fix amd tir&add examples #784
Merged
+665
−5
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…nd clarity (tile-ai#668) - Enhanced buffer index handling to address precision issues by removing redundant operations. - Streamlined the logic for determining buffer overlaps, ensuring more accurate conflict detection. - Updated related documentation to reflect changes in buffer management practices.
…ed flexibility - Introduced a new input.txt file for configurable parameters. - Modified the example_amd_flash_attn_fwd.py script to allow for a wider range of configurations, including additional options for num_stages, enable_rasterization, and k_pack. - Streamlined the main function for better clarity and organization. - Added a new test script to facilitate running the example with specified parameters.
… example with swizzle layout annotations - Deleted input.txt and test.sh files as they are no longer needed. - Updated example_amd_flash_attn_fwd.py to include swizzle layout annotations for shared memory, improving bank conflict avoidance. - Reintroduced swizzle usage in the kernel for better performance.
- Updated function names for clarity, changing `get_v2_configs` to `get_configs` and `fast_flashattn_v2` to `fast_flashattn`. - Streamlined the main function by renaming `main_v2` to `main` and adjusting the corresponding calls. - Removed outdated comments and improved code organization for better readability.
- Improved code readability by adjusting line breaks and indentation in the `fast_flashattn` function. - Streamlined the `main` function parameter formatting for consistency. - Removed unnecessary blank lines to enhance overall code organization.
- Improved the `example_amd_flash_attn_fwd.py` script for better clarity and organization. - Added new CI workflows for AMD and documentation publishing. - Updated various requirements files to include necessary dependencies. - Introduced new test cases and examples for better coverage and functionality. - Refactored existing code for improved readability and maintainability.
- Introduced `example_amd_flash_attn_bwd.py` for backward attention computation using TileLang. - Added `test.sh` script to facilitate running the new example with specified parameters. - Enhanced the overall structure and organization of the example for better clarity and usability.
- Reduced the number of threads and `num_split_q` options for improved performance. - Adjusted `panel_size` options to streamline configuration settings.
- Introduced a new example script `example_amd_flash_attn_bwd.py` demonstrating the forward and backward operations of Flash Attention using TileLang. - Implemented JIT-compiled functions for both forward and backward passes, including preprocessing and postprocessing steps. - Added a main function to facilitate testing and benchmarking of the attention mechanism with configurable parameters. - Included reference implementation for validation against PyTorch's attention mechanism. This addition enhances the examples directory by providing a comprehensive guide for users to understand and utilize Flash Attention in their applications.
- Updated `example_amd_flash_attn_bwd.py` to include more comprehensive testing features for the Flash Attention implementation. - Improved the main function to allow for better parameter configuration and benchmarking. - Added validation checks against PyTorch's attention mechanism to ensure accuracy and reliability of the example. This update aims to provide users with a more robust tool for understanding and utilizing Flash Attention in their applications.
- Updated file name from `intrin_rule_hip.cc` to `intrin_rule_cuda.cc` to reflect the change in focus from HIP to CUDA intrinsic rules. - Adjusted include paths for better organization and clarity in the code structure.
|
Note Other AI code review bot(s) detectedCodeRabbit has detected other AI code review bot(s) in this pull request and will avoid duplicating their findings in the review comments. This may lead to a less comprehensive review. WalkthroughAdds an AMD FlashAttention backward example with fused forward/backward kernels and a PyTorch autograd bridge; narrows the forward autotuner search space; adds test invocations for AMD examples; and implements HIP intrinsic lowering (math/fast-math, popcount, warp shuffles, activemask, fmod) with new HIP builtin registrations. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User
participant PyTorch as PyTorch (autograd)
participant Tile as tilelang kernels
participant Ref as Reference
User->>PyTorch: attention(q,k,v,causal,groups)
PyTorch->>Tile: compile flashattn_fwd(...) -> flash_fwd
PyTorch->>Tile: flash_fwd(Q,K,V) -> O, lse
PyTorch-->>User: O (saves Q,K,V,lse)
User->>PyTorch: backward(dO)
PyTorch->>Tile: flashattn_bwd_preprocess(...) -> Delta
PyTorch->>Tile: compile flashattn_bwd(...) -> flash_bwd
PyTorch->>Tile: flash_bwd(Q,K,V,dO,lse,Delta) -> dQ,dK,dV
PyTorch->>Tile: flashattn_bwd_postprocess(dQ) -> dQ_out
PyTorch-->>User: grads(dQ_out,dK,dV)
Note over PyTorch,Ref: Optional validation path via ref_program(Q,K,V)
sequenceDiagram
autonumber
participant TIR as TVM TIR
participant HIPRule as HIP lowering
participant HIPBuiltin as HIP builtins
TIR->>HIPRule: lower(intrinsic)
alt math / fast-math / tan / fmod / popcount
HIPRule->>HIPBuiltin: map to HIP symbol (e.g., expf, __expf, tan, __popc, fmod)
HIPBuiltin-->>HIPRule: call expr
else warp shuffle / activemask
HIPRule->>HIPBuiltin: __activemask()
HIPRule->>HIPBuiltin: __shfl_*_sync(mask,var,arg,width)
HIPBuiltin-->>HIPRule: call expr
end
HIPRule-->>TIR: replaced HIP-specific call
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60–90 minutes Possibly related PRs
Poem
✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
CodeRabbit Configuration File (
|
|
👋 Hi! Thank you for contributing to the TileLang project. Please remember to run We appreciate you taking this step! Our team will review your contribution, and we look forward to your awesome work! 🚀 |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @Alex4210987, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request significantly enhances AMD GPU support within the tilelang framework. It introduces a new set of intrinsic rules for HIP, which are crucial for optimizing low-level operations on AMD hardware. Additionally, it expands the example suite by adding a complete Flash Attention backward pass implementation and refines the autotuner configurations for the forward pass, aiming for improved performance and easier benchmarking.
Highlights
- AMD HIP Intrinsic Rules: Introduced a new C++ file (
src/target/intrin_rule_hip.cc) that defines intrinsic rules for HIP (Heterogeneous-compute Interface for Portability). This enables better compilation and optimization of mathematical functions and warp-level operations (like shuffle instructions) for AMD GPUs within the TVM framework, directly addressing the 'fix amd tir' aspect of the PR title. - Flash Attention Backward Pass Example: Added a comprehensive example (
examples/amd/example_amd_flash_attn_bwd.py) demonstrating the Flash Attention backward pass implemented usingtilelang. This example integrates with PyTorch's autograd system and includes benchmarking against a reference PyTorch implementation. - Autotuner Configuration Refinement: Modified the existing Flash Attention forward pass example (
examples/amd/example_amd_flash_attn_fwd.py) to refine the autotuner configurations. This change narrows down the search space for optimalthreads,num_split_q, andpanel_sizeparameters, potentially leading to more efficient tuning and improved performance. - New Test Script for AMD Examples: Added a new shell script (
examples/amd/test.sh) to facilitate testing and benchmarking of the AMD Flash Attention forward pass example. This script provides a convenient way to run thetilelangimplementation and compare its performance against a referencecomposable_kernelexecutable.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request introduces a new example for the backward pass of flash attention on AMD GPUs, along with a corresponding test script and new HIP intrinsic rules. It also refines the autotuner configurations for the existing forward pass example. My review focuses on improving code quality, correctness, and portability. Key suggestions include fixing incorrect argument parsing, removing dead code, replacing magic numbers with constants in the new Python example, improving the portability of the test script by removing hardcoded paths, and correcting comments and file formatting in the new C++ code.
| parser.add_argument('--n_ctx', type=int, default=1024, help='Context size') | ||
| parser.add_argument('--d_head_qk', type=int, default=192, help='Head dimension for Q/K') | ||
| parser.add_argument('--d_head_v', type=int, default=128, help='Head dimension for V') | ||
| parser.add_argument('--causal', type=bool, default=False, help='Causal flag') |
There was a problem hiding this comment.
The use of type=bool for command-line arguments is problematic and can lead to unexpected behavior (e.g., --causal False is interpreted as True). The standard way to handle boolean flags is with action='store_true', which is also consistent with other examples in the repository.
Suggested change
| parser.add_argument('--causal', type=bool, default=False, help='Causal flag') | |
| parser.add_argument('--causal', action='store_true', help='Causal flag') |
|
|
||
| @tilelang.jit(out_idx=[3, 4]) | ||
| def flashattn_fwd(batch, heads, seq_len, dim_qk, dim_v, is_causal, block_M, block_N, groups=1): | ||
| scale = (1.0 / dim_qk)**0.5 * 1.44269504 # log2(e) |
There was a problem hiding this comment.
The magic number 1.44269504 (which is log2(e)) is used here and on line 140. It would be better to define it as a constant at the module level for readability and maintainability, for example: LOG2_E = 1.44269504.
Comment on lines
+110
to
+113
| def make_dq_layout(dQ): | ||
| # atomicAdd can not be vectorized, so we need to reorder dq to match the 8x8 gemm fragment | ||
| return T.Layout(dQ.shape, | ||
| lambda b, l, h, d: [b, l // 8, h, d // 8, (d % 2), 4 * (l % 8) + (d % 8) // 2]) |
There was a problem hiding this comment.
The function make_dq_layout is defined but never used in this file. It appears to be dead code and should be removed to improve maintainability.
| @@ -0,0 +1,10 @@ | |||
| /root/miniconda3/envs/py312/bin/python3 examples/amd/example_amd_flash_attn_fwd.py \ | |||
There was a problem hiding this comment.
The hardcoded path /root/miniconda3/envs/py312/bin/python3 makes the script not portable. Please use python3 and assume it's in the user's PATH.
Suggested change
| /root/miniconda3/envs/py312/bin/python3 examples/amd/example_amd_flash_attn_fwd.py \ | |
| python3 examples/amd/example_amd_flash_attn_fwd.py \ |
| --is_causal \ | ||
| --groups 2 | ||
|
|
||
| /root/composable_kernel/build/bin/tile_example_fmha_fwd \ |
There was a problem hiding this comment.
The hardcoded path /root/composable_kernel/build/bin/tile_example_fmha_fwd makes the script not portable. Please consider making this configurable, for example via an environment variable, or assume it's in the PATH.
Suggested change
| /root/composable_kernel/build/bin/tile_example_fmha_fwd \ | |
| tile_example_fmha_fwd \ |
src/target/intrin_rule_hip.cc
Outdated
Comment on lines
2
to
3
| * \file intrin_rule_cuda.cc | ||
| * \brief CUDA intrinsic rules. |
There was a problem hiding this comment.
The file header comment seems to be copied from the CUDA equivalent and was not updated. It should refer to HIP instead of CUDA.
Suggested change
| * \file intrin_rule_cuda.cc | |
| * \brief CUDA intrinsic rules. | |
| * \file intrin_rule_hip.cc | |
| * \brief HIP intrinsic rules. |
src/target/intrin_rule_hip.cc
Outdated
|
|
||
| } // namespace intrin | ||
| } // namespace codegen | ||
| } // namespace tvm No newline at end of file |
There was a problem hiding this comment.
The file is missing a newline at the end. It's a common convention to end files with a newline for better compatibility with various tools.
Suggested change
| } // namespace tvm | |
| } // namespace tvm | |
…installation - Removed the installation of `flash_attn==2.5.8` to streamline the CI process. - Added a step to uninstall `torch`, `torchvision`, and `torchaudio` prior to installing pre-release versions, ensuring compatibility and reducing potential conflicts.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
examples/amd/example_amd_flash_attn_fwd.py (1)
20-21: Fix device mismatch in reference scaling.Dividing a CUDA/HIP tensor by a CPU tensor will error. Use a Python scalar or place the tensor on the same device.
- scores = scores / torch.sqrt(torch.tensor(dim, dtype=scores.dtype)) + # Avoid device mismatch; both of the following are fine: + # 1) Python scalar: + # scores = scores / (dim ** 0.5) + # 2) Tensor on same device: + scores = scores / torch.sqrt(torch.tensor(dim, dtype=scores.dtype, device=scores.device))examples/amd/test.sh (1)
1-11: Make the test script portable and fail-fast.Avoid absolute paths, add a shebang and strict mode, and resolve repo root so it runs in CI and local envs.
+#!/usr/bin/env bash +set -euo pipefail +cd "$(git rev-parse --show-toplevel)" + -/root/miniconda3/envs/py312/bin/python3 examples/amd/example_amd_flash_attn_fwd.py \ +python3 examples/amd/example_amd_flash_attn_fwd.py \ --batch 2 \ --heads 16 \ --seq_len 4096 \ --dim 128 \ --is_causal \ --groups 2 -/root/composable_kernel/build/bin/tile_example_fmha_fwd \ +${CK_BIN:-tile_example_fmha_fwd} \ -b=2 -h=16 -s=4096 -d=128 -mask=t -v=1 -warmup=5 -repeat=20
🧹 Nitpick comments (8)
examples/amd/example_amd_flash_attn_fwd.py (2)
31-41: Search-space narrowing looks reasonable; consider guarding for future devices.Hard-coding threads to [128, 256, 512] may miss optimal configs on some ASICs (e.g., 64 or 1024). Consider deriving candidates as multiples of the wavefront/warp size and capping by occupancy.
Example tweak inside get_configs (illustrative):
- threads = [128, 256, 512] + wave = 64 + threads = [t for t in (64, 128, 256, 512, 1024) if t % wave == 0 and t <= 1024]
63-74: Static fallback config deviates from the enumerated space; please confirm intent.The appended config uses panel_size=64 and num_stages=1, while the grid enumerates panel_size in [7,8] and num_stages=[0]. If this is a known good fallback, add a short comment; otherwise align it with the search space.
- 'num_stages': 1, - 'panel_size': 64, + 'num_stages': 0, # align with enumerated search or document why 1 is preferred + 'panel_size': 8, # align or document why 64 is requiredsrc/target/intrin_rule_hip.cc (2)
1-4: Header comment says CUDA; update to HIP.Minor doc inconsistency; avoid confusion.
-/*! - * \file intrin_rule_cuda.cc - * \brief CUDA intrinsic rules. +/*! + * \file intrin_rule_hip.cc + * \brief HIP intrinsic rules.
121-128: Width/warp_size handling in shuffle lowering.You drop
warp_size(arg[4]) when lowering. That’s fine if width ≤ hardware warp_size, but unsafe otherwise. Add a sanity check or document the assumption.- ICHECK_EQ(call->args.size(), 5); // mask, value, warp_id, width, warp_size + ICHECK_EQ(call->args.size(), 5); // mask, value, warp_id, width, warp_size + // Sanity: enforce width <= warp_size at compile time when both are IntImm. + if (const auto* w = call->args[3].as<IntImmNode>()) { + if (const auto* ws = call->args[4].as<IntImmNode>()) { + ICHECK_LE(w->value, ws->value) << "shuffle width exceeds warp_size"; + } + }examples/amd/example_amd_flash_attn_bwd.py (4)
1-6: Avoid wildcard imports for better maintainability.Wildcard imports can lead to namespace pollution and make the code harder to understand.
Apply this diff to explicitly import needed symbols:
-from tilelang.autotuner import * +# Import specific symbols from autotuner when needed +# from tilelang.autotuner import specific_symbolIf autotuner symbols are not actually used in this file, consider removing this import entirely.
110-114: Consider clarifying the comment about 8x8 gemm fragment.The comment mentions an 8x8 gemm fragment reordering, but the lambda function implementation appears more complex than just handling 8x8 blocks. Consider adding more details about the specific memory layout transformation being performed.
def make_dq_layout(dQ): - # atomicAdd can not be vectorized, so we need to reorder dq to match the 8x8 gemm fragment + # atomicAdd cannot be vectorized, so we reorder dQ to match the 8x8 GEMM fragment layout + # This transformation: [b, l, h, d] -> [b, l//8, h, d//8, d%2, 4*(l%8) + (d%8)//2] + # ensures coalesced memory accesses during atomic operations return T.Layout(dQ.shape, lambda b, l, h, d: [b, l // 8, h, d // 8, (d % 2), 4 * (l % 8) + (d % 8) // 2])
178-180: Remove unnecessary blank lines.Multiple consecutive blank lines don't add value here.
Apply this diff to clean up the formatting:
dv_shared = T.alloc_shared([block_N, dim_v], dtype) dk_shared = T.alloc_shared([block_N, dim_qk], dtype) - - T.copy(K[bz, by * block_M:(by + 1) * block_M, bx // groups, :], K_shared)
89-89: Consider renaming ambiguous variable names.Variable names like
O(capital O) can be confused with the number zero, which reduces code readability. The PEP 8 style guide recommends avoiding such ambiguous names.Consider using more descriptive names like
outputoroutinstead ofOthroughout the file. Similarly, consider renamingltolengthorseq_idxin line 113.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (4)
examples/amd/example_amd_flash_attn_bwd.py(1 hunks)examples/amd/example_amd_flash_attn_fwd.py(1 hunks)examples/amd/test.sh(1 hunks)src/target/intrin_rule_hip.cc(1 hunks)
🧰 Additional context used
🧬 Code graph analysis (3)
examples/amd/example_amd_flash_attn_fwd.py (1)
tilelang/language/kernel.py (1)
threads(195-199)
src/target/intrin_rule_hip.cc (2)
tilelang/language/tir/op.py (3)
tvm_warp_shuffle(549-570)tvm_warp_shuffle_up(573-595)tvm_warp_shuffle_down(598-620)src/tl_templates/cuda/reduce.h (1)
T(75-147)
examples/amd/example_amd_flash_attn_bwd.py (9)
src/tl_templates/cuda/reduce.h (1)
T(75-147)tilelang/jit/__init__.py (2)
jit(237-310)compile(33-86)tilelang/language/allocate.py (2)
alloc_shared(21-36)alloc_fragment(53-64)tilelang/language/copy.py (1)
copy(84-152)tilelang/language/fill.py (2)
fill(9-21)clear(24-48)tilelang/language/pipeline.py (1)
Pipelined(9-46)tilelang/language/parallel.py (1)
Parallel(8-28)tilelang/language/tir/op.py (2)
if_then_else(2804-2834)floordiv(2970-2989)tilelang/language/reduce.py (2)
reduce_max(50-68)reduce_sum(87-109)
🪛 Ruff (0.12.2)
examples/amd/example_amd_flash_attn_bwd.py
4-4: from tilelang.autotuner import * used; unable to detect undefined names
(F403)
89-89: Ambiguous variable name: O
(E741)
113-113: Ambiguous variable name: l
(E741)
175-175: Local variable dv_shared is assigned to but never used
Remove assignment to unused variable dv_shared
(F841)
176-176: Local variable dk_shared is assigned to but never used
Remove assignment to unused variable dk_shared
(F841)
278-278: Use of assert detected
(S101)
280-280: Use of assert detected
(S101)
324-324: Ambiguous variable name: O
(E741)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: bot-task
🔇 Additional comments (4)
examples/amd/example_amd_flash_attn_bwd.py (4)
262-264: Verify device placement for gradient tensors.The gradient tensors are initialized on
q.device, but k and v might be on different devices in some edge cases.Consider using the correct device for each tensor:
dq = torch.zeros(shape_q, dtype=torch.float32, device=q.device) - dk = torch.zeros(shape_k, dtype=torch.float32, device=q.device) - dv = torch.zeros(shape_v, dtype=torch.float32, device=q.device) + dk = torch.zeros(shape_k, dtype=torch.float32, device=k.device) + dv = torch.zeros(shape_v, dtype=torch.float32, device=v.device)
273-296: Well-structured reference implementation!The reference implementation provides a clear, readable PyTorch baseline for validation. The comments effectively document the tensor dimensions and the grouped attention logic is handled correctly.
194-196: Verify the causal mask condition.The condition
by * block_M + i <= k * block_N + jappears to be checking if row index is less than or equal to column index for lower triangular masking. This seems inverted from the typical causal mask pattern where we mask out future positions.Please verify this is the intended behavior for causal attention. Typically, causal masks use
row >= colto prevent attending to future positions. If this is intentional (perhaps due to your specific indexing scheme), consider adding a comment to clarify.
298-355: Comprehensive testing and benchmarking setup!The main function provides excellent validation by comparing against PyTorch's reference implementation and includes performance benchmarking. The gradient checking and FLOPS calculation are particularly valuable for ensuring correctness and measuring efficiency.
| return flash_bwd | ||
|
|
||
|
|
||
| @torch.compile |
There was a problem hiding this comment.
Remove unnecessary @torch.compile decorator.
The @torch.compile decorator on a class is not meaningful - it's designed for functions and methods. Since this is an autograd function class, torch.compile isn't applicable here.
Apply this diff to remove the decorator:
-@torch.compile
class _attention(torch.autograd.Function):📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| @torch.compile | |
| class _attention(torch.autograd.Function): | |
| # rest of the implementation unchanged |
🤖 Prompt for AI Agents
In examples/amd/example_amd_flash_attn_bwd.py around line 224, the
@torch.compile decorator is incorrectly applied to an autograd Function class;
remove the decorator line above the class definition so the class is not
decorated, ensuring torch.compile is only used on functions or methods where
appropriate.
Comment on lines
+278
to
+281
| assert Q.size(2) == K.size( | ||
| 2) * groups, f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}" | ||
| assert Q.size(2) == V.size( | ||
| 2) * groups, f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}" |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Replace assertions with proper error handling.
Using assert statements for input validation is discouraged in production code as they can be disabled with Python's -O flag.
