PowerShell (#677) - Fixed: a host ROCm version with a non-numeric build suffix, such as
7.2.4-98rc1, being silently read as7.2.4. It is now treated as unknown (#728) - Fixed: with
ROCM_CLI_CONFIG_DIRandROCM_CLI_DATA_DIRset but noROCM_CLI_CACHE_DIR,rocm storageandrocm uninstallacting on the default~/.rocm/cacheinstead of<data>/cache(#834) - Fixed: an unknown service id surfacing the raw OS error instead of saying the service was not found (#839)
- Fixed: automation stores (events, audit, proposals) losing records to concurrent writers (#738). A store whose torn tail could not be repaired by atomic replace, for example on a full disk or when Windows security software held the file open, could stop
rocmdfrom starting; it is now repaired in place (#822) - Fixed: automation stores growing without bound. They are now capped by rotation and compaction (#744)
- Fixed: on native Windows, a detached
rocm daemonstaying alive afterrocm serveandrocm services stop. It now runs only while an automation needs it, androcm services stopand the uninstall plan say what remains (#838) - Fixed:
rocmdreporting a timeout when a child wrote more output than the pipe holds, because it did not drain the pipes until the child exited (#526) - Fixed:
rocm engines install lemonadefailing on every run when a kept ROCm alignment pin names a download that is no longer served. The packaged backend pins are now restored (#754) - Fixed: vLLM installs on ROCm 10.x failing with an unsatisfiable
amd-aiterpin. Theflash-attnandamd-aiterpins now come from what the discovered vLLM depends on (#755) - Fixed:
/etc/os-releasebeing parsed differently by the driver plan,rocm examineand the OpenMPI hint, so a quotedID='ubuntu'read as an unsupported distro. It is now read by one parser, and/usr/lib/os-releaseis used when/etc/os-releaseis absent (#544, #675) - Fixed: on Windows,
rocm diagnose --report --sendand the dashboard send action never opening a mail client, becausecmdsplit the link at&(#733) - Fixed: command previews that did not read back as the exact arguments that run. Arguments are now quoted for a POSIX shell, including in copyable lines (#563)
- Fixed: CLI output such as
rocm modelshowing an em dash asΓÇöon Windows consoles using the OEM code page. Output text is now ASCII (#683) - Fixed: the dashboard Hardware tab labelling binary VRAM figures as GB instead of GiB (#699)
- Fixed: the dashboard header warning badge (
⚠ N) giving no way to read the warnings. It now opens their text (#793) - Fixed: the dashboard Services manager painting every service's status as healthy, including stopped and failed ones (#850)
- Fixed: the dashboard showing an idle server's all-time TTFT and TPOT indefinitely. Values are now held, marked with
*, and then cleared (#812) - Fixed: dashboard chat tool calls stalling other dashboard work while a slow probe ran (#801)
- Fixed: chat with the ChatGPT sign-in silently ignoring
--temperature,--top-pand--max-tokens. It now warns (#704) - Fixed:
rocm install sdk --dry-runfor a ROCm 10 SDK fetchinguv; the preview no longer downloads anything (#954) - Fixed: the SDK's torch stack being dropped or replaced when torch alignment is disabled, and the ROCm 10 SDK's torch and torchvision not being bounded to the series vLLM pairs (#995, #1031, #1080)
- Fixed: the installer not configuring fish when fish is not the login shell (#816)
- Fixed:
rocm versionreporting the frozen 0.1.0 instead of the release tag (#818) - Fixed:
rocm examinenot reportingamdkfd_loadedandamdgpu_loadedwhen the drivers are built into the kernel (#694) - Fixed: an unreadable
config.jsonbeing replaced with defaults (#541) - Fixed: log tails reported as empty when the log had content (#961)
- Fixed:
rocm uninstallprinting a wrong dry-run footer and answering--forcewith--force-dev-binaries(#916) - Fixed: the update check preparing a managed Python (#951)
- Fixed: a
[y/N]answer that is not understood being treated silently as no; the prompt now says so and ends the line on EOF (#982) - Fixed: managed-service recovery from the embedded and the standalone daemon, and stopping managed services through one routine shared by
rocmandrocmd, with Windows identity, liveness and tree-kill hardening (#948, #960) - Fixed: serving a Lemonade model registered by name on Linux, and discovering serving models in the dashboard (#963, #631)
- Fixed: a variantless Hugging Face repo being passed to Lemonade; it is now refused up front (#919)
- Fixed: the catalog's small assistant pointing at the wrong GGUF build for Lemonade (#1044)
- Fixed: local Qwen3 chat answers lost to unbounded reasoning in the dashboard, and dashboard chat tool output swamping small-context models (#910, #1041)
- Fixed: dashboard Home hero showing a
-0throughput, the two Observe heroes disagreeing on throughput, ADLX initialization on Windows, and CPU graphs not rendering continuously (#914, #1030, #1020, #856) - Fixed: Doctor skill and
rocm fixwording in several places (#976, #1014, #1053, #1093) - Fixed: a ROCm 10.1 SDK getting no usable Lemonade ROCm backend. Lemonade is now pinned to 2026.44.1, which has one (#855)
- Fixed:
rocm servenot saying in its summary when the post-launch smoke test fails (#1050) - Fixed: a managed runtime on WSL not loading its own
librocdxg, and not saying when it does not (#1091) - Fixed: a second
amd-smilibrary ending up in the runtime's torch (#1064) - Fixed: a vLLM install being called usable before its pinned
tensorizerimports (#1092) - Fixed: a timed-out
rocm fixstep not being reported as a timeout (#1062) - Fixed: installer download progress drawing outside the terminal (#847)
- Fixed: running ROCm 7 and ROCm 10 runtimes side by side, and a removed rollback target staying in the active-runtime marker (#1106)
- Fixed: the dashboard running a different tool call than the one its approval card shows. It now runs exactly the call you approved (#980)
- Fixed: Lemonade service state being written non-atomically, so concurrent writers could corrupt it (#1045)
- Fixed:
rocmdleaving an engine child running when the supervisor could not record it (#1039) - Fixed: Lemonade and vLLM teardown not escalating to
SIGKILLwhen an engine ignores the polite stop (#1088) - Fixed: the natural-language plan shown to you being built separately from the plan that runs, so a cloud planner's non-deterministic answer could make them differ. The plan is now built once per request (#741)
- Fixed: the service id being dropped from inline
key=valueservice rows (#698)
Contributors
Thanks to @jussielo-amd, @rominf, @volen-silo, @juhovainio, @tomastola, @r0x0r, @pmoutsias-amd, @johnl-amd, @siloteemu, @fredespi, @nowycondro, @michaelroy-amd and @haribabug for the changes since v0.2.1.
Installation
Install the latest stable release on Linux x86_64:
curl -fsSL https://raw.githubusercontent.com/ROCm/rocm-cli/main/install.sh | shInstall the latest stable release on Windows x86_64 from PowerShell:
$script = "$env:TEMP\install-rocm-cli.ps1"; irm https://raw.githubusercontent.com/ROCm/rocm-cli/main/install.ps1 -OutFile $script; powershell -ExecutionPolicy Bypass -File $scriptInstall a specific tagged release:
curl -fsSL https://raw.githubusercontent.com/ROCm/rocm-cli/main/install.sh | sh -s -- v0.3.0