mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-06-24 08:54:34 +00:00
test_async_v3
1058 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Enrico Degregori
|
2733e75900 |
[rocm-libraries] ROCm/rocm-libraries#6565 (commit d41715e)
[CK Tile] Async support pipeline V3 ## Motivation Optimize pipeline V3 for gfx950 by enabling buffer load to lds (async pipeline) ## Technical Details - Add `Async` bool to `Problem` struct to enable async pipeline in existing one - Add `static_move_ys` to load transpose. This generates offset in assembly instructions saving registers - Add `is_valid` to `async_get_vectorized_elements`. Before hard coded to true. It allows to support padding - Remove unnecessary restrictions to `is_a_load_tr` and `is_b_load_tr` (wider use of lds load transpose on gfx950) - Integrate async support in existing V3 pipeline (avoid pipelines duplication) - Create policy to support both async and default cases. This could be used by any async pipeline (next steps) - Define `wg_attr_num_access` separately for A and B. This allows to optimize ds_read instruction width for cases when one matrix is transposed and the other is not. Before in such cases, `ds_read_b64` was used instead of `ds_read_b128` - Add test for V3 async. Currently only supporting cases with A and B having the same type ## Test Plan New test `test_ck_tile_gemm_pipeline_compv3_async` ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Ville Pietilä
|
60b276647b |
[rocm-libraries] ROCm/rocm-libraries#8157 (commit b0d9d39)
[CK Tile] Rule-based configuration generation in CK Dispatcher codegen (#8157) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Motivation The CK Tile Dispatcher code generation for CK Tile Profiler relies on flat JSON files to list the generated configurations. This approach has the following problems - The JSON files are verbose - The JSON files get easily out of sync with the CK Builder .config files from which they were generated from. - The JSON file based configuration make it hard to list explicitly the rules that govern the instance generation. ## Technical Details Replaced the JSON files with a rule based configuration. To preserve the existing functionality, the `profiler` and the `tests` instance sets are generated directly from the CK Builder config files. The JSON config files are removed from source control, and the "on-the-fly" generation guarantees that the Dispatcher codegen uses up to date configurations. This is PR introduces six different rule sets for the CK Tile Dispatcher code generation 1. `profiler`: matches with the old JSON set of profiler configurations. 2. `tests`: matches with the old JSON set of tests configurations. 3. `full`: full configuration set created from a rule-based config selection 4. `full-tests`: a subset of `full` for generating configurations for convolution integration tests. 5. `tiny`: a subset of `full-tests` to produce the minimal set of configurations to test the Dispatcher codegen. 6. `default`: the default rules, which corresponds to the existing heuristic rules for configuration selection. This ensures that ML based kernel selection doesn't get broken. The main use of the `full` rule set is to define a reasonable solution space for the possible implicit GEMM configurations. We start from the configurations that allowed by the device architecture. The `full` rule set defines the relevant tile sizes for each convolution direction. From the tile size we have a curated mapping to the number of waves over the different GEMM axes, i.e., we describe how many waves each GEMM dimensions corresponds to. The GEMM-K wave tile dimension can be computed from the other parameters and does not need to be listed explicitly. An orthogonal axis to the tiling strategy is the vectorization strategy. This mainly defined by the data type and hardware as in general, we want to use the maximum possible load widths. The maximum sizes for each convolution direction variant are defined by the implicit GEMM matrix dimensions. For cases where have a low number of channels per convolution group, we need smaller vector load sizes. These are captured by the `VecStrategy` enumeration in the codegen rules. The problem with the rule based configuration selection is that we "over generate" configurations. The old JSON configurations compose approximately 25% of all configuration that the `full` rule set creates. The additional configurations are valid, but they many not provide any performance benefits. Hence, we keep the `profiler` and `tests` rule set for now to avoid building an excessive amount configurations by default. The `full` rule set can be taken into use by specifying CMake configuration flag `-D DISPATCHER_RULE_SET=full`. By default, the `tests` rule set is used, i.e., we don't change the existing bahaviour. ## Test Plan Added a new stage in the CI/CD pipeline that ensures the Dispatcher codegen rules are up to date. Otherwise the functionality is covered by the existing CI/CD tests. There are no functional changes to the convolution kernels. Only how the different instances are generated. ## Test Result If the CK Tile conv instances build without errors, the Dispatcher codegen is generating valid code. If all tests in CI/CD pipeline are passing, the Dispatcher codegen generates valid instances. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
damien-lejeune
|
5bebfd460f |
[rocm-libraries] ROCm/rocm-libraries#8492 (commit 46b6a06)
Add tile size for FMHA batch prefill bf16 for MI308X ## Motivation Adding a tile size adapted to MI308X, for the FMHA Batch Prefill BF16 input type case ## Technical Details N/A ## Test Plan Benchmarking from the Aiter side with: ``` python3 op_tests/test_batch_prefill.py -s 8000 -p 1 -q 4 -k 1 --head_dim 256 -c true -d bf16 --input_dtype bf16 --quant_method none --kv_layout linear -t sglang -l 0.0 --return_lse false --profile ``` ## Test Result We see an improvement with the new tile size on MI308X (both with PLT mode OFF and ON) ## Submission Checklist - [X] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Damien Lejeune <damien.lejeune@amd.com> |
||
|
Sami Remes
|
c1f7104852 |
[rocm-libraries] ROCm/rocm-libraries#6663 (commit f19fc01)
[CKTile] Fix MX GEMM: num_loop==3 dispatch, split-K, unsupported-shape guard (#6663) Three independent MX GEMM correctness bugs reported against example/ck_tile/42_mx_gemm (fp8xfp8, A=Row/B=Col) on MI350X, plus one host-side atomic-add accumulation bug in the example's repeat loop. - Pipeline (gemm_pipeline_ag_bg_cr_comp_async.hpp): BlockHasHotloop required num_loop > PrefetchStages, which let num_loop == 3 enter a hot loop that produced 5 gemm accumulations instead of 3 (K == 3*K_Tile, e.g. K=768, deterministically wrong). Require num_loop >= 4 instead: pre-pipeline + TailNumber::Three already totals exactly 3. - Kernel (gemm_mx_kernel.hpp): split-K was silently broken because GridSize did not thread k_batch into blockIdx.z and the scale tile windows were anchored at K=0 for every k_id. Every k_id >= 1 therefore read the wrong packed scales. Fix: * GridSize returns dim3(grid_x, 1, k_batch) (persistent and non-persistent). * MakeScaleA/BBlockWindows accept a k_elem_offset and translate it to a packed-scale K offset (also apply pad_tensor_view so OOB scale loads return zero, matching A/B padding). * operator() derives k_id from blockIdx.z, uses GetSplitKElemOffset (matches Underlying::SplitKBatchOffset's K1-aligned formula), and dispatches the epilogue with memory_operation_enum::atomic_add for k_batch > 1, set for k_batch == 1. Same fp16/bf16 even-vector-size guard as UniversalGemmKernel. * MakeCBlockWindows templated on DstInMemOp; unconditionally applies pad_tensor_view using kPadM/kPadN so partial trailing M/N tiles are handled correctly. - Compile- and runtime unsupported-shape guards (gemm_mx_kernel.hpp): add IsSupportedArgument and a static_assert for configurations that produce silent wrong results: * static_assert(!kPadK) -- the MX comp-async pipeline uses async_load_tile whose OOB check is per-vector-start, so a vector straddling the K pad boundary reads garbage. Until the async path learns per-element pad masking, reject kPadK at compile time. * Runtime: k_batch >= 1; M/N multiples of MPerBlock/NPerBlock when kPadM/kPadN are false; M >= MPerBlock and N >= NPerBlock always (CShuffleEpilogue cannot safely run with a single partial tile); K % (KPerBlock * k_batch) == 0; and for k_batch > 1, K must be a multiple of WarpTile_K * k_batch so every split lands on a packed-scale boundary. * All error paths log under CK_TILE_LOGGING with actionable messages. - Example (example/ck_tile/42_mx_gemm/mx_gemm_instance.hpp): * Call Kernel::IsSupportedArgument up front and throw a clear runtime_error for rejected shapes (was silently launching an unsupported kernel). * Switch to launch_kernel_time_mask with a clear_gemm_output preprocess that zeroes C between iterations when k_batch > 1 (mirrors universal_gemm_invoker). Without this the default -warmup=50 -repeat=100 accumulated 150 atomic_adds into C after the kernel-side split-K fix. Tests (test/ck_tile/gemm_mx/): - Add MXfp8_GemmConfig16_PadMN (kPadM = kPadN = true). - test_mx_gemm_fp8.cpp: HotLoopTailNumLoopThree (K=768 regression), SplitK (k_batch=2,4 across full_k/partial_k paths), TestMxGemmFp8PadMN::{MNPaddingAligned, MPadding, NPadding, MNPadding} covering trailing partial tiles along M, N, or both. - Run(...) now takes k_batch. - packScalesMNxK: guard against OOB (mn, k) reads from src and initialise e8m0 bytes to the zero exponent (0x00) instead of the default-constructed NaN (0xFF), so padded lanes don't poison the packed int32_t shared with in-range lanes. - test_mx_gemm_instance.hpp: call IsSupportedArgument before launch. Verification on gfx950, ROCm 7.2.0: - ctest -R test_ck_tile_mx_gemm -> 100% (2/2). - Example sweep over the original bug-report shapes: all K-aligned shapes now validate correct (including 4096^3 sk=2 and the K=768 cases); all K=128 shapes cleanly rejected with the new error message instead of producing silent wrong results. Made-with: Cursor ## Motivation <!-- Explain the purpose of this PR and the goals it aims to achieve. --> ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
damien-lejeune
|
aab1d219f5 |
[rocm-libraries] ROCm/rocm-libraries#8350 (commit f92ded1)
Add tile shape for FMHA batch prefill on MI308X (on fp8, hdim=256) (#8350) ## Motivation Add a tile size appropriate for FMHA batch prefill fp8/hdim256 on MI308X ## Technical Details Appending the tile shape to the existing factory such that it can be picked up by Aiter ## Test Plan Ran the performance test on both MI300X and MI308X ## Test Result MI300X performance seems unaffected by this change. MI308X does improve. ## Submission Checklist - [X] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Damien Lejeune <damien.lejeune@amd.com> |
||
|
ltqin
|
0954a8f3fa |
[rocm-libraries] ROCm/rocm-libraries#8262 (commit d4ff8fc)
[CK_TILE] Add graph capture support for FMHA backward(new branch) (#8262) ## Motivation Add HIP graph capture support for FMHA backward operations. The original implementation only supported normal execution mode and would cause use-after-free crashes when used with graph capture replay. When FMHA backward is captured into a HIP graph: - First replay: host callback executes and deletes the closure (as designed for normal mode) - Subsequent replays: use-after-free crash because the closure was already freed This PR enables `fmha_bwd_launcher::prepare_workspace_async()` to work correctly in both normal execution and graph capture modes. |
||
|
Chao
|
320a813d67 |
[rocm-libraries] ROCm/rocm-libraries#6533 (commit 5dcaa45)
