Skip to content

feat(uma): move to fairchem-core 2.25.0 and keep Nitro out of reach - #97

Merged
hydrogen1222 merged 4 commits into
mainfrom
feat/uma-runtime-upgrade
Nov 14, 2026
Merged

hydrogen1222 merged 4 commits into
mainfrom
feat/uma-runtime-upgrade

Conversation

@hydrogen1222

Copy link
Copy Markdown
Owner

Tracking: #61 (model-support spec §6.4). Draft: ready for independent review.

What was wrong

The pinned fairchem-core 2.21.0 cannot load the UMA-S 1.2.1 checkpoints. The bump is not
only a version change: 2.22+ made UMA Nitro the upstream default preset and removed
turbo, so re-pointing the pin would have silently changed the precision of every
mliport "default" run and could have selected Nitro (Ampere+, reduced precision).

What changes

  • fairchem-core==2.25.0 on torch 2.13.0 (torch~=2.13.0, Python >=3.11,<3.15).
    Volta keeps cu126; Turing+ uses cu130, the CUDA version of the default PyPI
    torch 2.13.0 build (cuda-toolkit==13.0.3); CUDA 13 has no kernels below Turing.
  • FRAMEWORK_PYTHON gains torch 2.13.0; _torch_modern_cuda maps it to cu130. An
    unknown version still fails closed.
  • mliport owns the preset mapping: default → upstream batch (tf32 off, no compile,
    no merged MOLE = the 2.21 default numbers), turbo → upstream multigpu
    (upstream's "turbo's former settings"). A preset that sets nitro is refused; a
    missing preset fails closed with an upgrade message.
  • The effective preset, its tf32 flag and its Nitro flag are added to the model
    identity. UMA Nitro is not exposed.
  • Docs regenerated; docs/backends/uma.md (+ .zh.md) states that the beta scientific
    validation of uma_omat ran on 2.21.0 and does not carry over.

Evidence (owner's V100)

./scripts/install_mliport.sh --engines uma --device auto re-resolved .venv and its
closing mliport doctor --engine uma --device auto passed: fairchem-core 2.25.0,
torch 2.13.0+cu126, real tensor product on cuda:0, arch list
sm_50/60/70/75/80/86/90. The UMA checkpoints themselves were not called: facebook/UMA
is gated and this host has no token (401), so uma-s-1p2p1*.pt is not_run, not passed.

Tests

Updated plan/doctor/gpu-setup/architecture tests for the new pins; new tests pin the
preset mapping, the Nitro refusal and the fail-closed missing-preset path.

DATA IMPACT

No stored scientific record invalidated. Runs produced on fairchem-core 2.21.0 keep their
recorded framework versions and remain valid for that runtime; they are not
interchangeable with runs on 2.25.0, which the docs now say.

Tracking: #61 (model-support spec §6.4)

What was wrong: the pinned fairchem-core 2.21.0 cannot load the UMA-S 1.2.1
checkpoints (uma-s-1p2p1*.pt); they need >= 2.22. The runtime bump is also the
first release whose upstream default changed semantics: 2.22+ made UMA Nitro
the upstream "default" preset (Ampere and newer, reduced precision) and
removed "turbo", so simply re-pointing the pin would have silently changed the
precision of every mliport "default" run and could have selected Nitro.

What changes:
- fairchem-core 2.25.0 (latest, Requires-Python >=3.11,<3.15) on torch 2.13.0.
  2.24 and 2.25 share the torch~=2.13.0 pin, and torch 2.13.0 still publishes
  cu126 wheels, so Volta keeps a supported route; 2.15 drops cu126 entirely.
- FRAMEWORK_PYTHON gains torch 2.13.0 and _torch_modern_cuda maps it to cu130,
  the CUDA version of the default PyPI torch 2.13.0 build (cuda-toolkit 13.0.3).
  CUDA 13 carries no kernel below Turing, which matches the modern-architecture
  set; an unknown torch version still fails closed.
- UMACalculator now owns the preset mapping: mliport "default" runs upstream
  "batch" (tf32 off, no compile, no merged MOLE = the 2.21 "default" numbers)
  and "turbo" runs upstream "multigpu" (upstream calls it "turbo's former
  settings"). A preset that would enable Nitro is refused, a missing preset
  fails closed with an upgrade message instead of degrading, and the effective
  preset, its tf32 flag and its Nitro flag are added to the model identity.
- Generated install docs and the UMA backend page are regenerated; the page
  records that the beta scientific validation of uma_omat ran on 2.21.0 and
  does not carry over.
- Tests updated for the new pins and new tests pin the preset mapping, the
  Nitro refusal and the fail-closed missing-preset path.

Evidence (owner's V100, Tesla V100-SXM2-16GB, driver 580.173.02):
./scripts/install_mliport.sh --engines uma --device auto re-resolved the
existing .venv and its closing `mliport doctor --engine uma --device auto`
passed: fairchem-core 2.25.0, torch 2.13.0+cu126 (CUDA 12.6), a real tensor
dot product on cuda:0, arch list sm_50/sm_60/sm_70/sm_75/sm_80/sm_86/sm_90.
The UMA checkpoints themselves are not called here: facebook/UMA is a gated
repository and this host has no HF token, so uma-s-1p2p1*.pt could not be
downloaded. That callable check remains not_run.

