Problem cache follow-on: pluggable backends, layered caches, compile-options API + offline aggregation/driver tooling - #47
Draft
danieyan-amd wants to merge 31 commits into
Conversation
Macro builders were recently added to better manage composed ops when parsing from outside libraries. This extends the torch kit with a number of missing ops required to fully migrate torch_migraphx to use this builder framework.
Adds gfx115 to hipblaslt_supported_impl() so gfx1150/1151/1152/1153 (Strix Halo/Point, RDNA3.5) use the hipBLASLt GEMM path.
…nd update tests (ROCm#5090) Rewrites the flash decoding kernel 2 recombination step in find_flash_decoding to use the exp-normalize form: `O = sum(O' * exp(LSE - max)) / sum(exp(LSE - max))` instead of normalizing weights first, then scaling and summing partial outputs. The result is mathematically equivalent but produces IR that fuses more cleanly downstream (e.g. with rewrite_broadcast in a follow-up PR).
danieyan-amd
force-pushed
the
feature/problem-cache-followon
branch
from
July 29, 2026 18:29
f33193b to
948355f
Compare
danieyan-amd
changed the title
danieyan-amd
force-pushed
the
feature/problem-cache-followon
branch
from
July 30, 2026 05:17
948355f to
caa0c6d
Compare
danieyan-amd
force-pushed
the
feature/problem-cache-followon
branch
2 times, most recently
from
August 5, 2026 21:32
3798b85 to
0ba5b7b
Compare
danieyan-amd
force-pushed
the
feature/problem-cache-followon
branch
from
August 7, 2026 00:31
8c119b5 to
6045418
Compare
added 3 commits
August 6, 2026 22:53
Introduce a type-erased problem_cache_backend (pluggable storage) with JSON and SQLite implementations; the runtime problem_cache routes has/insert/mark/get/load/save through it and canonicalizes keys so JSON round-trips match shipped caches. Move cache_device_key to its own header and route gpu/ DSL generation via generate.py. Signed-off-by: danieyan-amd <daniel.anieyan@amd.com>
Load an ordered list of read-only caches (first hit wins) alongside a writable cache; compile_ops and the gemm paths query them during tuning. Signed-off-by: danieyan-amd <daniel.anieyan@amd.com>
Read problem_cache_files from the GPU backend options and load them into the context; no dedicated API or environment variable is required. Signed-off-by: danieyan-amd <daniel.anieyan@amd.com>
danieyan-amd
force-pushed
the
feature/problem-cache-followon
branch
from
August 7, 2026 13:24
6045418 to
2911a33
Compare
Addresses review feedback: the multi-cache priority search lived in the context (read_only_caches + find_in_problem_caches). Fold it into problem_cache so has() and get() search the read-only layers (highest priority first) then the writable cache. The context now just delegates. Behaviour is unchanged: first hit wins; a single file is writable and multiple files are a read-only priority list. Signed-off-by: danieyan-amd <daniel.anieyan@amd.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
12 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
DRAFT — preview, stacked on ROCm#4835. Opened inside the fork against the
feature/problem-cache-schema-extension(ROCm#4835) branch, so the diff shows only the follow-on work. This will be retargeted toROCm/AMDMIGraphX:developand opened for review once ROCm#4835 merges — please don't merge this fork-internal draft.Builds on ROCm#4835 (device-keyed problem cache) to add the full production tooling around it. This is the second and final problem-cache PR — it turns the engine from ROCm#4835 into an end-to-end workflow for collecting, merging, shipping, and loading tuned kernel solutions. Depends on ROCm#4835 (must merge first).
What's included
problem_cache_backend+ a JSON concrete backend (json_problem_cache).sqlite_problem_cacheas a second backend (vendored SQLite amalgamation).compile_options.problem_cache_path+ aproblem_cache::load(path)overload.merge/validate/convert(device-keyed merge, conflict policy, dedup, reporting).aggregate_cache,validate_cache,convert_cachefor offline cache management.--remap-gfx).Testing
migraphx_gpu+migraphx-driverlink (only pre-existing warnings).problem_cache_backend(4),sqlite_problem_cache(5),problem_cache_path_override(2),problem_cache_aggregator(29),test_sqlite(1).aggregate_cache(first-wins) -> 672 entries;validate_cache-> valid;convert_cachejson->json and json->sqlite (672);--remap-gfxworks; error paths report cleanly (no crash) on bad options / conflicts.Notes
Single follow-on PR (all features in one), stacked on ROCm#4835. Includes two small reconciliation/hardening commits (adapt to ROCm#4835's
set_device_key(context)API; graceful driver error reporting) — squashable into their feature commits before the real upstream PR.