[CK_TILE] Add host-side Pack-GQA optimization for FMHA forward (#6533) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit [CK_TILE] Add host-side Pack-GQA optimization for FMHA forward ## Motivation Host-side Pack-GQA optimization for CK-Tile FMHA forward. Reshapes Q tensor from `[b, nhead_q, seqlen_q, d]` to `[b, nhead_kv, nhead_ratio * seqlen_q, d]` by adjusting strides, so grouped Q-heads sharing the same KV data are processed in a single tile. Zero kernel changes — runner-only. Phase 1: non-causal attention with GQA ratio packing. Phase 2: extends to dropout and split-kv paths, fixes stride edge cases. ## Technical Details Modified files (2): - `example/ck_tile/01_fmha/example_fmha_fwd.cpp` — Pack-GQA flag plumbing - `example/ck_tile/01_fmha/fmha_fwd_runner.hpp` — Q tensor reshape logic, stride adjustment for GQA ratio packing New files (1): - `example/ck_tile/01_fmha/test_pack_gqa_phase2.sh` — 53 test cases covering non-causal, dropout, split-kv, various GQA ratios ## Dependencies None — this PR is standalone. ## Test Plan - GPU validation on MI300X (gfx942, ROCm 6.4.1): - Command: `./build/bin/tile_example_fmha_fwd -b=2 -h=32 -h_k=8 -s=2048 -d=128 -prec=bf16 -mode=group -v=1 -warmup=1 -repeat=3` - GPU validation on MI350X (gfx950, ROCm 7.0), 53 parameterized test cases: - Command (GQA 4:1): `./build/bin/tile_example_fmha_fwd -b=2 -h=32 -h_k=8 -s=2048 -d=128 -prec=bf16 -mode=group -v=1 -warmup=1 -repeat=3` - Command (GQA 8:1): `./build/bin/tile_example_fmha_fwd -b=2 -h=64 -h_k=8 -s=2048 -d=128 -prec=bf16 -mode=group -v=1 -warmup=1 -repeat=3` - Command (decode): `./build/bin/tile_example_fmha_fwd -b=64 -h=32 -h_k=8 -s=1 -s_k=4096 -d=128 -prec=bf16 -mode=group -v=1 -warmup=1 -repeat=3` ## Test Result Benchmark results (MI350X, gfx950, ROCm 7.0): | Config | Without Pack | With Pack | Improvement | |--------|-------------|-----------|-------------| | GQA 4:1 prefill b=2 h=32 hk=8 s=2048 d=128 bf16 | 690.05 TFlops (0.199 ms) | 695.61 TFlops (0.198 ms) | +0.8% | | GQA 8:1 prefill b=2 h=64 hk=8 s=2048 d=128 bf16 | 706.25 TFlops (0.389 ms) | 729.35 TFlops (0.377 ms) | +3.3% | | GQA 8:1 decode b=64 h=32 hk=4 s_k=4096 d=128 bf16 | 305.20 GB/s (1.763 ms) | 1813.41 GB/s (0.297 ms) | **+5.9x** | | LLaMA-70B decode b=32 h=64 hk=8 s_k=4096 d=128 bf16 | 591.70 GB/s (0.909 ms) | 1820.65 GB/s (0.295 ms) | **+3.1x** | | MHA ratio=1 b=2 h=8 s=4096 d=128 bf16 | 695.16 TFlops | 702.72 TFlops | no regression | Benchmark results (MI300X, gfx942, ROCm 6.4.1): No regression on MI300X. Pack-GQA is a runner-only optimization (zero kernel changes), performance impact is within noise on MI300X. | Config | TFlops / GB/s | Time (ms) | Delta vs baseline | |--------|-------------|-----------|-------------------| | MHA bf16 b=2 h=8 s=4096 d=128 | 336.52 TFlops | 0.408 | -1.7% | | GQA 4:1 bf16 b=2 h=32 hk=8 s=2048 d=128 | 322.52 TFlops | 0.426 | -0.7% | | GQA 8:1 bf16 b=2 h=64 hk=8 s=2048 d=128 | 349.85 TFlops | 0.786 | +0.5% | | LLaMA-70B prefill b=1 h=64 hk=8 s=4096 d=128 bf16 | 381.29 TFlops | 1.442 | +1.2% | | Decode b=64 h=32 hk=8 s_k=4096 d=128 bf16 | 697.32 GB/s | 1.541 | +0.8% | All validation tests pass (`valid:y`) on both MI300X and MI350X. Additional validation: - 53 parameterized test cases pass (23 phase 1 + 30 phase 2) - GQA ratios tested: 1:1, 2:1, 4:1, 8:1, 32:1 - No regression on MHA (ratio=1) workloads - fp16 and bf16 validated |
||
|
Emily Martins
|
674f7cdc0e |
[rocm-libraries] ROCm/rocm-libraries#8141 (commit d3defa6)
[CK] Remove Stream-K from old CK ## Motivation Since Stream-K has a CK Tile implementation, we no longer need Stream-K in old CK. Hence, this PR removes Stream-K from old CK. ## Technical Details All Stream-K artifacts in old CK have been removed including examples, tests, kernels, and CK profiler artifacts. ## Test Plan Ran a CI run on the branch before publishing PR. ## Test Result All tests passed. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Claude Sonnet 4 <noreply@anthropic.com> |
||
|
Bartłomiej Kocot
|
2c363870d9 |
[rocm-libraries] ROCm/rocm-libraries#6744 (commit 9d056e8)
[Ck][CK Tile] Global Load/Store for Large Tensors support (#6744) ## Motivation Create solution to support large tensors in the entire ck tile. ## Technical Details - add possiblity to use global load - int64 indexing ## Test Plan conv fwd tests ## Test Result passed locally ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. AICK-913 |
||
|
Yung-sheng Tu
|
e826b2eb7e |
[rocm-libraries] ROCm/rocm-libraries#6768 (commit 43ca43f)
=?UTF-8?q?[CK=20TILE]=20Unification=20Work=20=E2=80=93=20?= =?UTF-8?q?Add=20MFMA=20specialisations=20for=20`tf32=5Ft`=20(#6768)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Motivation This PR adds two specialisations related to `tf32_t`. ## Technical Details This change treats `tf32_t` as a concrete type rather than an empty `struct`. It also adds two new specialisations for MFMA dense builtins and resolves existing circular include issues. ## Test Plan All the new wrappers were added to the test suite in test_amdgcn_mma_layout.inc. ## Test Result Test should pass. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Sami Remes
|
ad4e2e7624 |
[rocm-libraries] ROCm/rocm-libraries#7199 (commit 23f7320)
[CK_TILE] [QuantGEMM] Fix SplitK tail handling and other improvements (#7199) This pull request introduces improved and more robust split-K support for quantized GEMM. The main changes add runtime validation, utility functions for split-K batch calculations, pointer offset handling for split-K in grouped kernels, and enhanced support for various tensor layouts. The changes also improve error handling and provide more flexibility for runtime tail handling in split-K pipelines. **Split-K Support and Validation Enhancements:** * Added runtime validation to ensure `k_batch` is a positive integer and that split-K configurations do not produce empty final batches or mismatched pipeline tails, with detailed error messages and logging for misconfiguration. [[1]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R1184-R1211) [[2]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L1161-R1250) * Introduced utility functions `get_splitk_batch_k_read` and `get_splitk_last_batch_k` to compute per-batch K read sizes and handle split rounding, ensuring correct and consistent split-K batch partitioning. [[1]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R206-R234) [[2]](diffhunk://#diff-635b89bdffa96b2b42f1632520cde36701d7d631e864185591f6b32f7645cf47L104-R107) [[3]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L388-R417) [[4]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L1161-R1250) * Changed the default value of `k_batch` in `QuantGemmHostArgs` to 1 (no split-K) for safer default behavior. **Pointer Offsets and Grouped Kernel Handling:** * Updated `QuantGroupedGemmKernel` to apply split-K per-batch offsets to all input pointers, mirroring the behavior of non-grouped kernels and ensuring correctness for split-K launches. * Modified AQ tensor view handling to correctly reflect the remaining K-groups from the split-K batch's offset position, improving accuracy for split-K in grouped kernels. **Pipeline and Layout Flexibility:** * Added support for runtime selection of split-K tail handling via a new template parameter `RuntimeSplitKTail_`, with new helper methods to dispatch GEMM pipelines accordingly. [[1]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R273) [[2]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R1496-R1567) [[3]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L1427) [[4]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L1447-R1629) [[5]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L1459-R1641) * Improved handling for tensor layout cases, including preshuffled B and both row-major and column-major AQ layouts, ensuring correct pointer arithmetic and compatibility checks. [[1]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R438-R454) [[2]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871L464-R516) [[3]](diffhunk://#diff-d000149a681cd42bfb9947872c603e556cea26cbd7fd4f8f60afc6595d975871R1184-R1211) |
||
|
Enrico Degregori
|
7b9245f18c |
[rocm-libraries] ROCm/rocm-libraries#5854 (commit 8e2d46d)
[CK Tile] Async support preshuffle GEMM ## Motivation Add async support to existing preshuffle GEMM pipeline ## Technical Details Notes: the implementation avoids previous strategy of duplicating pipelines for async support and instead add a switch `Async` to the ops Problem to enable async pipeline. Then, integrate the async pipeline in the existing one. This allows to avoid code duplication and facilitate the integration of buffer load to lds in existing pipelines. In my opinion, it should be used also for other pipelines which don't support buffer load to lds yet and it would also be a good idea to refactor the existing async GEMM pipelines with the same approach. Summary: - integrate buffer load to lds in existing pipeline - add optimal tensor descriptors for vmem loading and lds reading. They are currently optimized for 16x16 wave tiles but they also work for 32x32 wave tiles. Optimizations for 32x32 wave tile requires different lds layout and it will be done in a follow-up issue - Add async config to examples - Add test (gfx950 only) ## Test Plan New test for gfx950 `test_ck_tile_gemm_pipeline_wp_async` ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Enrico Degregori
|
bdd7a8333d |
[rocm-libraries] ROCm/rocm-libraries#6672 (commit bda3f97)
[CK Tile] PermuteN support MX GEMM ## Motivation Add PermuteN support to preshuffle MX GEMM ## Technical Details - Modify `shuffle_b_permuteN` to support MX preshuffled layout - Add `preShuffleScalePermuteN` with same functionality of `preShuffleScale` but layout consistent with PermuteN - Include MX pre-processing functions in the library ## Test Plan Add test configuration for permuteN with preshuffle (both FP4 and FP8) ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Cong Ma <congma13@amd.com> |
||
|
John Afaganis
|
96c39b331e |
[rocm-libraries] ROCm/rocm-libraries#7829 (commit 13af7da)
[ck] Enforce ASCII-only C/C++ sources for hipRTC compatibility (#7829) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Summary CK source files must be compilable via **hipRTC (HIP runtime compilation)**, whose preprocessor does not accept non-ASCII bytes anywhere in a translation unit — **including in comments**. Bytes that are harmless under `hipcc` (em-dashes, smart quotes, multiplication signs, Greek letters, box-drawing glyphs, etc.) cause hipRTC to fail at preprocessing time. These regularly leak in via LLM-assisted authoring or copy/paste from formatted documents and silently break hipRTC paths that are not exercised by the default `hipcc`-based build matrix. This PR (a) cleans every existing violation (53 files) and (b) adds a pre-checkin gate so new violations are rejected before merge. ## File extensions covered Both the cleanup scan and the new Jenkins enforcement stage use the same predicate: ``` *.h *.hpp *.cpp *.h.in *.hpp.in *.cpp.in *.inc *.cl ``` (excluding `*/build/*` and `*/include/rapidjson/*`). This is a strict superset of the existing `Clang Format` stage's predicate — `*.inc` is added so test-fixture include files are also gated. The local pre-commit hook's `c++/inc` type filter covers the same set. ## Why no enforcement today CK is opted out of the rocm-libraries root `.pre-commit-config.yaml`, so the existing `pre-commit` workflow doesn't touch CK. The local CK `.pre-commit-config.yaml` only runs for developers who installed hooks. The **authoritative gate is therefore the new Jenkins stage** in this PR; the local hook is convenience. ## Commit layout (bisect-friendly) 1. `79798aa6261` — **`[ck] Convert reflect/ rendering to ASCII for hipRTC compatibility`** Behavior change, isolated. `TreeFormatter` swaps `├─ / └─ / │ ` for `|- / +- / | ` (3-col width preserved so alignment is unchanged). `conv_description.hpp` swaps `×` for `x` as the dimension separator. `test_conv_description.cpp` expected strings updated in lockstep so the snapshot test stays green. This is the only commit in the series with observable runtime impact. 