DATA IMPACT: no stored scientific record invalidated. Runs produced on
fairchem-core 2.21.0 keep their recorded framework versions and stay valid for
that runtime; they are not interchangeable with runs on 2.25.0, which the
docs now say explicitly.
Tracking: #61 (model-support spec §11 task 4, review of #97)

What was missing: the docs said Turing and newer use "cu128 or later" and that
the UMA runtime is on cu130, but not that the CUDA 13 wheels require an NVIDIA
driver of 580 or newer, and not that Volta stays on cu126 precisely so older
drivers keep working.

What changes:
- `docs/installation.md` (+ `.zh.md`): the GPU-architecture section names the
  `cu130` driver floor and says a CUDA 13 build is not an alternative for
  pre-Turing GPUs because it has no kernels below Turing, whatever the driver.
- `docs/backends/uma.md` (+ `.zh.md`): the "Runtime pin" bullet says the same
  in one sentence.
- CHANGELOG entry.

Verified on the owner's host: driver 580.173.02 runs torch 2.13.0+cu126 on the
V100 (sm_70 kernels present in the wheel's arch list).

DATA IMPACT: no scientific record invalidated; documentation only.
@hydrogen1222

Copy link
Copy Markdown
Owner Author

§11 task 4 done, and task 1 has passed for the checkpoints that are readable on this host, so this PR is marked ready for review.

Changes requested by §11 task 4:

  • docs/installation.md (+ .zh.md): the GPU-architecture section now states that the CUDA 13 cu130 wheels (used on Turing and newer, and by the UMA runtime from torch 2.13) need an NVIDIA driver of 580 or newer, that Volta stays on cu126 precisely so older drivers keep working, and that a CUDA 13 build is not an alternative for pre-Turing GPUs because it has no kernels below Turing.
  • docs/backends/uma.md (+ .zh.md): the "Runtime pin" bullet says the same in one sentence.
  • origin/main merged in; both changelog sections kept; CI green (13 pass).

Task-1 boundary, for the reviewer: on the #97 runtime uma-s-1p2p1.pt and uma-s-1p2p1-ft-r2scan.pt load and pass the bounded check on CPU and GPU with the recorded identity inference_preset: batch, nitro_enabled: false, tf32_enabled: false; --inference-mode turbo records inference_preset: multigpu, nitro_enabled: false. The third checkpoint, uma-s-1p2.pt, cannot be called at all: the local copy fails the zip CRC (PytorchStreamReader failed reading file data/115 ... corrupted), the file size matches upstream, and facebook/UMA is gated (HTTP 401 here), so it could not be re-fetched. That is recorded verbatim in the new #75 comment and is an artifact-transfer problem, not a runtime defect of this PR.

@hydrogen1222
hydrogen1222 marked this pull request as ready for review November 14, 2026 15:00
Resolve the CHANGELOG `Unreleased` conflicts by keeping every entry (registry
corrections, MACE gate, and this branch's UMA runtime note) and dropping a
whitespace-only hunk. No code file conflicted.

DATA IMPACT: none.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GYQHUBfRm75zyaocAPWhEb
@hydrogen1222
hydrogen1222 merged commit 76c75ae into main Nov 14, 2026
14 checks passed
@hydrogen1222
hydrogen1222 deleted the feat/uma-runtime-upgrade branch November 14, 2026 15:31
hydrogen1222 added a commit that referenced this pull request Nov 14, 2026
Tracking: #61 (model-support spec §6.5 + §11 tasks 1 and 5)

Rebased onto `main` after #94, #96 and #97 merged, so this carries only the
status-table change.

What changes:
- The status table leaves its initial state and carries what was measured on
  2026-11-14 on the owner's V100 in the installer-built environments: 18
  profiles `verified` on CPU and GPU, each with finite energy/force/stress on
  the 50-atom LGPS cell, a matching recorded identity and the bounded probe
  (shared calculator, central differences, 10 NVE steps, five-image NEB).
- UMA rows from §11 task 1: `uma_omat`, `uma_s_1p2p1_omat` and
  `uma_s_1p2p1_ft_r2scan_omat` verified on both devices and recorded with
  `inference_preset: batch`, `nitro_enabled: false`, `tf32_enabled: false`;
  `--inference-mode turbo` records `multigpu` with `nitro_enabled: false`. The
  r2SCAN fine-tune's NEB leg is refused by the gradient preflight (energy
  resolution), which the table states instead of hiding. `uma_omat` needed the
  re-uploaded checkpoint: the first copy had the right size but failed the zip
  CRC; the new file matches the pinned manifest hash `ba5c0d912efa...`.
- The CPU-GPU claim quotes the measured maxima (6.1e-7 eV/atom, 7.5e-6
  eV/Angstrom, `mace_omat_float32`) instead of a rounded "1e-7/1e-5 or better".
- Every Evidence cell links the specific #75 comment (runs, deltas, UMA record,
  re-check, NEP89 record) instead of the pull request.
- The Chinese table is translated column by column; it previously carried the
  English dtype and evidence cells.

DATA IMPACT: no scientific record invalidated. Statuses describe calling the
models correctly; no accuracy claim is made and no stored campaign record
changed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants