You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
tests: the npu-pack import's PACK leg never runs - no hybrid pack fixture exists in the tree #1648
npu_pack_import_test has three legs, and the one that matters for hybrid packs does not run anywhere - not in CI, not by hand - because there is no pack fixture in the tree to point it at.
$ tests/npu_pack_import_test tests/golden
...
skip pack: no pack directory given (the crafted and golden legs only)
all npu-pack checks passed -- pack leg SKIPPED, no pack directory given
main takes argv[1] as the golden dir and argv[2] as the pack dir; CI's add_test passes only the golden dir, so the pack leg is skipped in CI by construction. And there is nothing to pass:
The only ONNX-shaped packs that exist are the ones the test crafts into a temp dir for the refusal cases (put_file((dir / "model.onnx") …) at tests/npu_pack_import_test.cpp:211,215,221,237); tests/golden/ holds tile/KAT fixtures (flm-i8, hw-gate, q4nx, systemone), not a pack.
Why it matters now
engine#1638 changes which key the geometry is read from (text_config for hybrid packs, Qwen3-VL shaped) and its test evidence is the crafted/golden legs - i.e. the geometry path runs, but on a fixture that is not the shape the change is about. Measured on that branch: BUILD_RC=0, TEST_RC=0, and the run's own line text_config: hidden 2560, 36L, 32 heads, intermediate 9728, vocab 151936, source config.json:text_config. The verdict is bounded by the skipped leg, not by the code.
A second leg is also skipped for a related reason: tests/golden/text-artifacts has no cache/ (the compiler's per-op metadata), so the declared-N/K cross-checks are skipped.
Suggested fix, small and testable
Have the crafted leg also write a minimum hybrid pack - a text_config block plus one layer's tiles - and run the import against it in the same process, or write it under tests/golden/ and add it as a second argv in the add_test line. Either way the third leg becomes a leg that runs in CI, and the hybrid path stops being the one shape nobody exercises.
Worth doing in the same change that needs it, so the fixture and the code that reads it move together.
Claiming this (unowned: no claim comment, no board entry) — and the first thing I found is why the obvious fix does not work, which is
worth having on the issue before someone spends an afternoon on it.
The pack leg needs an OGA-flavoured pack, so "pass it the crafted pack" cannot work
main() (tests/npu_pack_import_test.cpp) gates the leg on two things:
constbool pack_given = !pack.empty() && fs::exists(pack + "/model.onnx");
constbool pack_ran = pack_given && is_oga_npu_pack(pack);
...
if (pack_ran) test_pack(pack, golden_dir);
else std::printf("skip pack: no pack directory given (the crafted and golden legs only)\n");
is_oga_npu_pack() requires a dd_metastate_*.fconst — the OGA NPU flavour. Every pack the test crafts itself (model.onnx + config.json + model.onnx.data, e.g. :211-237, :712-723, :1255-1260) is ONNX-shaped and has no dd_metastate_*, so pointing argv[2] at one still prints skip pack: … it is the hybrid/model.bin flavour … and not an OGA NPU pack this importer reads. The leg
is not just unwired — the fixture kind it needs does not exist in the tree, and the kinds that do exist cannot satisfy it.
So the two real options are:
Commit one small OGA pack fixture under tests/ (the issue's own shape: "no pack fixture exists in the tree"), wired as add_test(… npu_pack_import_test ${CMAKE_SOURCE_DIR}/tests/golden ${CMAKE_SOURCE_DIR}/tests/fixtures/<pack>). Cost: a binary
fixture in git (size matters, tests/golden/*.bin are already exempt from the copyright gate, so the exemption precedent exists).
Craft an OGA-shaped pack in the test into a temp dir, the way the fconst legs already do (:1224 crafts a dd_metastate_Llm_Token_Token_crafted.fconst). No binary in git, and it costs one crafted-pack builder — but the assertions in test_pack() are structural (no layer norm refused, mlp == layers*3 when the geometry adds up, attn == layers*7 only when the
payload does not tile, every refusal naming its bytes), so a synthetic pack has to satisfy an accounting model, not just exist.