2. `738fdb0d81c` — **`[ck] Strip non-ASCII bytes from C++ sources for hipRTC compatibility`** Mechanical text cleanup across 53 files. Replacements happen in comments or in `std::cout` strings that are not asserted on by any test. None of the 174 `.inc` files in the tree required edits, but they were in the scan's predicate so the enforcement stage's predicate is a superset of what was scanned. Full replacement table in the commit message. 3. `1d7cd8ba235` — **`[ck] Enforce ASCII-only C/C++ sources for hipRTC compatibility`** - New `projects/composablekernel/script/check_ascii_only.sh` (modeled on `check_copyright_year.sh`). - New entry in `projects/composablekernel/.pre-commit-config.yaml` under the local-hooks block (`types_or: [c++, inc]`). - New `ASCII Only Check` parallel stage in `projects/composablekernel/Jenkinsfile`'s `Static checks` block, mirroring the existing `Clang Format` stage but with `*.inc` added to the find predicate. Always-on, no `RUN_CPPCHECK` gate. The tree is buildable at every commit boundary. Commit 1 leaves 50 known violations; commit 2 leaves 0; commit 3 wires the gate. ## Demo Script output on a synthesized violation: ``` $ printf '// em-dash test \xe2\x80\x94 here\n' > /tmp/bad.cpp $ projects/composablekernel/script/check_ascii_only.sh /tmp/bad.cpp ERROR: /tmp/bad.cpp contains non-ASCII bytes: 1:// em-dash test — here Fix: replace with ASCII (em-dash -> --, smart quotes -> ", arrows -> ->, etc.) $ echo $? 1 ``` Full repo scan after the cleanup commits (note the `-name '*.inc'` clause): ``` $ cd projects/composablekernel && find . -type f \( -name '*.h' -o -name '*.hpp' -o -name '*.cpp' \ -o -name '*.h.in' -o -name '*.hpp.in' -o -name '*.cpp.in' -o -name '*.inc' -o -name '*.cl' \) \ -not -path '*/build/*' -not -path '*/include/rapidjson/*' -print0 \ | xargs -0 -P 8 -n 64 script/check_ascii_only.sh $ echo $? 0 ``` ## Test plan - [ ] Jenkins PR build: confirm new `Static checks -> ASCII Only Check` stage runs green over the full predicate (incl. `*.inc`) and existing `Clang Format` stage is unaffected. - [ ] `test_conv_description` passes against the ASCII tree-formatter output (touched in commit 1). - [ ] Local: `pre-commit run ascii-only-checker --all-files` runs cleanly after installing CK pre-commit hooks via `script/install_precommit.sh`. - [ ] Manually inject a non-ASCII byte in any `.cpp/.hpp/.inc` file, push: confirm Jenkins fails the new stage with a clear error. - [ ] Spot-check a representative subset of touched files under hipRTC compilation to confirm no remaining hipRTC-blocking content (optional, since the static byte check is a sufficient condition for hipRTC preprocessor acceptance on this dimension). 🤖 Generated with [Claude Code](https://claude.com/claude-code) |
||
|
chris-tsiaousis-hpc
|
db05d61136 |
[rocm-libraries] ROCm/rocm-libraries#6212 (commit ccee58d)
=?UTF-8?q?[CK=20TILE]=20Unification=20Work=20=E2=80=93=20?= =?UTF-8?q?More=20accurate=20tests=20for=20MmaPipelines=20(#6212)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Motivation This PR solves several issues: #### More accurate tests for MmaPipelines The current tests for the MmaPipelines (test_amdgcn_sparse_mma, test_amdgcn_wavewise_mma) use explicit input fragment vectors filled with 1s, and only check the output of a single lane. We should have tests that actually use the MmaPipelines with non-trivial input matrices and verify the complete output. Some other aspects of the current MmaPipelines tests that I noticed and deserve some attention: 1. There is sometimes iteration over K outside of the pipeline, which is then included in WaveTileK or FragK, which is not correct. We should remove it, move K iteration inside of the pipeline, or be more clear about this outer-K loop size and how it propagates downwards. 2. There is very tight coupling between the kernel, gtest code, and test_pipeline helper, requiring a lot of information and functions to be passed back and forth. 3. The test_pipeline helper is doing a bunch of register-related logic on the host (related to point 1) 4. Without this register logic the only thing it does is check the device, call the kernel, and check the output, but with a lot of boilerplate. #### Test helper for detecting target arch at HOST runtime There is a really apparent issue we faced while writing tests: Scenario: 1. Compile a test that supports both gfx950 and gfx1201 for gfx950 2. Run the test on a server that only has gfx1201 GPU Actual: Segmentation fault Expected: The test can correctly detect from HOST runtime that the DEVICE target_id was different and skips the test. Notes: The only way of detecting the COMPILER_TARGET_ID in the existing "arch" framework is launching a kernel and calling `get_compiler_target()` (so, from a DEVICE code). This will create a segmentation fault if the current arch differs from the target arch. To cope with this issue, we propose to export the compiler target(s) (note they can be many) through `projects/composablekernel/test/ck_tile/core/arch/CMakeLists.txt` and define a test helper to deal with such cases. #### Add composition support to Transforms We have a small number of Transforms which act on MmaOp input and output data, before and after the MmaOp call respectively. These are currently implemented to work on an MmaTile level, but in theory they are also supposed to work at a WaveTile level, i.e. after composition of multiple MmaTiles to create larger effective MNK dimensions. Currently the composed MmaTiles look like 2D C-style arrays of the individual MmaTile level register vectors (see WaveWiseMmaPipeline). The transforms should be able to take these and perform the proper transforms to the whole WaveTile at once. This might allow for better performing transformations. Note: This PR handles the SparseTransform case and if we don't end up doing scale as a transformation, there isn't really much left to do. If we end up having only the sparse transform as a non-trivial transform, then we could also consider removing the Transform framework. |
||
|
Anton Gorenko
|
7ecbf82708 |
[rocm-libraries] ROCm/rocm-libraries#7500 (commit f5cd4fd)
[CK_TILE][FMHA] Optimize long-context decoding on gfx11/12 (#7500) ## Motivation Relevant issue: ROCM-22065 FMHA has less-than-optimal performance of long-context decoding (i.e. when seqlen_q = 1) on gfx11/12. This PR optimizes the splitkv pipeline and configs for such scenarios. ## Technical Details Optimizations applied in this PR: 1. use tiles with smaller M0 (16 vs 64), these tiles are used when seqlen_q <= 16 2. adapt qr_nwarp_sshuffle pipeline for gfx11, it allows to use more warps even for M0 = 16 (the qr pipeline parallelizes work between warps in M dim so with M0 = 16 it allows to use only 1 warp) 3. enable kMergeNumHeadGroupsSeqLenQ (an optimization that merges one group of heads in GQA) for all hdim values, not only 128 4. increase the number of splits (multiply by the number of head groups) if (3) is used 5. increase the number of splits for RDNAs (`multiProcessorCount` is the number of WGPs on RDNAs, not CUs, so it should be doubled to have meaning similar to CDNAs) Performance on gfx1151: | Case | develop (GB/s) | This PR (GB/s) | |:-------|-------:|-------:| | [fp16\|group\|bshd] b:1, h:32/32, s:1/45056, d:64/64 | 127.58 | 183.11 | | [fp16\|group\|bhsd] b:1, h:32/32, s:1/45056, d:64/64 | 153.64 | 215.02 | | [fp16\|group\|bshd] b:1, h:16/8, s:1/77184, d:128/128 | 120.51 | 225.76 | | [fp16\|group\|bhsd] b:1, h:16/8, s:1/77184, d:128/128 | 130.62 | 223.84 | | [fp16\|group\|bshd] b:1, h:32/32, s:1/9600, d:128/128 | 82.65 | 138.44 | | [fp16\|group\|bhsd] b:1, h:32/32, s:1/9600, d:128/128 | 105.75 | 220.45 | | [fp16\|group\|bshd] b:1, h:8/1, s:1/401024, d:256/256 | 16.27 | 187.89 | | [fp16\|group\|bhsd] b:1, h:8/1, s:1/401024, d:256/256 | 16.28 | 188.19 | ## Test Plan An additional test case is added to the exiting test. It uses seqlen_q = 1, GQA, no mask to trigger the changes ``` ninja test_ck_tile_fmha_fwd_fp16 && bin/test_ck_tile_fmha_fwd_fp16 --gtest_filter="*SplitKV* ninja test_ck_tile_fmha_fwd_bf16 && bin/test_ck_tile_fmha_fwd_bf16 --gtest_filter="*SplitKV* ``` Manual testing can be done with these commands: ``` bin/tile_example_fmha_fwd -prec=fp16 -mode=1 -page_block_size=128 -b=1 -h=32 -h_k=32 -d=64 -s=1 -s_k=$((352 * 128)) -lse=1 -mask=0 -num_splits=0 -kname=1 -v=1 bin/tile_example_fmha_fwd -prec=fp16 -mode=1 -page_block_size=128 -b=1 -h=16 -h_k=8 -d=128 -s=1 -s_k=$((603 * 128)) -lse=1 -mask=0 -num_splits=0 -kname=1 -v=1 bin/tile_example_fmha_fwd -prec=fp16 -mode=1 -page_block_size=128 -b=1 -h=32 -h_k=32 -d=128 -s=1 -s_k=$((75 * 128)) -lse=1 -mask=0 -num_splits=0 -kname=1 -v=1 bin/tile_example_fmha_fwd -prec=fp16 -mode=1 -page_block_size=128 -b=1 -h=8 -h_k=1 -d=256 -s=1 -s_k=$((3133 * 128)) -lse=1 -mask=0 -num_splits=0 -kname=1 -v=1 ``` ## Test Result All the tests must pass. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Hosang Yoon
|
e7e8801dc3 |
[rocm-libraries] ROCm/rocm-libraries#7586 (commit c18f2c7)
[CK_TILE] Use gfx11 float buffer atomics in FMHA Bwd ## Motivation FlashAttention CK backward on gfx11 can hit out-of-bounds/tail writes in the dQ accumulator atomic-add path when sequence rows are padded at the tile level but not marked invalid in the DQDKDV main tensor view. With the generic global atomic fallback, an incorrectly-valid tail element can issue an actual pointer-based `atomicAdd`. With the buffer atomic path, the write is issued through a buffer resource with bounds information and follows the same backend already used by gfx9/gfx12. This fixes the gfx11 FMHA BWD failure without changing the gfx11 default for unrelated CK Tile kernels. ## Technical Details This PR enables the existing CK Tile AMD buffer float atomic-add path only for generated FMHA BWD gfx11 translation units. gfx11 normally uses the generic global atomic fallback for floating-point `buffer_view::atomic_add`. That fallback performs the atomic through a raw computed pointer and depends on the software validity predicate to avoid invalid elements. In FMHA BWD dQ accumulation, padded tail rows can reach this path, so using the buffer atomic backend is safer: it uses a buffer resource with base pointer, bounds information, and an element offset, matching the backend already used by gfx9/gfx12. Enabling `CK_TILE_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT` globally for gfx11 is too broad and can break unrelated gfx11 CK builds such as GEMM. Instead, `config.hpp` now preserves an explicitly pre-defined `CK_TILE_USE_AMD_BUFFER_ATOMIC_ADD_FLOAT`, while keeping the existing default disabled for gfx11. ## Test Plan Validated the change with the FlashAttention CK full test suite with backward pass enabled on gfx11. pytest -q -s tests/test_flash_attn_ck.py ## Test Result FlashAttention CK gfx11 test result: 260680 passed, 152076 skipped ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com> |
||
|
Andriy Roshchenko
|
d5c9215064 |
[rocm-libraries] ROCm/rocm-libraries#7359 (commit dd62f9f)
[CK_TILE][GFX1250] Enable MX GEMM FLATMM with ASYNC ## Motivation Enables MX GEMM FLATMM pipeline on gfx1250. The pipeline uses an async load instruction for tensor A, which complements the existing MX GEMM FLATMM pipeline with TDM load. At this time, only FLATMM MX pipelines are enabled on gfx1250. ## Technical Details The existing gfx950 implementation was extended to support gfx1250 architecture. All three MX FP data types are supported across the two ASICs. It should be noted that while the TDM pipeline uses an emulated 32x32x128 warp-tile instruction, the present submission relies on the built-in 16x16x128 instruction, called 4 times per warp. ## Test Plan Existing `test/ck_tile/flatmm` tests were extended to cover new gfx1250 functionality. To help facilitate the testing in development, `example/ck_tile/18_flatmm/script/smoke_test_mx.sh` script was introduced to verify various combinations of supported data types and pipeline versions. ## Test Result The present submission is expected to work on both gfx950 and gfx1250 hardware for all reasonable sizes and all MX FP8/FP6/FP4 data types. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. - [x] Relies on #6978 and should only be merged after the changes are merged to the `develop`. |