Apply this diff to use proper exceptions:
- assert Q.size(2) == K.size(
- 2) * groups, f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}"
- assert Q.size(2) == V.size(
- 2) * groups, f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}"
+ if Q.size(2) != K.size(2) * groups:
+ raise ValueError(f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}")
+ if Q.size(2) != V.size(2) * groups:
+ raise ValueError(f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}")📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| assert Q.size(2) == K.size( | |
| 2) * groups, f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}" | |
| assert Q.size(2) == V.size( | |
| 2) * groups, f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}" | |
| # Validate that Q’s feature dimension matches K’s feature dimension times the number of groups | |
| if Q.size(2) != K.size(2) * groups: | |
| raise ValueError( | |
| f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}" | |
| ) | |
| # Validate that Q’s feature dimension matches V’s feature dimension times the number of groups | |
| if Q.size(2) != V.size(2) * groups: | |
| raise ValueError( | |
| f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}" | |
| ) |
🧰 Tools
🪛 Ruff (0.12.2)
278-278: Use of assert detected
(S101)
280-280: Use of assert detected
(S101)
🤖 Prompt for AI Agents
In examples/amd/example_amd_flash_attn_bwd.py around lines 278 to 281, the two
input-validation assert statements should be replaced with explicit exception
handling because asserts can be disabled; raise a ValueError (or a more specific
exception used in the project) with the same formatted message for each check
instead of using assert, preserving the original diagnostic text that includes
Q.size(2), K.size(2)/V.size(2), and groups.
…rd example - Eliminated the allocation of shared memory for `dv_shared` and `dk_shared` in `example_amd_flash_attn_bwd.py` to streamline memory usage and improve performance. - This change focuses on optimizing the backward pass implementation by reducing unnecessary memory overhead.
- Eliminated the step to uninstall `torch`, `torchvision`, and `torchaudio` in the AMD CI workflow, as it is no longer required for the installation of pre-release versions. - This change simplifies the CI process and reduces potential overhead during package management.
- Updated the return statement to use std::string for concatenation in the case of 16-bit types, improving code clarity. - Added a null check for the CallNode pointer in DispatchHIPWarpActiveMask to enhance robustness and prevent potential dereferencing issues.
- Adjusted the formatting of TVM_REGISTER_OP calls for better readability by aligning method chaining. - No functional changes were made; this update focuses on code style improvements to enhance maintainability.
- Renamed the file from `intrin_rule_cuda.cc` to `intrin_rule_hip.cc` to accurately reflect the focus on HIP intrinsic rules. - Updated the file documentation to clarify its purpose as related to HIP rather than CUDA.
- Added NOLINTBEGIN and NOLINTEND comments to the DispatchHIPShuffle function to suppress clang-analyzer warnings related to inner pointer usage. - This change improves code clarity and maintains compliance with static analysis tools.
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (5)
examples/amd/example_amd_flash_attn_bwd.py (5)
11-11: Replace magic number 1.44269504 with a named constant.Define LOG2_E once and reuse for readability/maintainability. Mirrors a prior suggestion.
import argparse +LOG2_E = 1.4426950408889634 # math.log2(e) @@ - scale = (1.0 / dim_qk)**0.5 * 1.44269504 # log2(e) + scale = (1.0 / dim_qk) ** 0.5 * LOG2_E # log2(e) @@ - scale = (1.0 / dim_qk)**0.5 * 1.44269504 # log2(e) + scale = (1.0 / dim_qk) ** 0.5 * LOG2_E # log2(e)Also applies to: 140-140
110-114:make_dq_layoutis defined but unused. Remove or integrate.Either drop it or apply the layout when accumulating dQ to match the 8x8 fragment comment. Mirrors an earlier note.
-def make_dq_layout(dQ): - # atomicAdd can not be vectorized, so we need to reorder dq to match the 8x8 gemm fragment - return T.Layout(dQ.shape, - lambda b, l, h, d: [b, l // 8, h, d // 8, (d % 2), 4 * (l % 8) + (d % 8) // 2]) +# Removed unused helper `make_dq_layout`. Re-introduce if/when used.
274-277: Replace asserts with explicit validation errors.Asserts can be disabled with
-O. RaiseValueErrorinstead (also aligns with Ruff S101 from earlier feedback).- assert Q.size(2) == K.size( - 2) * groups, f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}" - assert Q.size(2) == V.size( - 2) * groups, f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}" + if Q.size(2) != K.size(2) * groups: + raise ValueError(f"Q.size(2): {Q.size(2)}, K.size(2): {K.size(2)}, groups: {groups}") + if Q.size(2) != V.size(2) * groups: + raise ValueError(f"Q.size(2): {Q.size(2)}, V.size(2): {V.size(2)}, groups: {groups}")
220-221: Remove@torch.compileon the autograd Function class.
torch.compileis for callables; decorating a class is meaningless and can mislead. This was flagged previously.-@torch.compile class _attention(torch.autograd.Function):
353-363: Use proper boolean CLI flag.
type=boolmakes--causal Falseparse as True. Preferaction='store_true'. This was already suggested previously.- parser.add_argument('--causal', type=bool, default=False, help='Causal flag') + parser.add_argument('--causal', action='store_true', help='Causal flag')
🧹 Nitpick comments (3)
examples/amd/example_amd_flash_attn_bwd.py (2)
4-4: Remove unused wildcard import (from tilelang.autotuner import *).Not referenced in this module and flagged by Ruff (F403). Drop it.
-from tilelang.autotuner import *
89-91: Avoid ambiguous variable names (O).Minor readability nit; consider
outinstead ofOin both places.Also applies to: 320-320
src/target/intrin_rule_hip.cc (1)
23-35: String construction is now safe; one remaining check.Using
"h" + nameandname + 'f'is valid (overloads exist). Keep it consistent: preferstd::string("h") + nameeverywhere for clarity across toolchains.- return "h" + name; + return std::string("h") + name; @@ - return name + 'f'; + return name + std::string("f")Also applies to: 58-67, 69-85
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (3)
.github/workflows/amd_ci.yml(0 hunks)examples/amd/example_amd_flash_attn_bwd.py(1 hunks)src/target/intrin_rule_hip.cc(1 hunks)
💤 Files with no reviewable changes (1)
- .github/workflows/amd_ci.yml
🧰 Additional context used
🧬 Code graph analysis (2)
src/target/intrin_rule_hip.cc (1)
tilelang/language/tir/op.py (3)
tvm_warp_shuffle(549-570)tvm_warp_shuffle_up(573-595)tvm_warp_shuffle_down(598-620)
examples/amd/example_amd_flash_attn_bwd.py (7)
tilelang/jit/__init__.py (2)
jit(237-310)compile(33-86)tilelang/jit/kernel.py (1)
out_idx(446-447)tilelang/language/allocate.py (2)
alloc_shared(21-36)alloc_fragment(53-64)tilelang/language/copy.py (1)
copy(84-152)tilelang/language/fill.py (2)
fill(9-21)clear(24-48)tilelang/language/pipeline.py (1)
Pipelined(9-46)tilelang/language/reduce.py (2)
reduce_max(50-68)reduce_sum(87-109)
🪛 Ruff (0.12.2)
examples/amd/example_amd_flash_attn_bwd.py
4-4: from tilelang.autotuner import * used; unable to detect undefined names
(F403)
89-89: Ambiguous variable name: O
(E741)
113-113: Ambiguous variable name: l
(E741)
274-274: Use of assert detected
(S101)
276-276: Use of assert detected
(S101)
320-320: Ambiguous variable name: O
(E741)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: bot-task
🔇 Additional comments (6)
examples/amd/example_amd_flash_attn_bwd.py (1)
40-49: Guard tail copies when seq_len % block_{M,N} != 0 to avoid OOB loads.Copies like T.copy(K[..., k*block_N:(k+1)block_N, ...], K_shared) and T.copy(Q[..., bxblock_M:(bx+1)*block_M, ...], Q_shared) rely on partial-tile semantics. Please confirm T.copy zero-fills/guards for shorter BufferRegions; otherwise, last tiles can read beyond bounds. If not guaranteed, clamp extents (e.g., valid = min(block, seq_len - base)) and limit dst slice or pre-clear shared buffers.
Do you want a patch that adds explicit tail handling for both fwd and bwd loops?
Also applies to: 56-56, 183-187, 193-199
src/target/intrin_rule_hip.cc (5)
116-120: Good fix:__activemasklowering passes zero args.The registration expects 0 inputs; forwarding an empty arg list here is correct.
218-233: HIP warp shuffle symbol names—verify availability across ROCm versions.
__shfl_*_syncexist in CUDA; HIP historically uses__shfl_*without*_sync. If these are aliases in your toolchain, great—otherwise this will fail to link.Would you like a follow-up patch that selects
__shfl_*vs__shfl_*_syncvia#ifdef __HIP_PLATFORM_AMD__and version checks?Also applies to: 242-286
186-217: Fast-math mapping coverage looks good.Broad set of math intrinsics wired to HIP fast variants; reasonable defaults for non-f32 paths.
87-101: Popcount mapping LGTM.32-bit →
__popc, 64-bit →__popcllis correct for HIP.
36-44: BF16 mapping likely mirrors half—please confirm intrinsics exist.Using
__habs/hrintforis_bfloat16()may not exist on HIP as BF16-specific builtins; ensure these resolve or gate them to half only.If needed, I can draft a fallback to call
fabsfvia cast for BF16.