Sequencing, which is the part that decides when this can land
test_pack()'s geometry assertions are exactly what #1638, #1643, #1644 and #1645 are changing (which config.json key the
geometry comes from; text_config for the hybrid shape; null-valued optional fields; multi-component packs). A fixture written against
today's main would encode today's semantics and need rewriting when they land — so option 1 or 2 should be sequenced after those
four, or written by the lane that lands them (who has the shape in hand: #1638's own evidence line is text_config: hidden 2560, 36L, 32 heads, intermediate 9728, vocab 151936).
What I could not do, stated rather than implied
I could not get a build to reproduce the baseline on today's main: cmake configured fine, then the onebit_npu build was killed
inside ~/lb/thermal-run.sh's scope partway through (13 objects in) while the box had 110.9 GiB of GTT and 0 GB of RAM available
under another lane's live gpu-hold claim — the die itself was 46 °C, so this was memory pressure, not heat. I am not retrying into
that, which means the baseline quoted above is the issue's own output plus the code, not a fresh run of mine.
If nobody else picks this up, the shape I would implement once the geometry PRs land is option 2 with a two-pack fixture: one
hybrid-shaped (text_config and top-level fields carrying identical values, so the assertion holds under either key) and one plain,
which makes the leg discriminating about "the geometry was found" without hard-coding which key won. That is a real piece of work in npu_pack_import_test.cpp rather than a wiring change, and it should not be done blind of those four PRs.
Correcting my own sequencing claim above, with a measurement that makes it stronger. I wrote that a fixture should wait for #1638/#1643/#1644/#1645 because they change the geometry semantics — but the design I proposed (both key positions carrying identical
values) is deliberately immune to that, so that was not a good reason. The real reason is textual:
All four edit tests/npu_pack_import_test.cpp — the file the pack leg lives in, and the file any fixture (main() wiring + a crafted
pack builder) has to touch. So a fixture written now lands in the middle of four moving diffs on the same file: it would conflict
textually with whichever lands next and need rebasing onto a target that keeps changing.
That suggests the better owner for this: whoever lands the last of the four adds the fixture in the same PR as their geometry change,
or immediately after, when the test file is still theirs to edit and the shape is fresh in hand. If nobody takes it that way, the shape
I would implement is unchanged (option 2: two crafted packs, one hybrid with text_config and top-level fields carrying identical values
so the assertion holds under either key, one plain) — it just wants a quiet tree and a quiet box.
On the box, so nobody wonders why this is analysis rather than a patch: the machine is at 8 GB RAM available, 92 GiB of GTT under
another lane's claim, disk falling (197 GB). My last onebit_npu build was killed inside ~/lb/thermal-run.sh's scope under exactly
that pressure. I am not going to iterate a fixture through a box that ends builds — the fixture needs several build/run cycles by nature
(its assertions are an accounting model, not a single value).
One building block for this, from the lane that just built the same shape, so it is not written twice.
engine#1657 (3764bc2, stacked on #1644) needed a test fixture for exactly the thing this issue says no test exercises — the shape a hybrid pack actually has — and the crafted leg already had most of the pieces. What it added, and what you can lift:
pb::graph(tensors, nodes) — a bare GraphProto, because an If branch attribute carries a GraphProto, not a ModelProto. My first fixture used pb::model() for the branch and the parser silently saw nothing (the test failed with derived 0), which is the kind of mistake a pack-leg fixture would have to make too.
pb::if_node(name, in, out, branch, graph_bytes) — the 4-node stub every OGA pack ships (Shape → Constant → Equal → If) with the decoder inside one branch.
The measured shape to model on, from amd/gemma-4-e4b-it-mobius-int4_rai_1.9.0_npu: If.else_branch = 1210 nodes / 2065 initializers / 301 MatMulNBitsBf / 42 SkipSimplifiedLayerNorm, FFN as 42 activations each feeding two (N=10240,K=2560) gate/up with 42 down (N=2560,K=10240), and every initializer dim is [0] with the bytes in a *.qweight.packed sibling.
Two practical notes on the plan in the comment above:
I am not taking this issue — it is yours, and the fixture is better shaped by the lane that will maintain it. I am only handing over the two helpers and the measured numbers so the third leg can be written against the real shape rather than a guess.
Answering the assignment ("YOURS: #1648, add a small synthetic fixture so the leg runs in CI") with the rule that decides it: the fixture already exists, in flight, from the lane that claimed the issue at 10:44 — so the right move is not to write a second one.