||
|
Illia Silin
|
8bd8094012 |
[rocm-libraries] ROCm/rocm-libraries#7833 (commit 8a444cd)
[CK] Replace deprecated load_module function in python (#7833) ## Motivation Recent pytorch builds with python 3.15 failed in CK due to deprecation of load_module function. This should fix the issue. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
ltqin
|
bf07a0150e |
[rocm-libraries] ROCm/rocm-libraries#7723 (commit 4ed6c51)
[CK Tile] Enable LSE output for fp8bf16 V3 FMHA kernels (#7723) ### Motivation The V3 pipeline (qr_async_trload_v3) for fp8bf16 FMHA kernels did not support LSE (Log-Sum-Exp) output. This PR enables LSE output support for fp8bf16 V3 FMHA kernels, allowing users to retrieve attention statistics alongside attention outputs. ### Technical Details - StandardAttention: lse = softmax_scale * m + log(l) - LogitsSoftCap: lse = (m / log2(e)) + log(l) ### Test Plan Run FMHA forward example with fp8bf16 precision and LSE output enabled: - Test 1: Basic LSE functionality ./build/bin/tile_example_fmha_fwd -v=1 -b=1 -h=8 -s=1024 -d=128 -prec=fp8bf16 -init=3 -qscale=1 -lse=1 - Test 2: LSE with LogitsSoftCap (CMakeList should remove Logits filter) ./build/bin/tile_example_fmha_fwd -v=1 -b=1 -h=8 -s=1024 -d=128 -prec=fp8bf16 -init=3 -qscale=1 -lse=1 -logits_soft_cap=30.0 |
||
|
Zoltán Lakatos
|
58e2ab1fc7 |
[rocm-libraries] ROCm/rocm-libraries#6761 (commit d19f6f1)
[CK] Large tensor gemm workaround (#6761) ## Motivation Customer qeruested large tensor gemm support for 8bit and 4bit data types. Currently CK triggers “This GEMM not supported” error. The root cause appears to be the 2 GB limit on the input/output matrix, triggered by buffer offset constraints when testing a larger shape such as M = 699,904 (which is an exact multiple of MPerBlock = 256). ## Technical Details Quick workaround to have support ASAP. Split the tensors into inputs / outputs smaller than 2GB limit. Iterate on host and call all subproblems without device code change. Support is restricted to rowise layout in A, Ds and E All changes were implemented in DeviceGemm structures to avoid secondory affect on grouped convolutions. Got lots of AI generated comments. Addressed the ones that seemed relevant on the functionality. ## Test Plan Within CK the following examples can be used with modified input sizes: example_gemm_multiply_multiply_xdl_fp8 example_gemm_mx_fp4 Tested with Aiter tuning on provided shapes. ## Test Result All gemms run and provide correct results. ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Zoltán Lakatos <zoltan.lakatos@streamhpc.com> Co-authored-by: Márton Bidlek <marton.bidlek@streamhpc.com> Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com> |
||
|
Illia Silin
|
c24e528481 |
[rocm-libraries] ROCm/rocm-libraries#7760 (commit a61bc76)
[CK] suppress compiler warnings while building pytorch. (#7760) ## Motivation Recently added compiler flags that are required to suppress false warnings by latest staging compiler are not recognized by older compiler versions and are triggering an avalanche of warnings. Previous attempt to suppress them by using -Wno-unknown-warning-option flag didn't help, because that flag wasn't recognized either and just added more warnings. I've verified that current approach by checking the clang version actually works as intended and makes the warnings go away. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
JH-Leon-KIM-AMD
|
00e1d82ae7 |
[rocm-libraries] ROCm/rocm-libraries#7732 (commit b0e29d9)
[CK] Fix grouped conv bwd data stride>1 silent miscompute (ALMIOPEN-1959) (#7732) ## Motivation Fix silent miscompute in the grouped convolution backward-data kernel (`DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1`) when stride > dilation (ALMIOPEN-1959). PR #6208 introduced a flat-descriptor fast path that dropped all but the first sub-GEMM, producing zeroed slices of `dx` on the (G=1, stride>1, 2D, NumDTensor=0) intersection. Restore correctness without giving up the perf gains PR #6208 delivered on stride=1 shapes. ## Technical Details - Tighten the flat-descriptor fast-path gate to require `arg.gemms_count_ == 1` (i.e. a single sub-GEMM per dispatch — its original purpose). For stride > 1, the implicit GEMM is split into `gemms_count_` sub-GEMMs whose output cells tile `dx` disjointly; routing them through the flat path required dropping all but the first, which was the source of the bug. - Stride > 1 now falls through to the existing grouped CShuffle path, which packs all sub-GEMMs into one descriptor array and walks them on-device in a single kernel launch. This is the pre-PR-6208 production path; correctness is established and per-dispatch launch count is minimised. - Add regression coverage for the (G=1, stride>1, 2D, NumDTensor=0) intersection in `test/grouped_convnd_bwd_data/test_grouped_convnd_bwd_data.cpp` with `gemms_count` ∈ {4, 9, 36}. Pre-existing cases did not hit this intersection (all stride>1 cases used G=2; all G=1 cases used stride=1), which is why PR #6208's regression slipped past CI. ## Test Plan - `ctest -L SMOKE_TEST -R 'grouped_convnd_bwd_data'` on gfx942 (smoke tier — runs on every PR via `smart_build_and_test.sh`). - End-to-end verify (`verify=1`) via `example_grouped_conv_bwd_data_xdl_fp16` on stride 1/2/3/6 shapes including the original ALMIOPEN-1959 case and a cross-bucket (`gemms_count=36`) case spanning two `MaxGroupedGemmGroupsNum=32` buckets. - ckProfiler A/B sweep on MI300X (gfx942) toggling the flat-path gate via an environment variable: full kernel-family enumeration, winning kernel + its avg_time reported under each gate. 33/41 shapes completed before the sweep was stopped; the remaining 8 were the largest i2v/synthetic shapes where ckProfiler exceeded its 300s per-shape enumeration budget (not relevant to the verdict). ## Test Result ### Correctness | Test | Result | |---|:---:| | `test_grouped_convnd_bwd_data` (12 type parameterizations × Test2D, includes 3 new regression shapes) | **12/12 PASSED** in 14.18 s | | `test_grouped_convnd_bwd_data_interface` (API checks) | **PASSED** in 0.28 s | | ALMIOPEN-1959 stride=2 (`verify=1`) | **PASSED** | | stride=1 K3 (`verify=1`) | **PASSED** | | stride=3 K3 `gemms_count=9` (`verify=1`) | **PASSED** | | stride=6 K6 `gemms_count=36` cross-bucket (`verify=1`) | **PASSED** | ### Performance (ckProfiler A/B on gfx942 / MI300X) Comparing the **post-fix gate** (flat path only when `gemms_count_==1`, column "B") vs the **inner-loop variant** that keeps the flat path on stride>1 (column "A") across 25 stride>1 shapes where production picks a `_v1` instance (so the gate actually fires): | Stride | Shapes | A wins | Tie | B wins | Notes | |:------:|:------:|:------:|:---:|:------:|---| | 1 (sanity, gate moot) | 3 | 0 | 3 | 0 | gate doesn't differentiate — A == B as expected | | > 1 (gate fires) | 25 | **0** | 11 | **14** | B wins +6% to +32%; A never wins | Highlights from the firing-gate cases: | Shape (G=1, stride=2 unless noted) | A ms | B ms | B vs A | |---|---:|---:|---:| | ALMIOPEN-1959 (N=16, K=256, C=128, 5×5, 40×175) | 0.183 | 0.171 | **B +6%** | | Retinanet-L61 (N=32, K=C=256, 3×3, 25×25) | 0.054 | 0.045 | **B +17%** | | i2v-010 (N=1, K=C=384, 3×3, 277×209) | 0.174 | 0.125 | **B +28%** | | Synthetic 50×50 K3 N=32 K=C=256 | 0.131 | 0.088 | **B +32%** | Why B wins everywhere the gate fires: for `gemms_count = N`, the flat path needs N kernel launches (one per sub-GEMM), while the grouped path loops over the same N sub-GEMMs on-device in 1 launch. The (N−1) × launch-tax is a structural disadvantage A can't recover from. ### Diff | File | Lines | |---|---:| | `include/.../device_grouped_conv_bwd_data_multiple_d_xdl_cshuffle_v1.hpp` | +14 / −8 (one extra condition + expanded dispatch comment) | | `test/.../test_grouped_convnd_bwd_data.cpp` | +9 / −0 (3 new shapes) | ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Aviral Goel
|
0df3523ef1 |
[rocm-libraries] ROCm/rocm-libraries#6807 (commit ddda8ac)
[CK_TILE] Add save_matrix_txt() and extract HostTensor I/O to free functions (#6807) ## Summary - Extract `loadtxt`, `savetxt`, and `save_matrix_txt` from `HostTensor` member functions into standalone free functions in `host_tensor_io.hpp` (Single Responsibility Principle) - Add `save_matrix_txt()` for writing 2D tensors to space-separated text files with configurable output limit (default 256x256, pass 0 to dump all) - Supports float, int, and int8_t output formats via a `dtype` parameter - Validate dtype early and throw on unsupported values in all three functions - Update callers in `15_fused_moe/main.cpp` to use free function syntax |
||
|
Anton Gorenko
|
66d6714376 |
[rocm-libraries] ROCm/rocm-libraries#5388 (commit 45583bd)
[CK_TILE][FMHA] Improve precision of mxfp4 FMHA with fp6 for matrix P (#5388) ## Motivation Improve precision of mxfp4 without performance penalties. ## Technical Details Since performance of scale MFMAs is the same when neither A nor B is fp8/bf8, it is possible to use fp6 x fp4 instead of fp4 x fp4 for the second GEMM, while types of Q, K, V stay the same. This allows to improve overall precision significantly because fp6 has 32 non-negative values used for P quantization compared to just 8 values for fp4. It was found that there is a compiler bug with `__builtin_amdgcn_cvt_scalef32_2xpk16_fp6_f32` (described in LCOMPILER-561) but a workaround seems to fix all failing instances. ## Test Plan ``` ninja test_ck_tile_fmha_fwd_mxfp4 && bin/test_ck_tile_fmha_fwd_mxfp4 ``` ## Test Result The tests must pass. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Yi DING
|
6a9c03f692 |
[rocm-libraries] ROCm/rocm-libraries#7450 (commit 402dbad)
[CK_TILE] Use Persistent Scheduling for FMHA BWD Group Deterministic (#7450) ## Motivation FMHA BWD group-mode deterministic currently uses a non-persistent scheduler: each `(batch, head, K-row)` work-item is launched as its own block, with no work-stealing across CUs. On uneven workloads (varlen, GQA, many heads with few K-rows) this leaves CUs idle and forces a larger dq_acc workspace than necessary. This PR ports the persistent + deterministic scheduling already used in batch mode to group mode: a fixed-grid kernel that pre-computes per-CU work ranges on the host and uses sparse dq_acc slot indexing so multiple K-rows handled by the same CU share one accumulator slot via intra-CU atomic adds. Stacked on #7331; merge that first. ## Technical Details Single file changed: `ops/fmha/kernel/fmha_bwd_kernel.hpp`. A new `kUsePersistent` path is added to the group-mode deterministic kernel, mirroring the batch-mode persistent scheduler. The host pre-computes a fixed per-CU partition of the total `(batch, head, K-row)` work and packs it into `cu_states[]` so the GPU consumes it in a single launch. Host preparation happens in four steps: 1. Build per-batch `seqstart` prefix sums. 2. Fill per-batch `(sq_w, nc)` with a placeholder `nsplits` (bumped in step 3). 3. Two-pointer scan over CUs to fill `cu_states[c]` (`isplit`, `head_start`, `c_start`, `w_lo`, `w_hi`), accumulating `nsplits[b]` as `max(cs->isplit + 1)`. 4. Compute compact per-batch dq_acc offsets from the finalized `nsplits`. `isplit` is the sparse dq_acc slot index — one CU's multi-K-row writes share slot `ceil(wc_start / denom)`, enabling intra-CU atomic accumulation instead of one slot per K-row. `denom = max(sq_w, target_w)`, splitting two regimes: - `target_w >= sq_w` (large work): `denom = target_w`, intra-CU atomic optimization engaged. - `target_w < sq_w` (sub-K-row sharding, multiple CUs sharing one K-row): `denom = sq_w` collapses to per-K-row indexing (`= c_start`), keeping `isplit ∈ [0, nc-1]` and matching the `nsplits_max = ceil(s_k/kN0) = nc` upper bound that #7331's `GetWorkspaceDeviceSizeUpperBound` assumes for group+det. `isplit` is additionally clamped to `nc-1` to absorb empty CUs (rounded-up `wc_start` past the last K-row); they don't write dq_acc on GPU so the slot value is harmless. `nsplits[b]` is accumulated dynamically in step 3 rather than via a closed form so it tightly matches the actual sparse slots used; step 4 (offsets) follows step 3 since offsets now depend on the dynamic `nsplits`. Group mode also allows batches with `seqlen_q == 0`. The persistent scheduler skips them on the dQ path (no work) but dK/dV are still zero-filled. ## Test Plan Built `tile_example_fmha_bwd` with receipt 5 (fp16, no-bias, no-dropout, `dpad == dvpad`, group + batch) on gfx950 (MI355X). - 8-case smoke (shapes that exercise the sub-K-row regime). - 44-case sweep covering: mask 0/1/2, GQA, var seqlen, `d != d_v`, extreme small seqlen / `nc=1`, CU >> work, huge batch, batch-mode regression. - 12-case perf comparison vs the non-persistent baseline (warmup=10, repeat=50). ## Test Result - All 8 + 44 cases `valid:y`. - Perf: ±5% noise, average -0.4% across the 12 cases — neutral. - Batch-mode deterministic / non-deterministic regression unchanged. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Michal Kulikowski