LeiWang1999
approved these changes
Sep 4, 2025
This was referenced Sep 25, 2025
This was referenced Oct 5, 2025
Merged
This was referenced Oct 14, 2025
chengyupku
added a commit
to tile-ai/tilescale
that referenced
this pull request
Oct 24, 2025
* [Index] Relocate Int64 Auto Promoter to ConfigBitWidth Pass, removing it from FlattenBuffer (#714) * Update submodule 'tvm' to commit e11521e6936a827efa334588d29571fbb4620107 * Refactor inject_pipeline.cc to enhance pipeline body rewriting and condition handling - Introduced a new function to replace IfThenElse nodes with their then_case while preserving attributes. - Streamlined the PipelineBodyRewriter to improve buffer access rewriting and async state management. - Enhanced the handling of pipeline loop conditions and added support for predicate conditions in the pipeline body. - Removed obsolete code and improved overall code clarity and maintainability. * lint fix * Refactor return statements in inject_pipeline.cc to remove unnecessary std::move calls - Updated return statements in multiple methods to return objects directly instead of using std::move, improving code clarity and potentially avoiding unnecessary moves. - Ensured consistent handling of BufferStore and BufferLoad nodes during pipeline transformations. * test fix * Enhance global read detection in pipeline planning - Updated the handling of global reads to account for condition expressions within IfThenElse nodes, ensuring accurate identification of global memory accesses. - Introduced a new flag to track whether the visitor is within a condition expression, improving the correctness of buffer access analysis. - Refactored the VisitStmt_ method to properly handle the structure of IfThenElse nodes, enhancing the clarity and maintainability of the code. * Add IndexLegalizer to enforce int64 for out-of-bound indices - Introduced the IndexLegalizer class to ensure that indices in BufferStore and BufferLoad nodes are promoted to int64 when they exceed their type bounds. - Refactored the Int64Promoter logic from flatten_buffer.cc into IndexLegalizer, improving code organization and reusability. - Updated the ConfigIndexBitwidth pass to apply IndexLegalizer after rewriting the body, enhancing the handling of index bitwidths in transformations. * [CI] Bind build-test CI to NVIDIA as AMD runners are being introduced (#718) * Update submodule 'tvm' to commit e11521e6936a827efa334588d29571fbb4620107 * Rename build-test job to build-test-nvidia and specify nvidia as a runner label in CI workflow. * Update CI workflow to specify 'nvidia' as an additional runner label for the format-check job. * fix: NVRTC backend (#717) * fix: NVRTC backend * fix: CI --------- Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [CUDA] Init support for sm_120 (#716) * Init support for sm120 * fmt * resolve comments * unify mma gemm * fmt --------- Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [CI] fix docs ci (#720) * [Chore] fix typos (#719) * chore: fix typos * chore: fix ruff * chore: fix clang-format * [CI][AMD] Add AMD GPU CI and fix some related bugs (#694) * [Enhancement] Refactor buffer index handling for improved precision and clarity (#668) - Enhanced buffer index handling to address precision issues by removing redundant operations. - Streamlined the logic for determining buffer overlaps, ensuring more accurate conflict detection. - Updated related documentation to reflect changes in buffer management practices. * Remove obsolete test script for AMD example, streamlining the examples directory. * Remove unused dtype_size variable in AMD example script to streamline code. * Add input configuration file and update AMD example script for enhanced flexibility - Introduced a new input.txt file for configurable parameters. - Modified the example_amd_flash_attn_fwd.py script to allow for a wider range of configurations, including additional options for num_stages, enable_rasterization, and k_pack. - Streamlined the main function for better clarity and organization. - Added a new test script to facilitate running the example with specified parameters. * Remove input configuration file and obsolete test script; enhance AMD example with swizzle layout annotations - Deleted input.txt and test.sh files as they are no longer needed. - Updated example_amd_flash_attn_fwd.py to include swizzle layout annotations for shared memory, improving bank conflict avoidance. - Reintroduced swizzle usage in the kernel for better performance. * Refactor AMD example script for FlashAttention-2 - Updated function names for clarity, changing `get_v2_configs` to `get_configs` and `fast_flashattn_v2` to `fast_flashattn`. - Streamlined the main function by renaming `main_v2` to `main` and adjusting the corresponding calls. - Removed outdated comments and improved code organization for better readability. * Refactor formatting in AMD FlashAttention example script - Improved code readability by adjusting line breaks and indentation in the `fast_flashattn` function. - Streamlined the `main` function parameter formatting for consistency. - Removed unnecessary blank lines to enhance overall code organization. * Update example_amd_flash_attn_fwd.py * Update AMD FlashAttention example and TVM submodule - Added a new example script `example_amd_flash_attn_fwd_k_block.py` for FlashAttention with K-blocking support. - Enhanced `example_amd_flash_attn_fwd.py` by expanding configuration options for block sizes and threads. - Updated the TVM submodule to the latest commit for improved functionality. - Introduced a new test script `test.sh` to facilitate running the new example with specified parameters. * Add CI workflow for automated format checking and testing - Introduced a new GitHub Actions workflow in `amd_ci.yml` to automate format checks and testing for pull requests. - The workflow includes steps for setting up a Python environment, running format checks, and executing tests. - Removed obsolete example script `example_amd_flash_attn_fwd_k_block.py` and test script `test.sh` to streamline the examples directory. * Rename CI workflow from "CI" to "AMD CI" for clarity and specificity. * Update AMD CI workflow to include copying PyTorch, TorchVision, and Torchaudio packages to the virtual environment for improved dependency management. * Update AMD CI workflow to install pytest directly instead of using requirements-test.txt * Update AMD CI workflow to remove 'flash-attn' from requirements and install dependencies from requirements-test.txt * Refactor AMD CI workflow to enhance clarity in removing 'flash-attn' from requirements-test.txt before installation * Remove Torchaudio package copying from AMD CI workflow to streamline dependency management. * Refactor AMD CI workflow to remove the format-check job and streamline the build-test process by directly copying PyTorch and TorchVision packages to the virtual environment. * Add installation of ROCm in AMD CI workflow - Included a step to execute the `install_rocm.sh` script for improved setup. - Removed unnecessary blank line for better readability in the workflow script. * Remove installation step for ROCm in AMD CI workflow to simplify the setup process. * Update AMD CI workflow to run specific test file with verbose output instead of all tests. * Add new tilelang built-in operations for AMD architecture - Introduced `tvm_mfma`, `tvm_mfma_store`, `tvm_rdna_wmma`, and `tvm_rdna_wmma_store` built-in operations to enhance support for matrix multiplication and storage in tilelang. - Each operation is configured with the appropriate number of inputs and marked as opaque in terms of call effects. * Enhance autotuner configurations and GEMM operations in AMD example - Updated block sizes and num_split_q parameters in `get_configs` for improved autotuning. - Modified `T.gemm` calls in `fast_flashattn` to utilize `GemmWarpPolicy.FullRow`, optimizing performance for matrix multiplications. * Update autotuner configurations in AMD example for enhanced performance - Refined block sizes, thread counts, and added new parameters in `get_configs` to optimize autotuning. - Adjusted `fast_flashattn` function to incorporate new parameters for panel size and coalesced widths, improving memory access patterns. * Enhance autotuner configurations and memory handling in AMD example - Expanded block sizes and thread counts in `get_configs` for improved autotuning capabilities. - Updated `fast_flashattn` to utilize a new shared memory allocation strategy, optimizing memory access patterns during GEMM operations. * Refine autotuner configurations and memory usage in AMD example - Reduced block sizes and adjusted thread counts in `get_configs` for optimized autotuning. - Updated `fast_flashattn` to utilize register fragments for accumulation, minimizing LDS usage and enhancing performance during GEMM operations. * Update autotuner configurations in AMD example for enhanced performance - Expanded block sizes and thread counts in `get_configs` to improve autotuning capabilities. - Adjusted `num_split_q` and `v_coalesced_width` parameters for better optimization during GEMM operations. * Enhance autotuner configurations and GEMM operations in AMD example - Expanded thread counts in `get_configs` to include higher values for improved autotuning. - Updated `fast_flashattn` to adjust accumulation logic and ensure proper handling of causal conditions, optimizing performance during matrix multiplications. * Update AMD CI workflow and remove obsolete test script - Modified the CI workflow to run on multiple environments: self-hosted, amd, and gpu. - Deleted the outdated `test.sh` script from the examples directory, streamlining the project structure. * Remove TVM subproject from 3rdparty directory * Refactor configuration generation and accumulation logic in AMD example - Reformatted the `get_configs` function for improved readability by aligning parameters. - Adjusted the `fast_flashattn` function to enhance clarity in the conditional logic for accumulation, ensuring better handling of causal conditions. * Enhance AMD CI workflow with additional logging and setup steps - Added echo statements to provide feedback during the CI process, indicating when the environment is running on an AMD GPU, copying necessary packages, and installing requirements. - Improved clarity in the workflow by explicitly stating when the project is being installed and when tests are being executed. * Comment out package copying in AMD CI workflow to prevent potential issues during environment setup * Update AMD CI workflow to install nightly versions of PyTorch and remove obsolete package copying steps * Enhance BuildTileLangHIP function by adding whitespace for improved readability * Refactor kTVMGridConstant definition for clarity and remove unnecessary comment * Update TVM subproject to latest commit a64a5926a6e59f5417ef2501f9d88b467337cf6a * lint fix * Update AMD CI workflow to use requirements-rocm.txt for dependency installation * fix ci * Remove dependency on format-check from AMD CI workflow * fix ci * fix ci * fix ci * Remove format-check job from AMD CI workflow * Add torch to requirements-rocm.txt and remove explicit pip install commands from AMD CI workflow * Add dependency on format-check job in AMD CI workflow * Add format-check job to AMD CI workflow * Update format-check job in AMD CI workflow to run on self-hosted environment * Enhance format-check job in AMD CI workflow with improved Python environment setup and automatic commit of lint changes * Update amd_ci.yml --------- Co-authored-by: xinxyxiao <xinyxiao@amd.com> Co-authored-by: Lei Wang <34334180+LeiWang1999@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Carver][Bugfix] Correct score function for warp tile selection in tensorcore policy (#724) * [Carver][Bugfix] Correct score function for warp tile selection in tensorcore policy * [Typo] Correct architecture selection for CUDA and CDNA * [Refactor] Refactor CUDA code generation to simplify eviction policy handling (#721) * Update submodule 'tvm' to commit e11521e6936a827efa334588d29571fbb4620107 * Refactor CUDA code generation to simplify eviction policy handling - Updated `VisitExpr_` methods in `codegen_cuda.cc` to use default eviction policy for `tma_load`, `tma_load_im2col`, and `tma_store` functions, reducing complexity. - Removed conditional assembly code for `EVICT_NORMAL` in `copy_sm90.h`, streamlining the assembly calls for tensor memory operations. * lint fix * [Language] Introduce `StridedTensor` to support non contigious torch inputs (#722) * Update submodule 'tvm' to commit e11521e6936a827efa334588d29571fbb4620107 * Support strided tensors * Refactor target attribute helper functions for improved clarity * No code changes made in proxy.py and setup.py * lint fix * lint fix via gemini * lint fix * test fix * test fix * lint fix * Update wrapper.py * test fix * Enhance test for InjectSoftwarePipeline by adding