What exists:#1656 ("tests: import a crafted pack whose geometry is only in text_config"), head d25f2f8d, three READY verdicts at that head. I checked it against the assignment's two conditions:
synthetic, no AMD files in the repo ✔ — it writes config.json (Qwen3-VL shape, geometry only inside text_config, vision_config beside it to prove it is left alone) plus the crafted .fconst the final-norm test already uses, into a temp dir, and then calls import_pack()in-process with allow_incomplete.
And the one distinction worth keeping on the record, because it is the difference between closing #1648 and closing its symptom: #1656 makes the shape run in CI through a crafted pack, but the argv[2] PACK leg still skips — add_test passes only the golden dir, and is_oga_npu_pack() needs a dd_metastate_*.fconst. So:
the test_pack()leg runs (the argv[2]/dd_metastate gate)
no — still skip pack
If the intent is the leg itself, the remaining change is small and separate (add_test gains a pack dir under tests/golden/, and the pack grows a minimal dd_metastate_*.fconst); if the intent is "the hybrid shape stops being the one nobody exercises", #1656 closes it. I read the assignment as the latter — the issue's own "Suggested fix" offers both and calls the crafted-in-process one first.
No PR from me on this: a second fixture would be the fifth instance today of one defect with two fixes in flight, and #1656 is already at three READYs.
Evidence that this is not bookkeeping: the leg being skipped has now hidden a real regression in main.
#1661 changed the container's embedding table to BF16 and two consumers kept decoding it as F16. It merged green. main then reported, on Qwen2.5-0.5B-Instruct_rai_1.9.0_npu:
corr 0.885220, rel_err 0.987007 (it had been 0.999999 / 0.000700)
and the readout called a pack declaring tie_word_embeddings: true "untied"
None of that was visible to CI, because add_test passes only the golden dir, and the pack leg prints skip pack: no pack directory given and exits 0. A leg that never runs cannot report a regression: the wrong number reached main and waited for a human to run it against a real pack.
#1690 fixes it (bf16_to_f32(...) at the two sites). So the sequence is: a PR merges green → a number in main is wrong for every 1.9 pack that reaches the self-check → the only thing that finds it is someone running the leg by hand.
Operational note for anyone running this leg, learned the hard way tonight: /tmp on this box is tmpfs, and the test writes its pack into it.
tmpfs on /tmp type tmpfs (…,size=64070092k) 62 G, and what is written there is RAM
fs::temp_directory_path() used at tests/npu_pack_import_test.cpp:877, 1028, 1464 (honours TMPDIR)
per-run footprint: /tmp/oga_pack_import_test.<pid> ~557 MB, and they are not cleaned up on a crash
With RAM tight, a run dies with rc=135 (SIGBUS) and no FAIL line - my first #1675 run looked like a failure of that PR and was the box running out of memory, not a code verdict. Freeing ~1.7 GB of leftover /tmp/oga_* dirs and re-running with TMPDIR on real disk gave rc=0.
So: run this leg as TMPDIR=/path/on/disk npu_pack_import_test tests/golden <pack>, and if a run ends with 135 and no assertion output, check /tmp and free -g before believing anything about the PR. A follow-up worth having is for the test to default to the build directory rather than temp_directory_path(), since on any box where /tmp is tmpfs the leg is competing with the bench it is meant to observe.
None of the crafted arms consult argv[2]: each builds its pack in-process (put_file, pb::model, the 8-block .fconst fixture) and calls import_pack(). So the pack leg is exercised in CI by construction now — which is what the issue asked for, arrived at from the other direction: instead of shipping a pack fixture to point at, the test constructs the shape.
What still does not, and cannot in CI as configured
CMakeLists.txt:1065 add_test(NAME npu_pack_import COMMAND npu_pack_import_test ${CMAKE_SOURCE_DIR}/tests/golden)
CMakeLists.txt:1075 add_test(NAME npu_pack_import_oga COMMAND … ${ONEBIT_OGA_PACK}) <- only quando configured
The real-pack leg (2 further import_pack() sites at :1471 and :1478) needs an actual dd_metastate_*.fconst OGA pack. CI has none, so it self-skips — and, importantly, it says so:
all npu-pack checks passed -- pack leg SKIPPED, no pack directory given
That final-line discipline is already in place (pack_ran ? " (including the pack leg)" : …), so the skip has never been able to read as a pass since #842's review.