|
8de4cb72fb |
[rocm-libraries] direct push (commit 49b73ad)
[CK][CK_TILE] POC for Instruction Cache prefetch. Signed-off-by: Michal Kulikowski <Michal.Kulikowski@amd.com> |
||
|
JP-Fernando
|
74bc86240b |
[rocm-libraries] ROCm/rocm-libraries#5647 (commit 490437a)
[CK Tile] Add gemm universal preshuffle to MX GEMM (#5647) ## Motivation Add gemm universal preshuffle support to existing MX GEMM pipeline. The straightforward way to do this is to port the `mx_flatmm` pipeline to the existing `gemm_mx` framework. ## Technical Details The `mx_flatmm` pipeline was not deleted, to allow for back-compatibility. ## Test Plan Add `preshuffle` option to example: `tile_example_mx_gemm`. Add new configurations with enabled preshuffle to the existing `test/ck_tile/gemm_mx` tests. ## Test Result Example and tests were successful on `gf950` architecture in the `Alola` cluster. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Fernando Jiménez <fernando.jimenez@streamhpc.com> |
||
|
Bartłomiej Kocot
|
ebb97044f4 |
[rocm-libraries] ROCm/rocm-libraries#7664 (commit de5d6b1)
Revert "[CK] Enable grouped conv bwd data to match non-grouped perf" (#7664) ## Motivation Incorrect results has been introduced for some conv bwd cases. ## Technical Details This reverts commit 33424f65346d6330d0fd94b5a4e6f843f24e52c3. ## Test Plan CI ## Test Result Pending ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. ALMIOPEN-1959 |
||
|
Illia Silin
|
e02c566795 |
[rocm-libraries] ROCm/rocm-libraries#7612 (commit 5427d24)
[CK] upgrade CI to rocm7.13 as default compiler (#7612) ## Motivation Upgrade the default docker and compiler version in CI to rocm7.13. In order to pass all the checks I had to also clean up a lot of non-ascii characters in the source code comments and modify a couple of tests that were affected by a new compiler logic. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Aviral Goel <aviral.goel@amd.com> |
||
|
kensclin
|
fc2862d712 |
[rocm-libraries] ROCm/rocm-libraries#6846 (commit 377def4)
[CK_TILE] Add fmha forward hdim 256 support (#6846) ## Motivation Enable Composable Kernel FMHA forward kernel for **hdim=256 BF16** on AMD gfx950 (MI350X). Prior to this change the (256, 256) head-dim configuration either failed to compile, was filtered out by the compatibility rules, or produced incorrect kernel output due to an LDS layout accounting bug. ## Technical Details Four files changed, all to enable hdim=256 BF16 on gfx950. - **`fmha_fwd.py`** — Allow `(256, 256)` in gfx950 compatibility rule; set `(256,256)` BF16 tile to `M0=128, N0=64` (the LDS-feasible shape on gfx950); emit minimal valid instance set for d=256 to bound compile time. - **`fmha_fwd_kernel.hpp`** — Gate Prefill launch path off for d=256 (`PrefillCase = kM0 > 64 && kQKHeaddim < 256`); the double-buffer Prefill variant overflows the 160 KB LDS budget. - **`trload_policy.hpp`** — **Critical correctness fix**: the LDS layout accounting in `GetSmemSize` was wrong (`max(Q, K+S+V)` instead of `max(Q, K) + V + S`), under-allocating LDS and silently corrupting d=256 output (~2% wrong values). - **`trload.hpp`** — Thread `LoadOnce=true` through all d=256 K-LDS descriptors so the compiler picks the matching XOR swizzle period; recompute the S-tile LDS offset to match the corrected `GetSmemSize` formula. ## Test Plan Built and ran `tile_example_fmha_fwd` on gfx950 (MI350X) with the canonical d=256 BF16 configurations: ```bash cd build && ninja tile_example_fmha_fwd ./bin/tile_example_fmha_fwd -prec=bf16 -d=256 -d_v=256 -b=1 -h=32 -h_k=2 -s=1024 -s_k=1024 -bias=n -mask=t -lse=0 -p_drop=0 -warmup=3 -repeat=10 -kname=1 -v=1 ./bin/tile_example_fmha_fwd -prec=bf16 -d=256 -d_v=256 -b=8 -h=32 -h_k=2 -s=16384 -s_k=16384 -bias=n -mask=t -lse=0 -p_drop=0 -warmup=3 -repeat=10 -kname=1 -v=1 ``` ## Test Result ```bash -b=1 -s=1024 [bf16|batch|bhsd] b:1, h:32/2, s:1024/1024, d:256/256, scale_s:0.0625, bias:n, p_drop:0, lse:0, qscale:n, mask:t(-1:0), v:r, fmha_fwd_d256_bf16_batch_b128x64x32x256x32x256_r4x1x1_r4x1x1_w32x32x16_w32x32x16_qr_async_trload_vr_psddv_nlogits_nbias_mc_nlse_ndropout_nskip_nqscale_ntrload_nsink, 0.058 ms, 298.42 TFlops, 618.68 GB/s, valid:y -b=4 -s=16384 [bf16|batch|bhsd] b:8, h:32/2, s:16384/16384, d:256/256, scale_s:0.0625, bias:n, p_drop:0, lse:0, qscale:n, mask:t(-1:0), v:r, fmha_fwd_d256_bf16_batch_b128x64x32x256x32x256_r4x1x1_r4x1x1_w32x32x16_w32x32x16_qr_async_trload_vr_psddv_nlogits_nbias_mc_nlse_ndropout_nskip_nqscale_ntrload_nsink, 42.797 ms, 822.18 TFlops, 106.63 GB/s, valid:y ``` ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: poyenc <1132573+poyenc@users.noreply.github.com> |
||
|
JP-Fernando
|
e7798e9560 |
[rocm-libraries] ROCm/rocm-libraries#7112 (commit a6e5eac)
Add asynchronous XOR shuffle support to the Async GEMM pipeline and the MX GEMM pipeline (#7112) ## Motivation The goal of this work is to apply XOR shuffle (swizzle) to the current `comp_async` GEMM pipeline and the `gemm_mx` pipeline. XOR swizzling has been helpful to avoid LDS bank conflicts, as data are redistributed across LDS banks, such that simultaneous threads accessing different rows land on different LDS banks. ## Technical Details A similar approach to the work in the existing eight-waves pipeline was followed. Currently, XOR swizzle support is available for FP8 and BF8 types. FP4 support is also available for MX GEMM. Should the types not match, or should the async vector width be of an unsupported size, then the pipeline falls through to the previously existing ('unswizzled') path. ## Test Plan Execute `test_ck_tile_gemm_pipeline_comp_async` for the Async GEMM pipeline. Execute `test_ck_tile_mx_gemm_fp8` and `test_ck_tile_mx_gemm_fp4` for the MX GEMM pipeline. ## Test Result The tests passed successfully in the `Alola` cluster with MI350 hardware. ## Submission Checklist - [X] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Fernando Jiménez <fernando.jimenez@streamhpc.com> Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> |
||
|
Kiefer van Teutem
|
b5f8bef97f |
[rocm-libraries] ROCm/rocm-libraries#6088 (commit 6ac353c)
[CK Tile][MFMA/WMMA unification] Add support for packed datatypes (tiny types) (#6088) ## Motivation This MR makes all the changes required for the unified architecture to be able to deal with packed datatypes i.e. int4, fp4, fp6, and bf6. The crux is that layout parameters should be interpreted as describing the pure mathematical matrix fragments, while the ext_vectors and tile distribution encodings describe everything in terms of packed datatype units. This matches how packed types are dealt with in ck_tile and should play nicely with the load and store tile ops once we integrate the unified framework into CK tile. The bf6 datatype was added to CK tile in the form of pk_bf6x16_t and pk_bf6x32_t, which did not exist before. The ext_vector implementations of pk_fp6x16_t and pk_bf6x16_t (vec size 1 and 2) were extended to make the subscripting operator work as expected. The layout test was adapted to be compatible with all packed datatypes, and all new intrinsics were added to the test. This MR adds ALL intrinsics across ALL architectures which use packed datatypes, as well as ALL scale intrinsics: mfma_scale_f32_16x16x128_f8f6f4 gfx950 (F8xF8, BF8xBF8, F4xF4, F6xF6, BF6xBF6) mfma_scale_f32_32x32x64_f8f6f4 gfx950 (F8xF8, BF8xBF8, F4xF4, F6xF6, BF6xBF6) wmma_i32_16x16x16_iu4_w32 wmma_i32_16x16x16_iu4_w32_gfx12 wmma_i32_16x16x32_iu4_w32_gfx12 ## Testing All intrinsics were tested on all architectures. |
||
|
Enrico Degregori
|
9565ca21ec |
[rocm-libraries] ROCm/rocm-libraries#5552 (commit 369c7a2)
[CK Tile] Eight Waves pipeline for MX GEMM (#5552) ## Motivation Integrate Eight Waves pipeline in MX GEMM ## Technical Details - EightWaves pipeline: - Add pipeline, policy and block gemm (internally using existing implementation used by GEMM and ABQuant) - Extend support of EightWaves policy for FP4 (packed types) - Async pipeline: - Fix pipeline with packed scales (requires MRepeat and NRepeat to be contiguous) - block gemm specific for MX GEMM is defined because distribution encodings have changed - CShuffle: - Add new functionality to support MRepeat and NRepeat contiguous (defined by `TilesPacked`) - Examples: - Refactor examples to easily switch different configurations (similar to GEMM universal) - Scales values generated consistently with other microscale implementations in CK Tile - Add configuration for EightWaves pipeline - Tests: - Unify existing FP8 and FP4 tests - Add tests for EightWaves pipeline - Scales values generated consistently with other microscale implementations in CK Tile Note: FP6 support for MX GEMM was added later and the support for the Eight Waves pipeline will be done in following PR ## Test Plan Add new pipeline to tests: `test_ck_tile_mx_gemm_async` for both FP4 and FP8 ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Po Yen Chen
|
a11f53564f |
[rocm-libraries] ROCm/rocm-libraries#7530 (commit 378e049)
[CK] Fix FMHA sink dispatch when init_sink_value is set (#7530) ## Summary - Fix `traits.has_sink` in `fmha_fwd_runner.hpp` to also check `init_sink_value != 0`, so the GPU kernel dispatches with sink support when `-init_sink=1` is passed. - Gate `run_sink_mask_tests` (StreamLLM) and `run_sink_init_tests` (GPT-OSS) behind opt-in flags `-m` and `-g` in `smoke_test_fwd.sh`. These tests require sink=true kernel instances which are excluded by the `BUILD_TESTING` CMake filter (`*_nsink*`), causing unconditional "not supported yet" failures (48 tests in CI). The opt-in flag approach was borrowed from PR #6057. ## Why gate tests instead of compiling sink=true kernels? The `BUILD_TESTING` filter in `CMakeLists.txt` uses `*_nsink*` glob patterns for the `fwd` and `fwd_splitkv` APIs, excluding sink=true kernel instances from compilation. We chose opt-in flags over widening the filter because: - **Compile time**: Enabling sink=true kernels doubles the kernel variants for `fwd` and `fwd_splitkv` APIs. The filter exists specifically to reduce CI build times. - **Incremental enablement**: Sink support (StreamLLM / GPT-OSS) is still maturing. Gating lets teams opt in explicitly (`smoke_test_fwd.sh -g`) while keeping the default CI path fast. - **Precedent**: splitkv (`-s`) and appendkv (`-a`) tests already follow this opt-in pattern. ## Test plan - [ ] Run `smoke_test_fwd.sh -g` with sink=true kernels compiled and verify sink-enabled kernels are dispatched - [ ] Verify `smoke_test_fwd.sh` still passes without `-m` / `-g` flags - [ ] Confirm CI no longer fails on sink tests (they are now opt-in) |