LowerOpaqueBlock transformation and updating expected function signature to use match_buffer for better clarity. * lint fix --------- Co-authored-by: Chenggang Zhao <chenggangz@deepseek.com> * [Enhancement][Bugfix] Fix bug in warp specialized pass and add gemm_sr fallback support for Hopper (#712) * bug fix and support gemm_sr fallback for hopper * Update gemm.cc --------- Co-authored-by: Lei Wang <34334180+LeiWang1999@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * 📝 Add docstrings to `fix` (#726) Docstrings generation was requested by @LeiWang1999. * https://github.com/tile-ai/tilelang/pull/712#issuecomment-3190680851 The following files were modified: * `src/op/gemm.cc` * `src/tl_templates/cuda/gemm_sm90.h` * `src/transform/warp_specialized_rewriter.cc` Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * [CI] Fix AMD CI (#729) * [Enhancement] Refactor buffer index handling for improved precision and clarity (#668) - Enhanced buffer index handling to address precision issues by removing redundant operations. - Streamlined the logic for determining buffer overlaps, ensuring more accurate conflict detection. - Updated related documentation to reflect changes in buffer management practices. * Remove obsolete test script for AMD example, streamlining the examples directory. * Remove unused dtype_size variable in AMD example script to streamline code. * Add input configuration file and update AMD example script for enhanced flexibility - Introduced a new input.txt file for configurable parameters. - Modified the example_amd_flash_attn_fwd.py script to allow for a wider range of configurations, including additional options for num_stages, enable_rasterization, and k_pack. - Streamlined the main function for better clarity and organization. - Added a new test script to facilitate running the example with specified parameters. * Remove input configuration file and obsolete test script; enhance AMD example with swizzle layout annotations - Deleted input.txt and test.sh files as they are no longer needed. - Updated example_amd_flash_attn_fwd.py to include swizzle layout annotations for shared memory, improving bank conflict avoidance. - Reintroduced swizzle usage in the kernel for better performance. * Refactor AMD example script for FlashAttention-2 - Updated function names for clarity, changing `get_v2_configs` to `get_configs` and `fast_flashattn_v2` to `fast_flashattn`. - Streamlined the main function by renaming `main_v2` to `main` and adjusting the corresponding calls. - Removed outdated comments and improved code organization for better readability. * Refactor formatting in AMD FlashAttention example script - Improved code readability by adjusting line breaks and indentation in the `fast_flashattn` function. - Streamlined the `main` function parameter formatting for consistency. - Removed unnecessary blank lines to enhance overall code organization. * Update example_amd_flash_attn_fwd.py * Enhance AMD example script and update CI workflows - Improved the `example_amd_flash_attn_fwd.py` script for better clarity and organization. - Added new CI workflows for AMD and documentation publishing. - Updated various requirements files to include necessary dependencies. - Introduced new test cases and examples for better coverage and functionality. - Refactored existing code for improved readability and maintainability. * Remove redundant tool cache cleanup step in AMD CI workflow * Remove `torch` dependency from `requirements-rocm.txt` to streamline requirements. --------- Co-authored-by: xinxyxiao <xinyxiao@amd.com> Co-authored-by: Lei Wang <34334180+LeiWang1999@users.noreply.github.com> * [Feature] Low-bit twiddling dequantization and FP4 GEMM (#725) * [Dequant] Add bit-twiddling dequantize cuda for fp4-->bf16 * [Dequant] Add extern call and serial dequantization * [Dequant] Parallel Dequant wait for fence debug. * [Scale] Add scale matrix to mxfp4 gemm * [Remove] Remove fence-buggy example and some generated source cuda code * [MXFP4] Update initial version of MXFP4 GEMM * [Scale] Add scale to latest mxfp4 gemm * [Lint] * [BugFix] Load Scale, disabe TMA to recover performance * [Lint] * [Lint] * [Scale] Use L2 to hold Scale and enable TMA will slightly boost performance * [Lint] * Update example_dequant_gemm_bf16_fp4_hopper_serial.py * Remove deprecated dequantization examples for BF16 and MXFP4 in the dequantize_gemm directory. * Refactor dequantization examples for improved readability and consistency. Adjusted formatting in matmul function and added spacing for clarity. Updated function signatures and comments for better understanding. * Refactor index_to_coordinates usage in bitnet example and update dequantization example configurations. Removed the custom index_to_coordinates function and replaced it with the built-in version. Adjusted block_K parameter in dequantization example for consistency. * lint fix * ci fix * Remove non-existent example * [BugFix] Add smem swizzle to recover performance of TMA * [BugFix] Enough reg for producer when threads=512 --------- Co-authored-by: Lei Wang <34334180+LeiWang1999@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * 📝 Add docstrings to `mxfp4` (#732) * 📝 Add docstrings to `mxfp4` Docstrings generation was requested by @LeiWang1999. * https://github.com/tile-ai/tilelang/pull/725#issuecomment-3191656561 The following files were modified: * `examples/bitnet-1.58b/kernel_benchmark/tilelang_bitnet_158_int8xint2_prefill.py` * `examples/dequantize_gemm/example_dequant_gemm_bf16_fp4_hopper.py` * `examples/dequantize_gemm/example_dequant_gemm_bf16_mxfp4_hopper.py` * `examples/dequantize_gemm/utils.py` * `examples/gemm/example_gemm_autotune.py` * `tilelang/intrinsics/utils.py` * `tilelang/language/__init__.py` * `tilelang/language/utils.py` * `tilelang/quantize/mxfp.py` * `tilelang/quantize/quantization.py` * [Lint] More accurate docstring * [Lint] --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: tzj-fxz <tzjfxz@gmail.com> * [Refactor] Refactor env into a more flexible version (#740) * Fix environment variable name for compilation print setting in `env.py` * Remove deprecated test file for warp specialized pass configuration and refactor environment variable access in `env.py` to utilize a centralized `EnvVar` class for better management and clarity. * lint fix * Refactor cache check to use `env.is_cache_enabled()` for consistency in `tuner.py` * [Enhancement] Add stride index validation in CythonKernelWrapper (#743) * Introduced an assertion to ensure that the stride index is within the valid range of tensor dimensions in `cython_wrapper.pyx`. * This change prevents potential out-of-bounds errors when accessing tensor dimensions, enhancing the robustness of the code. * [Bugfix]:Fix atomic add auto vectorize memory access out of bound error (#742) * [Bugfix]:Fix atomic add auto vectorize memory access out of bound error * Update atomicadd_vectorize.cc * format * 📝 Add docstrings to PR #744 (#745) * 📝 Add docstrings to `main` Docstrings generation was requested by @LeiWang1999. * https://github.com/tile-ai/tilelang/pull/742#issuecomment-3205103559 The following files were modified: * `src/transform/atomicadd_vectorize.cc` * lint fix --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Refactor] Refactor barrier management (#744) * Introduce Barrier * Enhance CUDA kernel with new barrier management and post-processing support - Added a new CUDA kernel implementation in `example_mla_decode.py` for improved performance with shared memory barriers. - Refactored barrier handling in `codegen_cuda.cc` and `codegen_hip.cc` to utilize a more flexible mbarrier structure. - Updated intrinsic definitions from `ptx_stmatirx` to `ptx_stmatrix` across multiple files for consistency. - Introduced additional print statements for debugging in the lowering phase of the TileLang engine. - Enhanced the overall structure and readability of the codebase. * Remove unused barrier handling code in CUDA and HIP code generators to streamline the implementation. This change enhances code clarity and reduces complexity in the barrier management logic. * Enhance barrier management in TileLang - Introduced a new intrinsic `allocate_barrier` for dynamic barrier allocation in the TileLang framework. - Updated CUDA code generation to support the new barrier structure, allowing for improved synchronization in shared memory. - Refactored existing barrier handling logic to accommodate the new intrinsic and streamline code. - Added print statements for debugging purposes in various examples and the lowering phase of the TileLang engine. - Removed deprecated memory scope handling code to enhance clarity and maintainability. * lint fix * lint fix * Remove `allocate_barrier` intrinsic and related code from TileLang to streamline barrier management. This includes updates to CUDA code generation and the removal of associated Python wrappers, enhancing code clarity and maintainability. * Refactor logging in JITKernel to improve kernel compilation tracking - Removed unused import of `torch.backends` in the example file. - Introduced logging for kernel compilation in `JITKernel`, replacing print statements with structured logging for better traceability and debugging. - Added an assertion to ensure the presence of the `global_symbol` attribute in the kernel function. * Refactor dequantization tests and update barrier function - Removed the test for `example_dequant_gemm_bf16_fp4_hopper_serial` to streamline the testing suite. - Updated the `mbarrier_cp_async_arrive` function to support both pointer and non-pointer types, enhancing flexibility in barrier management. * Update CI configuration to increase pytest parallelism from 4 to 8 threads for improved test execution speed. * Fix typos in rasterization parameters and update import path for cached module - Corrected the spelling of `enable_rasteration` to `enable_rasterization` in the matmul function and its usage. - Updated the import statement for the `cached` module to reflect the new path in the cache submodule. - Added `StridedTensor` import in the language module for enhanced tensor functionality. * Update ci.yml * [Refactor] Merge bulk copy into copy and improve layout inference for bulk copy (#746) * [Refactor] Merge bulk copy into copy and refactor layout inference for bulk copy * Deleted the `bulk_copy` operator implementation and its header file as it is no longer needed. * Introduced a new function `cuTensorMapType()` to return the data type for CUDA tensor mapping. * Updated related files to reflect these changes, ensuring that the codebase remains clean and maintainable. * lint fix * Fix typos in intrinsic names and remove unused print statement in block_sparse_attn_tilelang.py. Updated references from `ptx_ldmatirx` to `ptx_ldmatrix` across multiple files for consistency. * remove bulk copy * Refactor copy and atomic add operations to support TMA lower configuration - Updated `GetCopyInst` to accept a `disable_tma_lower` parameter, allowing for conditional usage of TMA in bulk load/store operations. - Modified `Lower` method in `Copy` to incorporate the new TMA configuration. - Refactored `AtomicAdd::Lower` to streamline layout inference and vectorization logic. - Removed unused `disable_tma_lower` field from `LowerArgs` structure for clarity. - Enhanced atomic add vectorization by replacing the buggy implementation with a more robust loop vectorization approach. * Enhance TMA bulk copy logic in `LowerBulkCopy` method - Added a condition to set `desc.swizzle` to `CU_TENSOR_MAP_SWIZZLE_NONE` when `shared_layout` matches `linear_layout`, improving clarity in layout handling. - Updated warning log to provide more detailed information about fallback scenarios, including source and destination buffer names and shapes, enhancing debugging capabilities. * lint fix * Remove fallback logging for non-swizzled global layout in `LowerBulkCopy` method to streamline the bulk copy logic. This change enhances code clarity by eliminating unnecessary warning messages related to inner box dimensions. * Enhance reshape kernel compilation in `run_reshape` and `run_reshape_smem_1d_2_2d` functions - Updated the `tl.compile` method to include `pass_configs` that disable TMA lower and warp specialization, addressing shared memory layout transformation limitations. - Added TODO comments to indicate the need for further improvements in shared memory handling. * Update `native_sparse_attention` function to include TMA configuration options - Added `pass_configs` to the JIT decorator to disable TMA lower and warp specialization, addressing potential issues with shared memory layout transformations. - Updated comments to clarify modifications in tensor shapes for inference, specifically setting `q` sequence length to 1. * Refactor JIT decorator formatting in `native_sparse_attention` function - Improved readability by reformatting the JIT decorator parameters for `native_sparse_attention`, ensuring consistent style across the codebase. - No functional changes were made; this update focuses on code clarity and maintainability. * Enhance thread management and logging in TileLang compilation - Added a method to check if printing is enabled during compilation, improving control over logging behavior. - Updated the JIT kernel class to utilize the new method for logging compilation status, ensuring consistent and clear output. - Added comments to clarify the purpose of changes and improve code readability. * Add warp specialization scope and refactor register management in TileLang - Introduced a new constant `kWarpSpecializationScope` in `builtin.h` for better attribute management. - Removed the `SetMaxNRegCollector` class and its related logic from `warp_specialized_rewriter.cc`, streamlining the warp specialization process. - Added functions `annotate_producer_reg_dealloc` and `annotate_consumer_reg_alloc` in `builtin.py` to facilitate register management. - Implemented `AnnotateWarpGroupRegAlloc` in `__init__.py` to inject register allocation calls into warp-specialized functions, enhancing the overall register handling in the compilation process. * Refactor test for InjectSetMaxNReg pass in TileLang - Improved readability by restructuring conditional checks and assertions in the test cases. - Enhanced clarity in the collection of `set_max_nreg` calls by simplifying the logic. - Ensured consistent formatting and spacing throughout the test functions for better maintainability. * Enhance bulk copy and store checks in `Copy` class - Updated scope validation for source and destination tensors in `CheckBulkLoad` and `CheckBulkStore` methods to include both `shared.dyn` and `shared` as valid options. - Modified `CheckLDSMCopy` and `CheckSTSMCopy` methods to accommodate the new scope validation, ensuring compatibility with shared memory configurations. - Improved logging in `LowerBulkCopy` to provide clearer warnings regarding unsupported swizzle layouts, including source and destination names for better debugging. * lint fix * [Refactor] Merge ThreadPartialSync and ThreadStorageSync (#741) * Remove `thread_partial_sync.cc` and refactor `thread_storage_sync.cc` to streamline synchronization handling. Introduce `thread_sync_types.h` for thread-bound key definitions and reserved named barriers. Update related logic in `ThreadSyncInserter` and `TileLangThreadSync` for improved clarity and efficiency. * Remove `sync_thread_partial` references and related documentation from the codebase. Update CUDA and HIP code generation files to eliminate calls to the removed function. Refactor `__sync_thread_partial` to `sync_thread_partial` in CUDA common header for consistency. * Remove unused import of `bulk_copy.h` in `codegen_hip.cc` to enhance code clarity and maintainability. * Add import of `bulk_copy.h` in `codegen_hip.cc` to support new functionality. * typo fix * Update data type in reduce_sum tests from float16 to float32 for consistency and clarity. Remove redundant dtype tests and streamline run functions. Enhance reshape kernel compilation with pass configurations to address shared memory layout issues. * lint fix * test fix * Enhance CI configuration by adding verbose output to pip install command for better visibility during installation. * use ninja instead of make * Add CMake configuration step for Ninja build system in setup.py * Update pyproject.toml to include additional build dependencies: build, torch, tox, auditwheel, patchelf, and ninja. * Enhance CI configuration by adding verbose output to pytest commands for improved test visibility. * Update pyproject.toml to add Cython as a build dependency. Enhance thread storage synchronization in thread_storage_sync.cc by introducing new thread variable handling and improving index disjointness checks. * Update data type in cumulative sum tests from float16 to float32 for consistency. Modify run_cumsum function to utilize the updated dtype and enhance result validation with assertions. Adjust test cases accordingly. * Refactor storage access handling by introducing buffer data mapping in TileLangStorageAccessVisitor. Enhance access entry structure to include pointer access flag. Update thread storage synchronization to accommodate new buffer data mappings. Adjust quickstart example to print kernel source for debugging purposes. * Refactor linear index conversion in TileLangStorageAccessVisitor to utilize the analyzer for simplification. Update buffer index calculations to ensure consistent simplification of range expressions. * bugfix * Refactor buffer index calculation in TileLangStorageAccessVisitor to simplify access handling. Removed unused buffer mapping logic, ensuring consistent buffer index generation with a default ramp. * Refactor TileLangStorageAccessVisitor to replace buffer indices with buffer ranges for improved pointer access handling. Update AccessEntry structure to include buffer_ranges and adjust thread storage synchronization logic to account for pointer access conflicts. * Refactor thread storage synchronization to replace 'shared.dyn' with 'shared' for consistency in memory allocation. Update related test cases to reflect this change and ensure proper functionality. * [Enhancement] Optimize loop body handling in IR (#749) - Updated the loop body construction in `ir.cc` to conditionally include an output statement based on the analyzable condition of the `waves` variable. - This change enhances performance by avoiding unnecessary statement wrapping when the condition is met, improving the efficiency of loop execution. Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [MXFP4] Fix bugs and optimize exponential operation (#750) * [MXFP4] Fix bugs - Optimize exp2 with shift operation to boost performance - Fix bug of simple dequantization function call - Fix bug of scaling factor with bias * [Lint] --------- Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Enhancement] Add DispatchInstruction specialization for fp8 types in gemm_sm90.h (#751) - Introduced specialized DispatchInstruction templates for fp8_e4_t and fp8_e5_t types, enhancing support for new data formats in CUDA GEMM operations. - Each specialization defines the corresponding MMA and MMA_Group types, optimizing performance for specific configurations. * [Enhancement] Add shape checking for reduce options (#748) * Add shape checking for reduce options * lint fix * Handle special case reducing into shape-1 tensor Allow reducing [X, d, Y] into [X, Y] or [X, 1, Y] --------- Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Bugfix] Add missing FP8 header include (#752) * [Enhancement] Add DispatchInstruction specialization for fp8 types in gemm_sm90.h - Introduced specialized DispatchInstruction templates for fp8_e4_t and fp8_e5_t types, enhancing support for new data formats in CUDA GEMM operations. - Each specialization defines the corresponding MMA and MMA_Group types, optimizing performance for specific configurations. Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Enhancement] Include cuda_fp8.h in gemm_sm90.h - Added the inclusion of the "cuda_fp8.h" header file to support new data formats in CUDA GEMM operations, enhancing compatibility with recent updates for fp8 types. Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * lint fix * [Refactor] Remove unused tl_shuffle_elect and related functions from common.h - Deleted the `tl_shuffle_elect` function and its associated comments to streamline the codebase. - Added inclusion of "intrin.h" for improved intrinsic support in CUDA operations. - Cleaned up the file by removing unnecessary template parameters and functions, enhancing clarity and maintainability. * lint fix * [Refactor] Update header inclusions in common.h and gemm_sm90.h - Removed the inclusion of "intrin.h" from common.h to streamline dependencies. - Added "intrin.h" inclusion in gemm_sm90.h to ensure intrinsic support for CUDA operations, enhancing functionality and maintainability. * bug fix * [MXFP4] Add bias to MXFP4 GEMM kernel (#753) * [MXFP4] Add bias to gemm kernel * [Lint] * [Lint] Rename "bias" to "Bias" * [Bugfix][WS] Consider loop min extent when computing phase id (#754) * Update test parameters and remove debug print statement - Adjusted test cases in `test_tilelang_dynamic_symbolic_bench.py` to use smaller matrix sizes (1024x1024) for improved performance and quicker execution. - Removed a debug print statement from `phase.py` to clean up the code and enhance clarity. * Refactor loop stack management in warp_specialized_rewriter - Introduced a new `LoopInfo` struct to encapsulate loop variable details, including `loop_var`, `extent`, and `min`, enhancing clarity and maintainability. - Updated the `loop_stack_` to utilize `LoopInfo` instead of a pair, improving type safety and readability. - Adjusted linear index calculations to account for the new structure, ensuring correct behavior in loop transformations. * [Typo] Remove `disable_cache` in some tests (#755) * Update test parameters and remove debug print statement - Adjusted test cases in `test_tilelang_dynamic_symbolic_bench.py` to use smaller matrix sizes (1024x1024) for improved performance and quicker execution. - Removed a debug print statement from `phase.py` to clean up the code and enhance clarity. * Refactor loop stack management in warp_specialized_rewriter - Introduced a new `LoopInfo` struct to encapsulate loop variable details, including `loop_var`, `extent`, and `min`, enhancing clarity and maintainability. - Updated the `loop_stack_` to utilize `LoopInfo` instead of a pair, improving type safety and readability. - Adjusted linear index calculations to account for the new structure, ensuring correct behavior in loop transformations. * Remove unused `torch.backends` import and `tilelang.disable_cache()` calls from multiple test files to enhance code clarity and maintainability. * [README] Update GDN README for clarity and add acknowledgements (#758) - Improved formatting and clarity of the GDN kernel implementation description. - Updated requirement section to list dependencies in a clearer format. - Added an acknowledgements section to credit the developers and the Xiaomi LLM-Core Team for their contributions. * cutlass v4.2.0 supporting cuda 13 (#760) * [Feature] Add 1D TMA support (#761) * [Feature] Add 1D TMA support - Check the contiguous conditions of 1D TMA copy - Add new interface and params order of `tma_load` and `tma_store` call - Add 1D `tma_store` interface in sm90 template - Add elementwise kernel for 1D TMA example * [Lint] * [BugFix] Add conditions for 1D TMA copy on non-swizzle shared tensors * [Lint] * [BugFix] 1D TMA load * [README] Update GDN README for clarity and add acknowledgements (#758) - Improved formatting and clarity of the GDN kernel implementation description. - Updated requirement section to list dependencies in a clearer format. - Added an acknowledgements section to credit the developers and the Xiaomi LLM-Core Team for their contributions. * cutlass v4.2.0 supporting cuda 13 (#760) * [Lint] * [Lint] * [MXFP4] Add test for bf16&mxfp4 gemm * [BugFix] * [Lint] --------- Co-authored-by: Yu Cheng <54519279+chengyupku@users.noreply.github.com> Co-authored-by: Johnny <johnnync13@gmail.com> * [Example] Add vertical slash sparse attention pattern (#762) * upd sparse attn * lint * rename * update test file * update benchmark * lint * update benchmark * [Bugfix] Address PassContext contamination from CI and fix incorrect rewrites in warp specialized pass (#767) * fix ci and pass bug * fix * try * lint * [MXFP4] Add 1D TMA copy for Scale tensor in MXFP4 GEMM (#766) * [TMA] Add 1D TMA copy for Scale tensor * [Lint] * [Test] Add test for kernel * [BugFix] * hot fix blackwell (#768) * [Refactor] Refactor `Operator` into `TileOperator` and with tvm reflection (#763) * Refactor operator classes to inherit from TileOperator and update layout inference methods - Changed base class of several operator classes (AtomicAdd, Copy, Gemm, etc.) from Operator to TileOperator for better alignment with tile operations. - Updated InferLayout and Lower methods to use 'override' specifier for clarity and consistency. - Adjusted header inclusions to replace "op.h" with "operator.h" across multiple files for improved organization. - Added missing layout inference implementations for Fill and Conv2DIm2ColOp. - Removed deprecated op.cc and op.h files to streamline the codebase. * lint fix * Refactor operator classes to use Node pattern and improve memory management - Updated several operator classes (AtomicAdd, Copy, Gemm, etc.) to utilize the Node pattern for better memory management and encapsulation. - Changed constructors to initialize member variables through a node object, enhancing clarity and reducing direct member access. - Updated Clone methods to return TileOperator instances instead of unique pointers, aligning with the new