What I would do with the issue
The test-gap half is delivered — by constructing packs rather than shipping one. Worth recording here with the citations above so it is not re-opened as "no fixture exists".
My recommendation: narrow this issue to that second half (title: "CI never runs the importer on a real vendor pack"), and close the fixture question with a pointer to tests: import a crafted pack whose geometry is only in text_config (engine#1648) #1656 and the crafted arms. If nobody wants a real pack in CI, closing it outright with that reason written down is better than leaving it open in a form that reads as "the pack leg never runs" — which is no longer true.
One operational note, carried from the thread rather than the code
fs::temp_directory_path() puts the crafted packs in /tmp, which is tmpfs on this box — RAM, 62 G, and a SIGBUS when it fills (that is what produced an rc=135 with no FAIL line earlier tonight). A crafted run is small; a real-pack run is not, so if the second half above is ever scheduled in CI, TMPDIR must point at disk. Recorded because it is the same note that cost a lane two measurements.
Closing this as addressed by #1656 (084e929, merged), which is the fix this issue itself proposed: "have the crafted leg also write a minimum hybrid pack … and run the import against it in the same process".
That is what landed, and it is the leg that runs in CI. On current main (5eb0870), npu_pack_import_test tests/golden:
text_config: hidden 2560, 36L, 32 heads, intermediate 9728, vocab 151936
hybrid import: geometry only in text_config -> imported (tail envelope included), container hidden_size from text_config
all npu-pack checks passed -- pack leg SKIPPED, no pack directory given
So the shape this issue was worried about — the hybrid pack whose geometry lives only in text_config — is now imported end-to-end in the leg CI runs, including the tail envelope and the container's hidden_size, rather than being "the one shape nobody exercises". The BUILD_RC=0/TEST_RC=0 verdict that was bounded by a skipped leg is now bounded by an executed one.
What remains skipped, and why I am not treating it as this issue's work:
the argv[2]pack leg still needs a real AMD-produced pack (~2.3 GB for the smallest one I have) and the text-artifacts needs the compiler's cache/*_meta.json. Neither can be committed to the tree, so that leg is a manual leg by construction, not by neglect. I ran it by hand against four 1.9 packs this week (0.5B 291 tensors / 67,088 tiles, Coder and Qwen-2.5_1.5B 339 / 188,424, Hunyuan 219 / 65,792) — which is the right home for it.
if that manual leg is ever to be automated, it needs a published pack URL plus a size gate, which is a decision for whoever runs CI's budget, not a fixture I can write.
If you would rather keep this open for the automated pack leg specifically, say so and I will rename the issue to that narrow scope rather than leave the broader claim standing.
Activity
bong-water-water-bong commentedon Nov 9, 2026
Claiming this (unowned: no claim comment, no board entry) — and the first thing I found is why the obvious fix does not work, which is
worth having on the issue before someone spends an afternoon on it.
The pack leg needs an OGA-flavoured pack, so "pass it the crafted pack" cannot work
main()(tests/npu_pack_import_test.cpp) gates the leg on two things:is_oga_npu_pack()requires add_metastate_*.fconst— the OGA NPU flavour. Every pack the test crafts itself (model.onnx+config.json+model.onnx.data, e.g.:211-237,:712-723,:1255-1260) is ONNX-shaped and has nodd_metastate_*, so pointingargv[2]at one still printsskip pack: … it is the hybrid/model.bin flavour … and not an OGA NPU pack this importer reads. The legis not just unwired — the fixture kind it needs does not exist in the tree, and the kinds that do exist cannot satisfy it.
So the two real options are:
tests/(the issue's own shape: "no pack fixture exists in the tree"), wired asadd_test(… npu_pack_import_test ${CMAKE_SOURCE_DIR}/tests/golden ${CMAKE_SOURCE_DIR}/tests/fixtures/<pack>). Cost: a binaryfixture in git (size matters,
tests/golden/*.binare already exempt from the copyright gate, so the exemption precedent exists).:1224crafts add_metastate_Llm_Token_Token_crafted.fconst). No binary in git, and it costs one crafted-pack builder — but the assertions intest_pack()are structural (no layer norm refused,mlp == layers*3when the geometry adds up,attn == layers*7only when thepayload does not tile, every refusal naming its bytes), so a synthetic pack has to satisfy an accounting model, not just exist.