||
|
Johannes Graner
|
3727d5220a |
[rocm-libraries] ROCm/rocm-libraries#5652 (commit 7dc7d1d)
[CK Conv] Wavelet gemm pipeline for bwd_weight convolution (#5652) ## Motivation In the current CShuffleV3 backward weight kernel, the in-kernel conv-to-GEMM transform generates significant INT32 VALU pressure per MFMA instruction. On VALU-heavy shapes (e.g., G=1, 3×3, C=256), these index computation ops compete with MFMA for VALU issue slots, creating a bottleneck that cannot be resolved by pipeline prefetching alone. This PR adds a wave-specialized ("wavelet") convolution backward weight kernel that splits workgroup threads into two roles: - **Load waves**: conv-to-GEMM address computation + global memory loads + LDS writes (all VALU/VMEM) - **Math waves**: LDS reads + MFMA + CShuffle epilogue (no index computation) By physically separating the two instruction classes onto different waves, VALU and MFMA execute on different hardware functional units without contention. ## Technical Details **Core kernel (new files):** - `gridwise_gemm_xdl_waveletmodel_cshuffle_conv_v3.hpp` — wave-specialized gridwise GEMM for conv bwd weight (2-way split: load + math) - `device_grouped_conv_bwd_weight_xdl_waveletmodel_cshuffle_v3.hpp` — device op following CShuffleV3 patterns; `BlockSize = TileMathThreadGroupSize` for MFMA wave assignment, `LaunchBlockSize = TileLoad + TileMath` for kernel launch **Wave pipeline (modified):** - `gridwise_gemm_waveletmodel.hpp` — load/math wave pipeline structs with `sched_group_barrier` scheduling hints to front-load VMEM reads before address-advance VALU **Two wave ratios:** - **(4,4)**: 256 load + 256 math = 512 threads (8 waves). Best on large shapes. - **(4,2)**: 256 load + 128 math = 384 threads (6 waves). Best on small shapes (fewer sync barriers, denser MFMA per math wave). **Instance coverage (F16 and BF16 symmetric):** | Ratio | Tiles | Layouts | ConvSpecs | |-------|-------|---------|-----------| | (4,4) | M128×N128, M64×N64, M128×N64, M64×N128 | 2D NHWGC, 3D NDHWGC | Default, Filter1x1Stride1Pad0 | | (4,2) | M64×N64, M128×N64, M64×N128 | 2D NHWGC | Default, Filter1x1Stride1Pad0 | **Existing wavelet model fixes:** - `BlockSize` corrected from `math::max(TileLoad, TileMath)` to `TileMathThreadGroupSize` in the flat-GEMM wavelet device op and gridwise kernel ## Test Plan - `test_grouped_convnd_bwd_weight` GTest: 34 hardcoded test cases covering 1D/2D/3D, F16/BF16, G=1/2/16, various spatial sizes - Performance benchmark: all 37 RetinaNet bwd_weight shapes on gfx950 ```bash ninja -C build test_grouped_convnd_bwd_weight ./build/bin/test_grouped_convnd_bwd_weight ``` ## Test Result **Correctness:** 34/34 GTest cases passed (F16/BF16 × 1D/2D/3D × Default/Filter1x1Stride1Pad0 × various G/N/K/C combinations). **Performance:** Wavelet is the fastest overall instance on 12/37 RetinaNet shapes — all G=1, 3×3 convolutions with C=256 (the VALU-heavy target shapes): | Shape | Uplift vs best baseline | |-------|------------------------| | K=36, 7×7 | 1.91x | | K=36, 100×100 | 1.60x | | K=36, 13×13 | 1.43x | | K=36, 25×25 | 1.38x | | K=36, 50×50 | 1.38x | | K=256, 100×100 | 1.24x | | K=256, 13×13, s=2 | 1.20x | | K=256, 25×25, s=2 | 1.20x | | K=256, 7×7 | 1.17x | | K=256, 13×13 | 1.13x | | K=2376, 50×50 | 1.05x | | K=2376, 100×100 | 1.06x | Where wavelet does not win (25/37): 1×1 convolutions (explicit kernel does host-side transform), grouped convolutions with small per-group channels, and shapes where standard CShuffleV3 already amortizes VALU overhead. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: jakpiase <jakpia21@gmail.com> |
||
|
JH-Leon-KIM-AMD
|
9a5d1ea791 |
[rocm-libraries] ROCm/rocm-libraries#6208 (commit 33424f6)
[CK] Enable grouped conv bwd data to match non-grouped perf via NoShuffle + packed descriptors (#6208) ## Motivation Improve performance of grouped convolution backward-data kernels to match non-grouped kernel performance for G=1 cases. ## Technical Details - Add NoShuffle epilogue path (direct VGPR→Global writes) by setting `CDEBlockTransferScalarPerVector_NPerBlock = 1` - Add nongrouped-match instances with optimized BBlockTransfer parameters for better thread utilization - Add packed (flat) descriptor path for G=1 2D convolutions, using simpler tensor descriptors with fewer transform layers to reduce address computation overhead in the GEMM main loop - Cherry-pick PR #6090 for fair benchmarking (cache flush, include dX zeroing cost) ## Test Plan - Benchmark grouped vs non-grouped kernels on MI300X (589 shapes, BF16) - Verify correctness with existing conv bwd data tests ## Test Result | Metric | Before | After | |--------|--------|-------| | Mean ratio (grouped/nongrouped) | 1.159 | **1.028** | | Median ratio | 1.142 | **1.026** | | Cases within 2% | 26 (4.4%) | **186 (31.8%)** | | Cases >20% slower | 188 (32%) | **2 (0.3%)** | NoShuffle + nongrouped-match instances achieve **~2.8% average gap** with non-grouped kernels (down from ~16%). ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: root <root@ctr-cx64-mi300x-4.amd.com> Co-authored-by: root <root@ctr-cx71-mi300x-01.amd.com> Co-authored-by: root <root@ctr-cx63-mi300x-21.amd.com> Co-authored-by: Bartłomiej Kocot <barkocot@amd.com> Co-authored-by: root <root@gt-ccs-aus-h17-18.cs-aus.dcgpu> Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
Illia Silin
|
717f2efef7 |
[rocm-libraries] ROCm/rocm-libraries#6978 (commit e58096d)
[CK] add composable kernel support on gfx1250 (#6978) ## Motivation Add composable kernel support on gfx1250. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Qun Lin <qlin@amd.com> Co-authored-by: jialuo12_amdeng <jia.luo@amd.com> Co-authored-by: Andriy Roshchenko <andriy.roshchenko@amd.com> Co-authored-by: hsivasun_amdeng <haresh.sivasuntharampillai@amd.com> |
||
|
Illia Silin
|
ac18460782 |
[rocm-libraries] ROCm/rocm-libraries#7384 (commit 10e9d70)
[CK] Suppress new staging compiler errors (#7384) ## Motivation This should make new builds with staging compiler pass. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Meekail Zain
|
0937b002d8 |
[rocm-libraries] ROCm/rocm-libraries#6867 (commit 3cb0219)
Added custom FMHA codegen receipt for TransformerEngine (#6867) ## Motivation TE uses AITER to build static MHA libraries, which ultimately rely on CK kernels. We use the `600` receipt which generates more kernels than TE truly needs. This bespoke receipt allows us to minimize the kernel count, compile time, and memory footprint of our MHA library. ## Technical Details Extended the receipt mechanism to include a custom `700` receipt for TE's needs ## Test Plan Test by building TE using the same receipt profile ## Test Result Build validated in TE using a custom feature branches of AITER/CK to temporarily apply the patch ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com> |
||
|
Yi DING
|
af7118e342 |
[rocm-libraries] ROCm/rocm-libraries#7331 (commit 5692db0)
[CK_TILE] Add async workspace prepare to FMHA BWD launcher (#7331) ## Motivation `aiter::mha_bwd` in group mode currently issues two synchronous `hipMemcpy` D2H copies to read `seqstart_q/k` for launcher construction. These sync copies block the host (~10–30 µs each) and implicitly synchronize the device by draining the stream, breaking CPU/GPU overlap on hot training paths. This PR adds a fully stream-async workspace preparation path on the FMHA BWD launcher so callers can pre-allocate the device workspace from upper-bound shapes and stage seqstart-dependent metadata via D2H/host-pack/H2D entirely on the user's stream. ## Technical Details - `FmhaBwdWorkspaceManager::GetWorkspaceDeviceSizeUpperBound` (`include/ck_tile/ops/fmha/kernel/fmha_bwd_kernel.hpp`): computes the worst-case device dq_acc size from `(max_batch, hdim_q, nhead_q, max_seqlen_q, max_seqlen_k)` without dereferencing any seqstart array. Mirrors `PrepareWorkspaceHost`'s return value with worst-case bounds. - `fmha_bwd_launcher::prepare_workspace_async` (`example/ck_tile/01_fmha/fmha_bwd.hpp`): on the caller's stream, in order: 1. `hipMemsetAsync` of the dq_acc region (when `NeedsZeroDqAcc()`) 2. group mode: `hipMemcpyAsync` D2H of `seqstart_q/k` into a pinned host staging buffer 3. `hipLaunchHostFunc` runs `PrepareWorkspaceHost` on the pinned buffer 4. `hipMemcpyAsync` H2D of the packed metadata into `device_ws_ptr` The pinned staging buffer is held via `std::shared_ptr<void>` returned by a caller-provided `pinned_host_alloc` callback. Lifetime is extended past stream completion by a tail `hipLaunchHostFunc` scheduled in the launcher's destructor. - `ck_tile::pinned_host_releaser` (`include/ck_tile/host/pinned_host_releaser.hpp`): worker-thread utility for callers using bare `hipHostMalloc`. Defers `hipHostFree` off the HIP driver callback thread, which holds runtime locks and would deadlock against concurrent main-thread `hipFree`. PyTorch's `CachingHostAllocator` does not need this. - Example runner (`example/ck_tile/01_fmha/fmha_bwd_runner.hpp`): switched to the async path. ## Test Plan - `tile_example_fmha_bwd` (gfx950, dev preset `-Werror -Weverything`): - batch + nondet / batch + det / group + nondet / group + det - group + det 4-batch varlen (`-b=4 -h=8 -s=4096,3072,2048,1024 -d=128`) - FA (`flash-attention`) integration on ROCm 7.1.1 + PyTorch 2.9.1: - `tests/test_flash_attn_ck.py::test_flash_attn_varlen_deterministic` - `tests/test_flash_attn_ck.py::test_flash_attn_bwd_varlen_seqq_zero` ## Test Result - All CK runner cases `valid:y`. - FA pytest: **1952 passed in 44.82s**. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Copilot
|
c39bff93d0 |
[rocm-libraries] ROCm/rocm-libraries#6983 (commit f4e9a84)
Remove batch_prefill from FMHA_FWD_KNOWN_APIS (#6983) Remove `batch_prefill` from the `FMHA_FWD_KNOWN_APIS` list in `projects/composablekernel/example/ck_tile/01_fmha/CMakeLists.txt`. **Change:** ```cmake # Before set(FMHA_FWD_KNOWN_APIS "fwd;fwd_splitkv;fwd_appendkv;pagedkv_prefill;batch_prefill") # After set(FMHA_FWD_KNOWN_APIS "fwd;fwd_splitkv;fwd_appendkv;pagedkv_prefill") ``` Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: asleepzzz <4926646+asleepzzz@users.noreply.github.com> Co-authored-by: asleepzzz <hanwen.chang@amd.com> Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com> |
||
|
Linjun-AMD
|
5003f7ef8a |
[rocm-libraries] ROCm/rocm-libraries#7272 (commit d02f3c0)
[ck_tile][fmha_bwd] Fix sink_host OOB in group mode reference runner (#7272) ## Summary In `fmha_bwd_runner.hpp`, the `sink_host` `HostTensor` is allocated with first dimension `shape_batch` (= 1 in group mode), but the reference forward loop accesses `sink_host(wb, i_h)` with `wb ∈ [0, batch-1]`. For any `wb >= 1` this is an out-of-bounds heap read, silently corrupting the reference forward math chain (`lse_host`, `o_host`) and turning the bwd-side `d_sink_head_acc` reference into non-deterministic garbage. `HostTensor::operator()` does not bounds check, so the OOB is not caught at runtime. This manifests as intermittent `tile_example_fmha_bwd` failures (25–67% fail rate) when `-sink_grad=1` is combined with `-mode=1` (group mode), with bit-exact but spurious `max_err` values like 4.27 / 14.6. ## Fix One-line: allocate `sink_host` with `batch` (the real per-batch dim) instead of `shape_batch`, mirroring how `sink_host` is accessed by the loop. ```diff - sink_grad ? std::array<ck_tile::index_t, 2>{shape_batch, nhead} + sink_grad ? std::array<ck_tile::index_t, 2>{batch, nhead} Repro tile_example_fmha_bwd -b=2 -h=2 -s=516 -s_k=253 -prec=bf16 -d=72 \ -bias=n -dbias=0 -p_drop=0 -iperm=1 -operm=1 -deterministic=0 \ -v=3 -mode=1 -kname=1 -sink_grad=1 Verification - 0/30 fail on the repro config after fix - Baselines (before fix): - sink=1, mask=n: 25% fail rate (p ≈ 1.8e-4) - sink=1, mask=t: 67% fail rate (p ≈ 6e-15) Attribution Shape bug introduced together with sink_grad in #5504. Unrelated to #6914 (which is a fwd-only fix on a different code path) ``` ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Signed-off-by: junlin12 <junlin12@amd.com> Co-authored-by: Max Podkorytov <4273004+tenpercent@users.noreply.github.com> |