design. - Refactored InferLayout and Lower methods to ensure consistency across operator implementations. - Adjusted header files to reflect the new class structure and removed deprecated code for a cleaner codebase. * Enhance Clone methods in AtomicAdd and Copy classes to support parallel operation cloning - Updated the Clone methods in AtomicAddNode and CopyNode to ensure that the parallel operation (par_op_) is properly cloned when defined, improving the integrity of cloned objects. - Refactored the FillNode class to use ParallelOp directly instead of std::make_unique, streamlining the creation of parallel operations. - Made minor adjustments in layout inference and other related methods for consistency and clarity. * Refactor FillNode::Lower method to remove unused global function call - Eliminated the call to the global function "tl.fill.lower" in the FillNode::Lower method, streamlining the code and improving clarity. - Retained the core functionality of the method while enhancing maintainability by reducing unnecessary dependencies. * [Reducer] Introduce `alloc_reducer` to separate inter and intra warp reduction (#757) * [Enhancement] Introduce finalize_reducer operator and layout reducer support - Added `FinalizeReducer` operator to handle reduction finalization in the TileLang framework, allowing for efficient reduction operations. - Implemented layout inference for local.reducer buffers, enhancing the handling of layout mappings and reducing complexity in buffer management. - Updated `setup.py` to include logging for build directory paths, improving build process visibility. - Enhanced atomic operations with new functions for atomic max, min, load, and store, providing more robust atomicity control in memory operations. - Refactored parallel loop handling to incorporate reducer information, ensuring proper management of reduction operations in parallel contexts. - Cleaned up test cases by removing unnecessary cache disabling and optimizing test parameters for better performance. * Refactor code formatting and improve readability in multiple files - Cleaned up whitespace in `setup.py` to enhance logging clarity. - Reformatted `AtomicMax` and `AtomicMin` functions in `common.h` for better alignment and readability. - Adjusted `debug_print_var` function in `debug.h` to improve code structure and maintainability. - Enhanced readability of the `atomic_add` function in `customize.py` by breaking long lines for better clarity. * Remove debug print statements from `copy.cc` and `inject_tma_barrier.cc` to enhance code clarity and maintainability. * [Enhancement] Disable reuse of small arrays in shared memory allocation - Added logic to prevent the reuse of small arrays (<= 32 bits) in `merge_shared_memory_allocations.cc`, ensuring they are lowered to registers in LLVM for improved performance and memory management. * Refactor `setup.py` to remove duplicate logging statements and enhance clarity. Update `finalize_reducer` function documentation in `reduce.py` to include detailed parameter and return descriptions, improving code readability and maintainability. * Refactor `finalize_reducer` and `reduce` functions to remove redundant target checks. Simplified conditionals by retaining only the `TargetIsHopper` check, enhancing code clarity and maintainability. * bug fix * Add thread checks workaround for replicated cases * Remove the is_one check * fix lint error * lint fix * Update autotune tests to use smaller matrix sizes for improved performance and reliability * [Refactor] Update FinalizeReducer to FinalizeReducerOp and adjust related methods - Refactored FinalizeReducer class to FinalizeReducerOp, updating constructor and method signatures for consistency with the new TileOperator structure. - Enhanced layout inference and cloning methods in FinalizeReducerOpNode. - Updated test_example_flash_attention.py to call test_example_gqa_bwd instead of tilelang.testing.main. - Adjusted header inclusions for improved organization and clarity across multiple files. * [Refactor] Update atomic operations in common.h and modify test_example_flash_attention.py - Enhanced atomic operations (Add, Min, Max) in common.h to handle half and bfloat16 types more efficiently. - Updated test_example_flash_attention.py to call test_example_gqa_bwd instead of tilelang.testing.main, improving test organization. * [Refactor] Simplify CopyNode::LowerBulkCopy logic and update test execution - Removed redundant checks for contiguous memory access in CopyNode::LowerBulkCopy, streamlining the logic for TMA copy operations. - Updated test_tilelang_kernel_gemm.py to comment out the main testing function and call a specific test for i8i8i32 tensor operations instead, improving test focus. --------- Co-authored-by: Huanqi Cao <caohuanqi@deepseek.com> Co-authored-by: Freebase6912 <amid-gauze-racing@duck.com> * 📝 Add docstrings to `pytile_0826` (#770) * 📝 Add docstrings to `pytile_0826` Docstrings generation was requested by @LeiWang1999. * https://github.com/tile-ai/tilelang/pull/763#issuecomment-3224197814 The following files were modified: * `src/op/atomic_add.cc` * `src/op/atomic_add.h` * `src/op/copy.cc` * `src/op/copy.h` * `src/op/elem.cc` * `src/op/elem.h` * `src/op/gemm.cc` * `src/op/gemm.h` * `src/op/gemm_sp.cc` * `src/op/gemm_sp.h` * `src/op/operator.cc` * `src/op/operator.h` * `src/op/parallel.cc` * `src/op/parallel.h` * `src/op/reduce.cc` * `src/op/reduce.h` * `src/op/region.cc` * `src/op/region.h` * `src/transform/layout_inference.cc` * `src/transform/lower_tile_op.cc` * lint fix --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * [Bugfix]:Fix atomic add auto vectorize negative optimization (#765) * [Bugfix]:Fix atomic add auto vectorize negative optimization * fixbug * format * fix bug * 📝 Add docstrings to `reducer_0825` (#772) * 📝 Add docstrings to `reducer_0825` Docstrings generation was requested by @LeiWang1999. * https://github.com/tile-ai/tilelang/pull/757#issuecomment-3219088118 The following files were modified: * `setup.py` * `src/op/builtin.h` * `src/op/finalize_reducer.cc` * `src/op/finalize_reducer.h` * `src/op/parallel.cc` * `src/op/parallel.h` * `src/op/reduce.cc` * `src/target/codegen_cuda.cc` * `src/tl_templates/cuda/common.h` * `src/transform/layout_inference.cc` * `src/transform/layout_reducer.cc` * `src/transform/layout_reducer.h` * `src/transform/merge_shared_memory_allocations.cc` * `src/transform/storage_access.cc` * `src/transform/warp_specialized_rewriter.cc` * `testing/python/autotune/test_tilelang_autotune_with_inputs.py` * `tilelang/engine/phase.py` * `tilelang/language/customize.py` * `tilelang/language/reduce.py` * `tilelang/transform/__init__.py` * lint fix * lint fix --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: LeiWang1999 <leiwang1999@outlook.com> * Allow fill global buffer (#774) * Allow fill global buffer * fix lint error * [BugFix] Refactor the op check in LowerTileOp pass using the member function instead of string match (#771) * [BugFix] Refactor the op check in LowerTileOp pass using the member function instead of string match * [Lint] * add bf16 exp fallback (#776) * [Lint] Introduce clang-tidy into format.sh (#777) * [Refactor] Update Clang-Tidy Checks and Improve Code Consistency - Enhanced .clang-tidy configuration by adding specific checks for better bug detection and performance optimization. - Refactored function signatures across multiple files to use `const` references for parameters, improving performance and code clarity. - Updated various methods to ensure consistent handling of parameters, particularly in `AddPredicate`, `Substitute`, and `PlanLoopPartition` functions. - Improved readability by replacing size checks with `empty()` method calls in several locations, ensuring clearer intent in the code. - General code cleanup and adherence to best practices for better maintainability. * [Refactor] Enhance Code Consistency and Clang-Tidy Configuration - Updated .clang-tidy configuration to include additional checks for improved code quality and performance. - Refactored function signatures across multiple files to use `const` references, enhancing performance and clarity. - Replaced size checks with `empty()` method calls in various locations for clearer intent. - Improved handling of parameters in several functions, ensuring consistent usage of `std::move` where applicable. - General code cleanup to adhere to best practices and improve maintainability. * [Refactor] Integrate Clang-Tidy Checks and Enhance Code Consistency - Added clang-tidy checks to the format script for improved code quality assurance. - Refactored function signatures across multiple files to consistently use `const` references, enhancing performance and clarity. - Updated the requirements-lint.txt file to include clang-tidy as a dependency. - General code cleanup to adhere to best practices and improve maintainability. * [CI] Update AMD CI Workflow to Include Build Directory Creation - Added steps to create a build directory and configure CMake with ROCm support during the format check process. - Ensured cleanup of the build directory after the format check to maintain a clean workspace. * [Refactor] Remove Unused Member Variables in AtomicAddNode and CopyNode - Removed the `args_` member variable from both `AtomicAddNode` and `CopyNode` classes to streamline the code and eliminate unnecessary data members. - This change enhances code clarity and maintainability by focusing on relevant attributes for each class. * [Refactor] Update Clang-Tidy Integration and Code Improvements - Modified the format script to include the `-fix` option in the clang-tidy command for automatic code fixes. - Refactored the `AtomicAddVectorizePlanner` class to improve variable handling and consistency, including changes to member variable types and function signatures. - Enhanced code clarity by removing unnecessary `std::move` calls and ensuring consistent usage of types across the class. - General code cleanup to adhere to best practices and improve maintainability. * [Refactor] Improve Parameter Handling and Consistency in AtomicAddVectorize - Updated function signatures in `AtomicAddVectorizePlanResult` and `AtomicAddVectorizeRewriter` to use `const` references and `std::move` for better performance and clarity. - Enhanced the `UpdateVectorSize` method to accept `const Array<PrimExpr>&` for improved efficiency. - General code cleanup to maintain consistency and adhere to best practices. * [CI] Add Git Submodule Initialization to CI Workflow - Included a step to initialize and update git submodules recursively in the CI workflow. - This change ensures that all necessary submodules are available during the format check process, improving build reliability. * [CI] Add Git Submodule Update Step to Format Check - Included a command to initialize and update git submodules recursively in the CI workflow during the format check process. - This enhancement ensures that all required submodules are available, contributing to improved build reliability. * [Refactor] Update Function Signatures in AtomicAddVectorize - Modified the `VectorizeAtomicAdd` function signature to use `const` references for `thread_var` and `thread_bounds`, enhancing performance and code clarity. - This change aligns with previous refactoring efforts to improve parameter handling and consistency across the codebase. * [Cache] Introduce detailed target information for the disk kernel cache (#780) * Fix type hint for target_host parameter in compile function to allow None value * Refactor target handling in compile function to utilize determine_target for improved clarity and consistency * Update PrintConst function in codegen_cuda.cc to use hexfloat format for bfloat16 and float8/float4 types, while adding scientific notation comments for clarity. This change enhances the representation of floating-point constants in the generated code. * Refactor PrintType function in codegen_cuda.cc to remove unnecessary failure conditions for floating-point types with lane counts greater than 4. This change simplifies the logic and improves code clarity. * Enhance benchmark_matmul.py to conditionally print Reference TFlops only if ref_latency is not None. Update param.py to ensure target is converted to string for consistency. Refactor tuner.py to utilize determine_target for improved clarity in target handling. * Remove automatic commit and push step from AMD and NVIDIA CI workflows to streamline the process and avoid unnecessary commits. * [Example]Adds example for top-k operation (#775) * [Example]Adds example for top-k operation Adds an example demonstrating the top-k operation using tilelang * format * Adds topk tilelang example test * fix lint * [Math] Dispatch `T.rsqrt(x)` into cuda intrin instead of `1 / T.sqrt(x)` (#781) * Fix type hint for target_host parameter in compile function to allow None value * Refactor target handling in compile function to utilize