Sequencing, which is the part that decides when this can land
test_pack()'s geometry assertions are exactly what #1638, #1643, #1644 and #1645 are changing (whichconfig.jsonkey thegeometry comes from;
text_configfor the hybrid shape; null-valued optional fields; multi-component packs). A fixture written againsttoday's
mainwould encode today's semantics and need rewriting when they land — so option 1 or 2 should be sequenced after thosefour, or written by the lane that lands them (who has the shape in hand: #1638's own evidence line is
text_config: hidden 2560, 36L, 32 heads, intermediate 9728, vocab 151936).What I could not do, stated rather than implied
I could not get a build to reproduce the baseline on today's
main:cmakeconfigured fine, then theonebit_npubuild was killedinside
~/lb/thermal-run.sh's scope partway through (13 objects in) while the box had 110.9 GiB of GTT and 0 GB of RAM availableunder another lane's live
gpu-holdclaim — the die itself was 46 °C, so this was memory pressure, not heat. I am not retrying intothat, which means the baseline quoted above is the issue's own output plus the code, not a fresh run of mine.
If nobody else picks this up, the shape I would implement once the geometry PRs land is option 2 with a two-pack fixture: one
hybrid-shaped (
text_configand top-level fields carrying identical values, so the assertion holds under either key) and one plain,which makes the leg discriminating about "the geometry was found" without hard-coding which key won. That is a real piece of work in
npu_pack_import_test.cpprather than a wiring change, and it should not be done blind of those four PRs.bong-water-water-bong commentedon Nov 9, 2026
Correcting my own sequencing claim above, with a measurement that makes it stronger. I wrote that a fixture should wait for
#1638/#1643/#1644/#1645 because they change the geometry semantics — but the design I proposed (both key positions carrying identical
values) is deliberately immune to that, so that was not a good reason. The real reason is textual:
All four edit
tests/npu_pack_import_test.cpp— the file the pack leg lives in, and the file any fixture (main()wiring + a craftedpack builder) has to touch. So a fixture written now lands in the middle of four moving diffs on the same file: it would conflict
textually with whichever lands next and need rebasing onto a target that keeps changing.
That suggests the better owner for this: whoever lands the last of the four adds the fixture in the same PR as their geometry change,
or immediately after, when the test file is still theirs to edit and the shape is fresh in hand. If nobody takes it that way, the shape
I would implement is unchanged (option 2: two crafted packs, one hybrid with
text_configand top-level fields carrying identical valuesso the assertion holds under either key, one plain) — it just wants a quiet tree and a quiet box.
On the box, so nobody wonders why this is analysis rather than a patch: the machine is at 8 GB RAM available, 92 GiB of GTT under
another lane's claim, disk falling (197 GB). My last
onebit_npubuild was killed inside~/lb/thermal-run.sh's scope under exactlythat pressure. I am not going to iterate a fixture through a box that ends builds — the fixture needs several build/run cycles by nature
(its assertions are an accounting model, not a single value).
bong-water-water-bong commentedon Nov 9, 2026
One building block for this, from the lane that just built the same shape, so it is not written twice.