||
|
Illia Silin
|
22b9feb40f |
[rocm-libraries] ROCm/rocm-libraries#7111 (commit 651947f)
[CK] Fix latest batch of staging compiler warnings (#7111) ## Motivation Suppress the new batch of clang lifetimebound and invalidation warnings with the latest staging compiler. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Chao
|
3d8c21e838 |
[rocm-libraries] ROCm/rocm-libraries#6529 (commit 93a6097)
[CK_TILE] Enable V3 persistent kernel dispatch for FMHA forward on gfx950 (#6529) [CK_TILE] Enable V3 persistent kernel dispatch for FMHA forward on gfx950 ## Motivation Enable the existing V3 persistent kernel path for CK-Tile FMHA forward on gfx950 (MI350X/MI355X). The V3 kernel and codegen infrastructure already exist but are disabled via hardcoded `F_is_v3_enabled=False`. This change replaces the compile-time gate with a runtime environment variable `CK_FMHA_ENABLE_V3=1` (disabled by default, opt-in). When enabled: - **Prefill** workloads (seqlen_q > 1) dispatch to V3 persistent pipeline - **Decode** workloads (seqlen_q == 1) always use V2 (memory-bound, better suited) The V3 persistent kernel uses grid-stride scheduling, XCD-interleave tile assignment for L2 locality, LPT reversal for causal masks, and gfx950 async buffer loads. ## Technical Details Single file: `example/ck_tile/01_fmha/codegen/ops/fmha_fwd.py` - Add `#include <cstdlib>` and `<string>` for `std::getenv` - Replace `{F_is_v3_enabled}` template parameter with runtime env var check - Add `seqlen_q > 1` guard (decode always uses V2) - Remove `.format()` call in `write_fwd_api()` ## Dependencies Depends on https://github.com/ROCm/rocm-libraries/pull/6501 — builds on XCD-interleave and LPT scheduling infrastructure. ## Test Plan - GPU validation on MI300X (gfx942, ROCm 6.4.1): - Command: `./build/bin/tile_example_fmha_fwd -b=2 -h=8 -s=4096 -d=128 -prec=bf16 -v=1 -warmup=1 -repeat=3` - GPU validation on MI350X (gfx950, ROCm 7.0): - Command (V2): `./build/bin/tile_example_fmha_fwd -b=2 -h=8 -s=4096 -d=128 -prec=bf16 -v=1 -warmup=1 -repeat=3` - Command (V3): `CK_FMHA_ENABLE_V3=1 ./build/bin/tile_example_fmha_fwd -b=2 -h=8 -s=4096 -d=128 -prec=bf16 -v=1 -warmup=1 -repeat=3` - Command (decode, always V2): `./build/bin/tile_example_fmha_fwd -b=64 -h=32 -h_k=8 -s=1 -s_k=4096 -d=128 -prec=bf16 -mode=group -v=1 -warmup=1 -repeat=3` ## Test Result Benchmark results (MI350X, gfx950, ROCm 7.0): | Config | V2 (TFlops) | V3 (TFlops) | Speedup | |--------|-------------|-------------|---------| | Non-causal b=2 h=8 hk=2 s=4096 d=128 bf16 | 696.3 | 884.2 | **+27.0%** | | Causal b=2 h=8 hk=2 s=4096 d=128 bf16 | 371.3 | 494.9 | **+33.3%** | | GQA b=2 h=32 hk=8 s=2048 d=128 bf16 | 671.3 | 831.7 | **+23.9%** | | LLaMA-70B b=1 h=64 hk=8 s=4096 d=128 bf16 | 761.5 | 927.3 | **+21.8%** | | Causal GQA b=2 h=32 hk=8 s=2048 d=128 bf16 | 345.4 | 631.9 | **+82.9%** | | Long-seq b=1 h=16 s=16384 d=128 bf16 | 797.8 | 969.9 | **+21.6%** | | Decode b=64 h=32 hk=8 s=1 s_k=4096 bf16 | 1828 GB/s | — (V2 path) | unaffected | Benchmark results (MI300X, gfx942, ROCm 6.4.1): V3 has 0% effect on MI300X — V3 relies on gfx950 async buffer loads and falls back to the V2 code path on gfx942. No regression on any config. | Config | TFlops / GB/s | Time (ms) | Delta vs baseline | |--------|-------------|-----------|-------------------| | MHA bf16 b=2 h=8 s=4096 d=128 | 342.98 TFlops | 0.401 | +0.1% | | MHA fp16 b=2 h=8 s=4096 d=128 | 411.18 TFlops | 0.334 | +4.9% | | Causal MHA bf16 b=2 h=8 s=4096 d=128 | 232.61 TFlops | 0.296 | +2.4% | | GQA 4:1 bf16 b=2 h=32 hk=8 s=2048 d=128 | 320.07 TFlops | 0.429 | -1.4% | | GQA 8:1 bf16 b=2 h=64 hk=8 s=2048 d=128 | 353.91 TFlops | 0.777 | +1.7% | | LLaMA-70B prefill b=1 h=64 hk=8 s=4096 d=128 bf16 | 381.53 TFlops | 1.441 | +1.2% | | Long-seq bf16 b=1 h=16 s=16384 d=128 | 388.61 TFlops | 5.659 | +1.4% | | Decode b=64 h=32 hk=8 s_k=4096 d=128 bf16 | 693.40 GB/s | 1.550 | +0.3% | All validation tests pass (`valid:y`) on both MI300X and MI350X. Additional validation: - `CK_FMHA_ENABLE_V3=0` correctly falls back to V2 (default behavior unchanged) - `CK_FMHA_ENABLE_V3=1` dispatches to V3 for prefill, V2 for decode - Validation passes across fp16/bf16, batch/group mode, causal/non-causal - No regression on decode path --------- Co-authored-by: Chao Zhou <chaozhou@fb.com> Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com> |
||
|
Linjun-AMD
|
5a2a362c46 |
[rocm-libraries] ROCm/rocm-libraries#6914 (commit b791478)
[CK_TILE][FMHA] Fix sink un-mask under right-window and emit fp8bf16 batch_prefill sink kernels (#6914) ## Summary Two related fixes to `ck_tile` FMHA so that StreamLLM-sink + sliding-window batch-prefill works correctly for fp8 KV / bf16 compute. Review the commits in this order: 1. `fmha: emit sink kernels for fp8bf16 batch_prefill` Extends `example/ck_tile/01_fmha/codegen/ops/fmha_batch_prefill.py` so the fp8(KV) / bf16(QO) batch-prefill codegen also emits the `mask=mask_enum::generic_with_sink` variant. Without this the runtime could not dispatch to a sink-aware kernel for the fp8bf16 path. 2. `fmha: respect right-window in IsOutOfSinkBound` The sink un-mask in `GenericAttentionMask::IsOutOfSinkBound` (local-mask branch) used `(i_y + x) > 1` as the gate, which conditioned on the row index instead of the column index. As a result, queries `1..sink-1` could attend to *future* sink positions (violating causal / right-window), while query `0` fell back to the plain causal mask. The fix replaces the guard with `i_x < i_y + x` so every query only sees sink columns up to its own right-window boundary. 3. `fmha: clarify IsOutOfSinkBound predicate comment` Doc-only follow-up that rewrites the comment above the predicate as a clause-by-clause explanation (`i_x < sink`, `i_x < i_y + x`, `y < y_total`, `i_y < x_total`). ## Test plan - [x] Repro on aiter `op_tests/test_batch_prefill.py` (fp8 + bf16_dequant modes with `sink=4`, `win_left=1023`, `softcap=0.0`, `sal=True`) now passes for all parametrized shapes. - [x] Existing fp16/bf16 batch-prefill paths (no sink) unchanged — codegen diff only adds the `generic_with_sink` variant for fp8bf16; existing kernel object lists unaffected. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: fengjunda.aml <fengjunda.aml@bytedance.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: root <root@smci350-rck-g03-f12-31.rck.dcgpu> |
||
|
Yi DING
|
2c677e8471 |
[rocm-libraries] ROCm/rocm-libraries#6152 (commit 36b016a)
[CK_TILE] Use Unified Workspace for FMHA BWD (#6152) ## Motivation `dq_acc` is the intermediate accumulation buffer used in FMHA backward pass for deterministic mode. The current implementation allocates it as a **single rectangular tensor**: ``` shape = [shape_batch, nhead, nsplits, shape_seqlen_q, hdim_q] ``` where `nsplits = launcher.dq_acc_splits` (a single scalar), computed from `max_seqlen_k` and shared across all batches. ### Problems 1. **Memory waste**: In group mode, each batch may have a different `seqlen_k`, but `nsplits` is computed from `max_seqlen_k`, causing batches with shorter `seqlen_k` to over-allocate in the split dimension. 2. **Interface coupling**: `fmha_bwd_args` exposes internal layout details such as `stride_dq_acc`, `nhead_stride_dq_acc`, `batch_stride_dq_acc`, and `split_stride_dq_acc`. The caller is responsible for computing these strides, but this logic belongs inside the kernel. ### Goals 1. Switch `dq_acc` buffer to a **compact layout**: batches are concatenated contiguously, with each batch occupying `nhead * nsplits_i * seqq_i * hdim_q` elements (nhead outermost). 2. **Remove all `*_stride_dq_acc` fields** from `fmha_bwd_args`, replacing them with a single `workspace_ptr`; the kernel splits this internally using a fixed layout. 4. `fmha_bwd_launcher` provides a **workspace management interface**: the caller only needs to allocate GPU memory and call `prepare_workspace()` — no layout computation required. 5. **Isolate kernel internals from the caller API**: the `dq_acc` layout (nsplits, strides, buffer size) is determined entirely inside the launcher/kernel. Future changes to block shape, pipeline type, or persistent kernel strategy require no modifications to the caller's `fmha_bwd_args` or workspace allocation logic. ## Technical Details ### Interface Design #### New fields in `fmha_bwd_traits` ```cpp struct fmha_bwd_traits { int seqlen_q; int seqlen_k; int batch; int max_seqlen_q; int max_seqlen_k; int hdim_q; int hdim_v; int nhead_q; int nhead_k; std::string data_type; bool is_group_mode; mask_enum mask_type; bias_enum bias_type; bool has_dbias; bool has_dropout; bool is_store_randval; bool is_deterministic; // New: cumulative physical seqlen pointers for group mode (pass nullptr for batch mode). // seqstart_qs[i+1] - seqstart_qs[i] = physical seqlen_q of batch i (including padding); length = batch+1 // seqstart_ks[i+1] - seqstart_ks[i] = physical seqlen_k of batch i (including padding); length = batch+1 const int* seqstart_qs = nullptr; const int* seqstart_ks = nullptr; }; ``` #### `fmha_bwd_launcher` actual structure ```cpp struct fmha_bwd_launcher { std::function<float(fmha_bwd_args, const ck_tile::stream_config&)> run{}; // Total workspace size in bytes (host_ws_size + device_ws_size), computed by init(). // Zero for kUseQrQtrDorPipeline (writes dq directly, no acc buffer needed). size_t workspace_size = 0; fmha_bwd_launcher(const fmha_bwd_traits&); // Copies auxiliary data (nsplits[], offsets[]) via hipMemcpy to the head of the GPU workspace, // and zeros the dq_acc buffer portion (tail of workspace) if required. // The memory pointed to by device_ws must be >= workspace_size bytes. std::function<void(void* device_ws)> prepare_workspace{}; template <typename... Args> float operator()(Args&&... args) const { return run(std::forward<Args>(args)...); } private: size_t host_ws_size = 0; // CPU workspace size (nsplits[] + offsets[] arrays) size_t device_ws_size = 0; // GPU-only data size (dq_acc buffer) std::unique_ptr<char[]> ws_host; // host-side workspace buffer public: template <typename T0, typename T1, typename T2, typename Arch> void init(const fmha_bwd_traits& traits); }; ``` The `init<>()` template method (invoked by codegen dispatch branches as `this->init<...>(t)`) is responsible for: 1. Setting the `run` lambda 2. Calling `FmhaBwdDQDKDVKernel::GetWorkspaceHostSize(batch)` to obtain `host_ws_size` 3. Allocating `ws_host` (host memory) 4. Calling `FmhaBwdDQDKDVKernel::PrepareWorkspaceHost(ws_host.get(), ...)