determine_target for improved clarity and consistency * Update PrintConst function in codegen_cuda.cc to use hexfloat format for bfloat16 and float8/float4 types, while adding scientific notation comments for clarity. This change enhances the representation of floating-point constants in the generated code. * Refactor PrintType function in codegen_cuda.cc to remove unnecessary failure conditions for floating-point types with lane counts greater than 4. This change simplifies the logic and improves code clarity. * Enhance benchmark_matmul.py to conditionally print Reference TFlops only if ref_latency is not None. Update param.py to ensure target is converted to string for consistency. Refactor tuner.py to utilize determine_target for improved clarity in target handling. * Remove automatic commit and push step from AMD and NVIDIA CI workflows to streamline the process and avoid unnecessary commits. * Add intrin_rule source files to CMakeLists.txt and implement hrsqrt function for half_t in common.h * lint fix * remove cmake dep in pyproject as it may lead to different cmake paths in diff stages * lint fix * Add cmake dependency to pyproject.toml and improve build logging in setup.py * [CI] Adds pytest-durations for test timing (#782) * [Ci] Adds pytest-durations for test timing Adds `pytest-durations` to the test requirements and configures pytest to display test durations. This helps in identifying slow-running tests and optimizing the test suite for faster feedback. * add amd ci durations * Removes flash_attn installation from CI * [Refactor] Support python reflection for tile operators (#783) * Implement Fill operator and related reflection methods in TileLang - Added Fill operator implementation in `fill.cc` and `fill.h` for element-wise filling of buffers. - Introduced reflection methods for Fill, AtomicAdd, Copy, Conv2DIm2Col, FinalizeReducer, Gemm, and Parallel operators to enhance introspection capabilities. - Updated relevant files to register reflection methods and ensure proper initialization in static blocks. - Removed outdated comments and unnecessary code in various operator files to improve clarity and maintainability. - Added new Python bindings for the Fill operator in `tilelang/ir/fill.py` and updated the module imports accordingly. * Refactor operator reflection methods and improve code clarity - Updated reflection methods for AtomicAdd, Copy, FinalizeReducer, Gemm, and Parallel operators to enhance readability by using `empty()` instead of size checks. - Consolidated static initialization blocks for various operators to a single line for improved consistency. - Cleaned up whitespace and formatting in multiple files to adhere to coding standards and improve maintainability. - Added new Python bindings for operators in the `tilelang/ir` module, ensuring proper registration and organization of imports. * Refactor GEMM and AtomicAdd operations for improved clarity - Updated the `GetArchInt` function in `atomic_add.cc` to use `std::string` and `std::stoi` for better readability and type safety. - Removed unnecessary variables and comments in `gemm_sp.cc` and `gemm.cc` to streamline the `ComputeWarpPartition` method. - Cleaned up the `layout_reducer.cc` file by removing unused variable declarations, enhancing code clarity. - Added import for the `ir` module in `tilelang/__init__.py` to ensure proper organization of module imports. * Remove deprecated operator files from the tilelang IR module - Deleted files for Fill, AtomicAdd, Copy, Gemm, GemmSP, FinalizeReducer, Parallel, Reduce, and Region operators to streamline the codebase. - This cleanup enhances maintainability by removing unused code and improving overall organization of the module. * Refactor imports in tilelang IR module for improved organization - Updated import statements in `tilelang/ir.py` to reflect changes in the TVM library structure, enhancing clarity and maintainability of the codebase. * lint fix * Refactor GEMM and GEMM-SP operations to enhance clarity and maintainability - Updated the `Gemm` and `GemmSP` classes to utilize a new `GemmWarpPolicy` object for warp partitioning, improving encapsulation and readability. - Removed deprecated `ComputeWarpPartition` methods and replaced them with calls to the new policy object, streamlining the code. - Cleaned up comments and unnecessary code in `gemm.cc`, `gemm_sp.cc`, and related header files to enhance overall clarity. - Introduced a new `GemmWarpPolicyNode` class to manage warp policy attributes and methods, facilitating better organization of related functionalities. - Updated reflection methods to include the new policy structure, ensuring proper registration and introspection capabilities. * Refactor Reduce operation to utilize ReduceType class for improved clarity and maintainability - Replaced multiple conditional checks for reduce types with a single ReduceType object, simplifying the code structure. - Introduced a new ReduceTypeNode class to encapsulate reduce type logic and methods, enhancing organization. - Updated MakeInitValue, MakeReduce, and Lower methods to leverage the new ReduceType class, improving readability. - Added Python bindings for the ReduceType class in tilelang IR module to ensure proper registration and usability. * comment * Refactor operator header files for improved readability - Cleaned up formatting and whitespace in `atomic_add.h`, `copy.h`, `fill.h`, `reduce.cc`, and `reduce.h` to enhance code clarity. - Consolidated comments and adjusted line breaks for better organization and maintainability across multiple operator definitions. * Refactor MakeReduce method in ReduceOpNode for clarity - Updated the parameter name in the MakeReduce method from `rhs` to `b` and assigned it to `rhs` for improved readability. - This change enhances the clarity of the method's purpose and aligns with the overall refactoring efforts in the Reduce operation. * Update Reduce operation type checks for consistency - Changed string comparisons for reduce types in the MakeReduce method from "abs_sum" to "abssum" and "abs_max" to "absmax" for uniformity. - This adjustment enhances the clarity and consistency of the reduce type handling in the codebase. * [AMD] Fix amd tir&add examples (#784) * [Enhancement] Refactor buffer index handling for improved precision and clarity (#668) - Enhanced buffer index handling to address precision issues by removing redundant operations. - Streamlined the logic for determining buffer overlaps, ensuring more accurate conflict detection. - Updated related documentation to reflect changes in buffer management practices. * Remove obsolete test script for AMD example, streamlining the examples directory. * Remove unused dtype_size variable in AMD example script to streamline code. * Add input configuration file and update AMD example script for enhanced flexibility - Introduced a new input.txt file for configurable parameters. - Modified the example_amd_flash_attn_fwd.py script to allow for a wider range of configurations, including additional options for num_stages, enable_rasterization, and k_pack. - Streamlined the main function for better clarity and organization. - Added a new test script to facilitate running the example with specified parameters. * Remove input configuration file and obsolete test script; enhance AMD example with swizzle layout annotations - Deleted input.txt and test.sh files as they are no longer needed. - Updated example_amd_flash_attn_fwd.py to include swizzle layout annotations for shared memory, improving bank conflict avoidance. - Reintroduced swizzle usage in the kernel for better performance. * Refactor AMD example script for FlashAttention-2 - Updated function names for clarity, changing `get_v2_configs` to `get_configs` and `fast_flashattn_v2` to `fast_flashattn`. - Streamlined the main function by renaming `main_v2` to `main` and adjusting the corresponding calls. - Removed outdated comments and improved code organization for better readability. * Refactor formatting in AMD FlashAttention example script - Improved code readability by adjusting line breaks and indentation in the `fast_flashattn` function. - Streamlined the `main` function parameter formatting for consistency. - Removed unnecessary blank lines to enhance overall code organization. * Update example_amd_flash_attn_fwd.py * Enhance AMD example script and update CI workflows - Improved the `example_amd_flash_attn_fwd.py` script for better clarity and organization. - Added new CI workflows for AMD and documentation publishing. - Updated various requirements files to include necessary dependencies. - Introduced new test cases and examples for better coverage and functionality. - Refactored existing code for improved readability and maintainability. * Remove redundant tool cache cleanup step in AMD CI workflow * Remove `torch` dependency from `requirements-rocm.txt` to streamline requirements. * Add new AMD FlashAttention example and test script - Introduced `example_amd_flash_attn_bwd.py` for backward attention computation using TileLang. - Added `test.sh` script to facilitate running the new example with specified parameters. - Enhanced the overall structure and organization of the example for better clarity and usability. * Update configurations in `example_amd_flash_attn_fwd.py` for autotuner - Reduced the number of threads and `num_split_q` options for improved performance. - Adjusted `panel_size` options to streamline configuration settings. * Update submodule 'tvm' to commit 6ccc74f622c7ec4ac25d430d0f6546e7b9edb217 * Update submodule 'tvm' to commit 14ff70ab142b9e5a31bbf9c7923c8a697d41e86c * Add example for AMD Flash Attention backward pass implementation - Introduced a new example script `example_amd_flash_attn_bwd.py` demonstrating the forward and backward operations of Flash Attention using TileLang. - Implemented JIT-compiled functions for both forward and backward passes, including preprocessing and postprocessing steps. - Added a main function to facilitate testing and benchmarking of the attention mechanism with configurable parameters. - Included reference implementation for validation against PyTorch's attention mechanism. This addition enhances the examples directory by providing a comprehensive guide for users to understand and utilize Flash Attention in their applications. * Enhance AMD Flash Attention example with additional testing capabilities - Updated `example_amd_flash_attn_bwd.py` to include more comprehensive testing features for the Flash Attention implementation. - Improved the main function to allow for better parameter configuration and benchmarking. - Added validation checks against PyTorch's attention mechanism to ensure accuracy and reliability of the example. This update aims to provide users with a more robust tool for understanding and utilizing Flash Attention in their applications. * Update submodule TVM to commit a64a5926a6e59f5417ef2501f9d88b467337cf6a * Refactor HIP intrinsic rules to CUDA - Updated file name from `intrin_rule_hip.cc` to `intrin_rule_cuda.cc` to reflect the change in focus from HIP to CUDA intrinsic rules. - Adjusted include paths for better organization and clarity in the code structure. * Update AMD CI workflow to uninstall specific PyTorch packages before installation - Removed the installation of `flash_attn==2.5.8` to streamline the CI process. - Added a step to uninstall `torch`, `torchvision`, and `torchaudio` prior to installing pre-release versions, ensuring compatibility and reducing potential conflicts. * Remove unused shared memory allocations in AMD Flash Attention backward example - Eliminated the allocation of shared memory for `dv_shared` and `dk_shared` in `example_amd_flash_attn_bwd.py` to streamline memory usage and improve performance. - This change focuses on optimizing the backward pass implementation by reducing unnecessary memory overhead. * Remove unnecessary pip uninstall command from AMD CI workflow - Eliminated the step to uninstall `torch`, `torchvision`, and `torchaudio` in the AMD CI workflow, as it is no longer required for the installation of pre-release versions. - This change simplifies the CI process and reduces potential overhead during package management. * Refactor DispatchHIPWarpActiveMask function in HIP intrinsic rules - Updated the return statement to use std::string for concatenation in the case of 16-bit types, improving code clarity. - Added a null check for the CallNode pointer in DispatchHIPWarpActiveMask to enhance robustness and prevent potential dereferencing issues. * Refactor formatting of HIP intrinsic rule registrations - Adjusted the formatting of TVM_REGISTER_OP calls for better readability by aligning method chaining. - No functional changes were made; this update focuses on code style improvements to enhance maintainability. * Update file na…
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit
New Features
Performance
Tests
Chores