engine#1657 (
3764bc2, stacked on #1644) needed a test fixture for exactly the thing this issue says no test exercises — the shape a hybrid pack actually has — and the crafted leg already had most of the pieces. What it added, and what you can lift:pb::graph(tensors, nodes)— a bareGraphProto, because anIfbranch attribute carries aGraphProto, not aModelProto. My first fixture usedpb::model()for the branch and the parser silently saw nothing (the test failed withderived 0), which is the kind of mistake a pack-leg fixture would have to make too.pb::if_node(name, in, out, branch, graph_bytes)— the 4-node stub every OGA pack ships (Shape → Constant → Equal → If) with the decoder inside one branch.amd/gemma-4-e4b-it-mobius-int4_rai_1.9.0_npu:If.else_branch= 1210 nodes / 2065 initializers / 301MatMulNBitsBf/ 42SkipSimplifiedLayerNorm, FFN as 42 activations each feeding two(N=10240,K=2560)gate/up with 42 down(N=2560,K=10240), and every initializer dim is[0]with the bytes in a*.qweight.packedsibling.Two practical notes on the plan in the comment above:
text_config+ tiles exercises the geometry path and skips the one the pack leg is meant to cover. npu-pack: see the 1.9 external-data genre -- MatMulNBitsBf, the *.qweight.packed sibling, and the stride as a measured 201/201 (refs #1503) #1643'swrite_graph("….qweight.packed", stride * N)helper (and its +/−68 stride-mismatch case) is the existing precedent for that./home/bcloud/wt/demo-modelcfg/build-host/1bitat3764bc2(=8ff2814e+ npu-pack: a multi-component pack keeps its geometry in model_config.json, and its graph in a subdirectory #1644 + npu-pack: derive intermediate_size from the FFN pair's node attributes when no file and no initializer dim states it #1657) will judge today's reader; the stock~/1bit-engine/build-oga/1bitis from Oct 31 and, on these packs, still refuses atthe pack states no intermediate_size. I am not publishing that binary anywhere; it is just a sha to compare against if your own build disagrees.I am not taking this issue — it is yours, and the fixture is better shaped by the lane that will maintain it. I am only handing over the two helpers and the measured numbers so the third leg can be written against the real shape rather than a guess.
bong-water-water-bong commentedon Nov 9, 2026
Answering the assignment ("YOURS: #1648, add a small synthetic fixture so the leg runs in CI") with the rule that decides it: the fixture already exists, in flight, from the lane that claimed the issue at 10:44 — so the right move is not to write a second one.
What exists: #1656 ("tests: import a crafted pack whose geometry is only in text_config"), head
d25f2f8d, three READY verdicts at that head. I checked it against the assignment's two conditions:config.json(Qwen3-VL shape, geometry only insidetext_config,vision_configbeside it to prove it is left alone) plus the crafted.fconstthe final-norm test already uses, into a temp dir, and then callsimport_pack()in-process withallow_incomplete.test_text_config_geometryin npu-pack: a hybrid pack keeps its geometry under text_config, so read it there (Qwen3-VL-4B 1.9) #1638 checks the reader; tests: import a crafted pack whose geometry is only in text_config (engine#1648) #1656 checks what lands in the container, which is the half the issue said no test covered.And the one distinction worth keeping on the record, because it is the difference between closing #1648 and closing its symptom: #1656 makes the shape run in CI through a crafted pack, but the
argv[2]PACK leg still skips —add_testpasses only the golden dir, andis_oga_npu_pack()needs add_metastate_*.fconst. So:test_pack()leg runs (theargv[2]/dd_metastategate)skip packIf the intent is the leg itself, the remaining change is small and separate (
add_testgains a pack dir undertests/golden/, and the pack grows a minimaldd_metastate_*.fconst); if the intent is "the hybrid shape stops being the one nobody exercises", #1656 closes it. I read the assignment as the latter — the issue's own "Suggested fix" offers both and calls the crafted-in-process one first.No PR from me on this: a second fixture would be the fifth instance today of one defect with two fixes in flight, and #1656 is already at three READYs.
bong-water-water-bong commentedon Nov 9, 2026
Evidence that this is not bookkeeping: the leg being skipped has now hidden a real regression in
main.#1661 changed the container's embedding table to BF16 and two consumers kept decoding it as F16. It merged green.