` to fill nsplits/offsets; return value is `device_ws_size` 5. `workspace_size = host_ws_size + device_ws_size` 6. Setting the `prepare_workspace` lambda (captures `this`, calls `PrepareWorkspaceDevice`) When no kernel matches the given traits, both `run` and `prepare_workspace` are initialized to default lambdas that print a warning to `std::cerr` and return gracefully (no exception). #### Workspace overall layout The workspace is managed by `FmhaBwdWorkspaceManager` and consists of two segments: ``` Offset 0 (CPU-prepared segment, host_ws_size bytes; also hipMemcpy'd to the head of GPU workspace): index_t nsplits[batch or 1] — per-batch nsplits array group mode: batch elements batch mode / non-deterministic: 1 element [group mode only] long_index_t dq_acc_offsets[batch+1] — per-batch element offset (inclusive prefix sum) offsets[0]=0, offsets[i+1] = offsets[i] + nhead*nsplits_i*seqq_i*hdim_q Offset host_ws_size (device data segment, device_ws_size bytes): AccDataType dq_acc[total_elements] — compact dq_acc buffer (zeroed if required) total_elements = sum_i(nhead * nsplits_i * seqq_i * hdim_q) layout within each batch: [nhead, nsplits_i, seqq_i, hdim_q] note: seqq_i uses the physical length (including padding) ``` Alignment constant (`ALIGNMENT = 16`): ``` nsplits_size = align_up(sizeof(index_t) * N, 16) // N = batch (group) or 1 (batch/non-det) offsets_size = align_up(sizeof(long_index_t) * (batch+1), 16) // group mode only host_ws_size = nsplits_size + offsets_size dq_acc_offset = host_ws_size // GetDqAccDataOffset(batch) ``` **Key benefits**: - The kernel reads nsplits/offsets directly from the workspace head — no device-side recomputation. - `FmhaBwdConvertQGradKernel` is completely decoupled from the pipeline block shape (`kN0`): nsplits is read from `nsplits_ptr`, `kN0` is no longer a template parameter, and multiple dq_dk_dv tiles with different `F_bn0` values now share a single convert_dq kernel instance (under receipt 1/2, deterministic convert_dq kernel count drops from ~300 to 60). - nsplits/offsets are computed on the host and transferred in one `hipMemcpy`; the dq_acc buffer follows immediately, at the offset given by `GetDqAccDataOffset`. #### Workspace size by scenario | Scenario | `workspace_size` | Notes | |----------|-----------------|-------| | **kUseQrQtrDorPipeline** (any mode) | `0` | Writes dq directly; no acc buffer; `PrepareWorkspaceHost` returns 0 | | **Non-deterministic + batch mode** | `> 0` | nsplits[1]=1; dq_acc used for atomic add; `workspace_size = host_ws_size + batch*nhead*seqlen_q*hdim_q*ebytes` | | **Non-deterministic + group mode** | `> 0` | nsplits[1]=1; dq_acc contiguous layout; `workspace_size = host_ws_size + nhead*seqstart_qs[batch]*hdim_q*ebytes` | | **Deterministic + group mode** | `> 0` | nsplits[batch], offsets[batch+1], compact dq_acc; nsplits_i computed independently per batch | | **Deterministic + batch mode persistent** | `> 0` | nsplits[1] (uniform across batches); dq_acc `batch*nhead*nsplits*seqlen_q*hdim_q` | **NeedsZeroDqAcc** (determines whether `PrepareWorkspaceDevice` calls `hipMemset`): - Persistent kernel (deterministic batch mode) or non-deterministic: **must zero** (atomic add requires zero initialization) - Deterministic group mode + no mask: **no zeroing needed** (every tile writes its full region) - Deterministic + with mask: **must zero** (some blocks are skipped, leaving uninitialized tiles that would contribute to the reduction) #### Caller usage ```cpp // 1. Create launcher (traits include seqstart_qs/ks pointers; workspace_size is computed during construction) fmha_bwd_launcher launcher(fmha_traits); // 2. Read launcher.workspace_size directly const auto ws_size = launcher.workspace_size; // 3. Allocate a single GPU workspace ck_tile::DeviceMem ws_buf(ws_size); // 4. Copy nsplits/offsets to GPU head and zero dq_acc if required launcher.prepare_workspace(ws_buf.GetDeviceBuffer()); // 5. Build args with a single workspace pointer; the kernel splits it internally fmha_bwd_args args{ ..., ws_size > 0 ? ws_buf.GetDeviceBuffer() : nullptr, // workspace_ptr }; launcher(args, stream_config); ``` --- ### Key Code Structure #### FmhaBwdWorkspaceManager (`fmha_bwd_kernel.hpp`, new class) ```cpp template <typename AccDataType, bool kIsGroupMode, bool kIsDeterministic> struct FmhaBwdWorkspaceManager { static constexpr size_t ALIGNMENT = 16; // CPU workspace (nsplits + offsets) sizes static size_t GetDqAccSplitsSize(int batch); // align_up(sizeof(index_t)*N, 16) static size_t GetDqAccOffsetsSize(int batch); // group mode only: align_up(sizeof(long_index_t)*(batch+1), 16) static size_t GetWorkspaceHostSize(int batch); // = SplitsSize + OffsetsSize // Starting offset of dq_acc data within the full workspace (= host_ws_size) static size_t GetDqAccDataOffset(int batch); // = GetWorkspaceHostSize(batch) // Fills nsplits/offsets in the CPU workspace; returns device_ws_size (dq_acc buffer bytes) template <bool kUseQrQtrDorPipeline, index_t kN0> static size_t PrepareWorkspaceHost(void* cpu_ws, index_t batch_size, index_t hdim_q, index_t nhead_q, index_t seqlen_q, index_t seqlen_k, const index_t* seqstart_qs, const index_t* seqstart_ks); // hipMemcpy's cpu_ws to device_ws head; hipMemset's the dq_acc portion to 0 if required template <bool kUseQrQtrDorPipeline, bool kHasMask> static void PrepareWorkspaceDevice(void* device_ws, const void* host_ws, size_t device_ws_size, size_t host_ws_size); }; ``` #### workspace_ptr parsing (inside the kernel) The kernel parses three address regions from `kargs.workspace_ptr`: **Group mode (`FmhaBwdDQDKDVKernel::MakeKargs`)**: ```cpp const uint8_t* ws = reinterpret_cast<uint8_t*>(workspace_ptr); // dq_acc_ptr (stored in FmhaBwdCommonKargs) ws + WorkspaceManager::GetDqAccDataOffset(batch) // dq_acc_batch_offset_ptr (FmhaBwdGroupModeKargs field) reinterpret_cast<const long_index_t*>(ws + WorkspaceManager::GetDqAccOffsetsOffset(batch)) ``` **Batch mode**: ```cpp ws + WorkspaceManager::GetDqAccDataOffset(batch) // dq_acc_ptr // No offsets pointer; batch offset is computed inside run_() from nsplits ``` **`FmhaBwdConvertQGradKernel`** follows the same pattern: - Group mode: extracts `dq_acc_ptr`, `dq_acc_batch_offset_ptr`, and `nsplits_ptr` (`GetDqAccSplitsOffset(batch)`) from workspace - Batch mode: reads nsplits from `nsplits_ptr[0]`; batch offset computed internally ### Addressing in `run_()` (group mode) ```cpp // Per-batch processing: const long_index_t batch_offset_dq_acc = kargs.dq_acc_batch_offset_ptr[i_batch]; // seqq_i (physical length) derived from seqstart_q_ptr const index_t seqq_i = kargs.seqstart_q_ptr[i_batch+1] - kargs.seqstart_q_ptr[i_batch]; // nsplits_i read from nsplits_ptr (convert_dq kernel) or from GetDqAccSplits const long_index_t split_stride_i = static_cast<long_index_t>(seqq_i) * kargs.hdim_q; const long_index_t nhead_stride_i = static_cast<long_index_t>(nsplits_i) * split_stride_i; // Final address: dq_acc_base + batch_offset_dq_acc + i_nhead * nhead_stride_i + i_split * split_stride_i ``` #### nsplits computation (`PrepareWorkspaceHost`) `PrepareWorkspaceHost` is a template method of `FmhaBwdWorkspaceManager` that still takes `kN0` as a template parameter (from `BlockFmhaShape::kN0` of the dq_dk_dv pipeline). However, this parameter is **only used inside this host-side function** to compute nsplits — it is no longer passed into the convert_dq kernel. | Mode | nsplits computation | |------|---------------------| | kUseQrQtrDorPipeline | Writes dq directly; nsplits[0]=0; returns device_ws_size=0 | | Non-deterministic | nsplits[0]=1; dq_acc used for atomic add | | Deterministic + group mode | `ceil((seqstart_ks[i+1]-seqstart_ks[i]) / kN0)` computed per batch | | Deterministic + batch mode persistent | Same logic as the original `GetDqAccSplits` (`dqdqkdv_workers` based) | ### Removing kN0 dependency from `FmhaBwdConvertQGradKernel` `FmhaBwdConvertQGradKernel` previously required `kN0` as a template parameter (via `BlockFmhaBwdConvertQGradPipelineProblem`) for two purposes: 1. In batch mode `operator()`: self-computing `nsplits = ceil(seqlen_k / kN0)` 2. The `b{kM0}x{kN0}` component of the kernel name string Both have been removed in this refactor: - **Batch mode**: now reads `kargs.nsplits_ptr[0]` directly (guarded by `if constexpr(kIsDeterministic)` to avoid accessing a non-existent field in non-deterministic instances) - **Kernel name**: simplified to `b{kM0}`, no longer includes `kN0` - **Template parameters**: `BlockFmhaBwdConvertQGradPipelineProblem` drops the `kN0_` parameter; `fmha_bwd_convert_dq_traits_` drops the `kN0` parameter; `F_bn0`/`convert_dq_bn0` fields removed from codegen Effect: all dq_dk_dv tiles sharing the same `(hdim, dtype, mode, pad, deterministic)` combination — regardless of `F_bn0` value (16/64/128/192/256) — now share a **single** convert_dq kernel instance. --- ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
ltqin
|
501e7ef12a |
[rocm-libraries] ROCm/rocm-libraries#6574 (commit b3db057)
[CK_TILE] Add SageAttention v2 forward kernel with multi-granularity quantization (#6574) ## Summary Add a CK_TILE forward kernel implementing [SageAttention v2](https://arxiv.org/abs/2411.10958) — an attention algorithm that applies multi-granularity quantization to Q/K/V before computing attention, trading minimal accuracy loss for higher throughput on low-precision hardware. ### Quantization design | Tensor | Supported data types | Scale granularity options | |--------|---------------------|--------------------------| | Q | fp8 / int8 / int4 | per-tensor, per-block (128 tokens), per-warp (32 tokens), per-thread (4 tokens) | | K | fp8 / int8 / int4 | per-tensor, per-block (128 tokens), per-warp (64 tokens), per-thread (16 tokens) | | V | fp8 | per-channel (always) | | O | bf16 | — | Three precision combinations are supported: `fp8/bf16` (QKV fp8, O bf16), `i8/fp8/bf16` (QK int8, V fp8, O bf16), and `i4/fp8/bf16` (QK int4, V fp8, O bf16). ### Architecture support - **gfx9** (CDNA2/3, e.g. gfx90a, gfx942) — full tile set - **gfx950** (CDNA4) — restricted tile set (N-per-block capped at 64 for fp8-family dtypes) ### Implementation - Two pipeline variants: `QRKSVS` (synchronous) and `QRKSVS_ASYNC` (async copy) - Masking support: no mask, causal (top-left / bottom-right), and generic windowed - Batch and group (variable-length) modes - Head dimension: d=128, d_v=128 - Python codegen under `example/ck_tile/49_sageattention/codegen/` generates kernel instances per target/dtype/tile combination - Smoke tests included via `tile_example_sageattn_fwd` ### Test commands \`\`\`bash # fp8 QKV ./build/bin/tile_example_sageattn_fwd -v=1 -b=16 -h=8 -s=1024 -d=128 -kname=1 -prec=fp8bf16 -qscale=3 -init=3 # int8 QK, fp8 V ./build/bin/tile_example_sageattn_fwd -v=1 -b=16 -h=8 -s=1024 -d=128 -kname=1 -prec=i8fp8bf16 -qscale=3 -init=3 \`\`\` \`-qscale\` values: 1=per-tensor, 2=per-block, 3=per-warp, 4=per-thread |
||
|
ArthurLiu
|
67aa854621 |
[rocm-libraries] ROCm/rocm-libraries#6764 (commit 8c20d70)
[CK][CK_TILE] Fix FMHA codegen group mode dispatch (#6764) ## Motivation FMHA codegen had incorrect dispatch behavior in group mode. Two root causes: 1. Wrong field names in dispatch conditions — Used batch-mode fields (seqlen_q, seqlen_k) instead of group-mode fields (max_seqlen_q, max_seqlen_k), causing wrong kernel selection at runtime on gfx950. 2. Missing kernel variants — Group mode was overly filtered out from smaller-tile specializations (bwd) and lacked spatial-padding pipeline variants on gfx950 (fwd). gfx942 don't support trload pipeline. ## Technical Details fmha_bwd.py: - max_seq_q_cond and extra_cond now emit t.max_seqlen_q / t.max_seqlen_k for group mode. - Relaxed kernel filtering: group mode no longer skips tiles with max_seq_q != 0. fmha_fwd.py: - get_bm0_cond emits a.max_seqlen_q for group mode tile-size dispatch. - Added two qr_async_trload pipeline variants with spatial padding for gfx950 group mode. ## Test Plan Triggering AITER CI job: ## Submission Checklist - [ x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Wojciech Laskowski
|
640bd560ec |
[rocm-libraries] ROCm/rocm-libraries#5801 (commit 27f6d15)
[CK Tile] Adding WMMA wrappers for dense builtins (#5801) ## Motivation This PR is part of the [WMMA/MFMA] unification work. It's the first of the series of PRs that add all the necessary MMA builtins as a `amdgcn_mma` structs. ## Technical Details This change adds new specializations for WMMA dense builtins. In total, we have now 9 RDNA4 builtins and 3 RDNA3 builtins. ## Test Plan All the new wrappers were added to the test suite in `test_amdgcn_mma_layout.inc`. ## Test Result Test pass locally, waiting for the CI. ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. --------- Co-authored-by: Yung-sheng Tu <yung-sheng@streamhpc.com> |