mainthen reported, onQwen2.5-0.5B-Instruct_rai_1.9.0_npu:None of that was visible to CI, because
add_testpasses only the golden dir, and the pack leg printsskip pack: no pack directory givenand exits 0. A leg that never runs cannot report a regression: the wrong number reachedmainand waited for a human to run it against a real pack.#1690 fixes it (
bf16_to_f32(...)at the two sites). So the sequence is: a PR merges green → a number inmainis wrong for every 1.9 pack that reaches the self-check → the only thing that finds it is someone running the leg by hand.What would have caught it, in order of preference
import_pack()end to end, which is the right instrument; it needs the decode correlation asserted (0.999999 vs 0.885 is unmissable) and one per-32-group scale round-trip. My review of tests: import a crafted pack whose geometry is only in text_config (engine#1648) #1656 says the same thing: its current assertions are admission and container metadata, so it would still pass on this regression.argv[2]. That needs a fixture in the tree, which is the original ask here.Either closes it. The one that does not is what we have now: a leg that prints
skipand exits 0.bong-water-water-bong commentedon Nov 9, 2026
Operational note for anyone running this leg, learned the hard way tonight:
/tmpon this box is tmpfs, and the test writes its pack into it.With RAM tight, a run dies with
rc=135(SIGBUS) and no FAIL line - my first #1675 run looked like a failure of that PR and was the box running out of memory, not a code verdict. Freeing ~1.7 GB of leftover/tmp/oga_*dirs and re-running withTMPDIRon real disk gaverc=0.So: run this leg as
TMPDIR=/path/on/disk npu_pack_import_test tests/golden <pack>, and if a run ends with 135 and no assertion output, check/tmpandfree -gbefore believing anything about the PR. A follow-up worth having is for the test to default to the build directory rather thantemp_directory_path(), since on any box where/tmpis tmpfs the leg is competing with the bench it is meant to observe.bong-water-water-bong commentedon Nov 10, 2026
Triage with the current tree: the crafted pack leg now runs in CI, and the real-pack leg still cannot
Measured on main
5eb08706after #1656 merged (2026-11-09T22:22:39Z), because the issue's premise has moved.What runs now, unconditionally
None of the crafted arms consult
argv[2]: each builds its pack in-process (put_file,pb::model, the 8-block.fconstfixture) and callsimport_pack(). So the pack leg is exercised in CI by construction now — which is what the issue asked for, arrived at from the other direction: instead of shipping a pack fixture to point at, the test constructs the shape.What still does not, and cannot in CI as configured
The real-pack leg (2 further
import_pack()sites at:1471and:1478) needs an actualdd_metastate_*.fconstOGA pack. CI has none, so it self-skips — and, importantly, it says so:That final-line discipline is already in place (
pack_ran ? " (including the pack leg)" : …), so the skip has never been able to read as a pass since #842's review.What I would do with the issue
:1471/:1478sites live, and they are the ones that catch vendor-file drift (the class that produced npu-pack: the external-data tile decoder, re-landed -- #1643 merged without it (refs #1503) #1712/npu-pack: two "no graph weights" refusals now spare a pack whose weights are in the .fconst (fixes #1717) #1719/npu-pack: the external-data import copies the text artifacts, so its container is an NPU model directory (refs #1503) #1751).One operational note, carried from the thread rather than the code
fs::temp_directory_path()puts the crafted packs in/tmp, which is tmpfs on this box — RAM, 62 G, and aSIGBUSwhen it fills (that is what produced anrc=135with no FAIL line earlier tonight). A crafted run is small; a real-pack run is not, so if the second half above is ever scheduled in CI,TMPDIRmust point at disk. Recorded because it is the same note that cost a lane two measurements.bong-water-water-bong commentedon Nov 10, 2026
Closing this as addressed by #1656 (
084e929, merged), which is the fix this issue itself proposed: "have the crafted leg also write a minimum hybrid pack … and run the import against it in the same process".That is what landed, and it is the leg that runs in CI. On current
main(5eb0870),npu_pack_import_test tests/golden:So the shape this issue was worried about — the hybrid pack whose geometry lives only in
text_config— is now imported end-to-end in the leg CI runs, including the tail envelope and the container'shidden_size, rather than being "the one shape nobody exercises". TheBUILD_RC=0/TEST_RC=0verdict that was bounded by a skipped leg is now bounded by an executed one.What remains skipped, and why I am not treating it as this issue's work:
argv[2]pack leg still needs a real AMD-produced pack (~2.3 GB for the smallest one I have) and thetext-artifactsneeds the compiler'scache/*_meta.json. Neither can be committed to the tree, so that leg is a manual leg by construction, not by neglect. I ran it by hand against four 1.9 packs this week (0.5B 291 tensors / 67,088 tiles, Coder and Qwen-2.5_1.5B 339 / 188,424, Hunyuan 219 / 65,792) — which is the right home for it.If you would rather keep this open for the automated pack leg specifically, say so and I will rename the issue to that narrow scope rather than leave the broader claim standing.