Skip to content

Bump paddlepaddle from 2.6.0 to 3.3.0 in /tests#237

Open
dependabot[bot] wants to merge 1 commit intomasterfrom
dependabot/pip/tests/paddlepaddle-3.3.0
Open

Bump paddlepaddle from 2.6.0 to 3.3.0 in /tests#237
dependabot[bot] wants to merge 1 commit intomasterfrom
dependabot/pip/tests/paddlepaddle-3.3.0

Conversation

@dependabot
Copy link

@dependabot dependabot bot commented on behalf of github Mar 12, 2026

Bumps paddlepaddle from 2.6.0 to 3.3.0.

Release notes

Sourced from paddlepaddle's releases.

PaddlePaddle 3.3.0 Release Notes

重要更新

飞桨框架 3.3 版本在大模型训练效率、开发体验及国产硬件适配等关键领域持续突破创新,在计算显存高效利用、训推衔接转换、生态兼容性、调试效率、国产硬件适配等方面带来多项重要升级,全面提升大模型训推能力。

训练效率突破

  • FlashMaskV3升级:深度优化稀疏注意力掩码计算FlashMaskV3计算内核,性能全面超越FlexAttention,算子性能最高领先2.1 倍;原生支持上下文并行并引入计算负载均衡机制,分布式场景算子性能相比Megatron-LM快80%,全面强化长文训练能力。
  • FlexCheckpoint参数自动切分重组系统:基于首创的轻量级描述语言AOA (All in One Arrow),支持从单卡视角灵活描述复杂的权重转换关系,并自动推导分片映射;通过跨机通信的高并发与负载均衡调度,在大参数规模下权重转换性能相比Megatron-LM领先1.2倍以上,有效解决大模型训推不同阶段参数转换的成本与效率难题。
  • 虚拟内存动态碎片整理:引入基于虚拟内存技术的显存分配机制,可根据运行时显存占用情况动态进行碎片整理。在主流MoE模型训练中,显存碎片率从超过10%降至最低3%,显著提升显存利用率。

开发体验优化

  • 生态兼容:通过框架API、算子注册、执行调度等关键链路的兼容设计,实现无缝衔接使用外部生态算子,支持FlashInfer、FlashMLA、DeepGEMM、DeepEP、TorchCodec和SonicMoE等社区高性能模块的高效集成应用。
  • 动态图调试能力升级:新增动态图前反向计算图可视化功能,支持导出算子调用栈与张量MD5校验信息;系统优化关键路径日志,新增局部日志打印功能,提升调试信息丰富度与获取便捷性。
  • 显存观测工具:新增显存观测功能,可视化显存池中内存块分布,支持追踪特定代码段的显存申请/释放与全局状态,助力大模型显存异常的精准定位与处置。

国产硬件深度适配

  • 昆仑芯 XPU:系统完善MoE场景适配,为相关算子新增bool、bfloat16、complex64数据类型支持,并对FlashAttention、DeepEP、Profiler等模块进行深度适配。
  • 海光 DCU:支持 Hygon数学库后端,进一步优化海光芯片推理性能。

1. 执行调度机制

针对大模型预训练、后训练与推理阶段中,因分布式策略及组网差异导致的权重转换困难,飞桨FlexCheckpoint机制创新提出高效的权重重组方法与灵活的模型编辑原语AOA,为模型研发到生产全流程提供高效、统一的分布式参数转换与重组支持。该机制全面适配训推衔接、跨策略断点续训、生态兼容格式加载导出、强化学习参数同步等多种场景下的权重转换需求,并在大参数规模场景下实现超越Megatron-LM 1.2倍以上的转换性能,从根本上解决了分布式参数转换成本高、效率低的难题。 在显存管理方面,针对MoE模型因专家路由动态性导致的显存碎片率高、资源浪费严重等问题,本版本推出了基于虚拟内存管理技术的高性能VMM Allocator。该分配器可在模型运行期间根据系统显存使用情况,动态自适应执行碎片整理,显著提升显存利用效率。 飞桨3.3版本持续深耕自动并行与AI编译器技术。在自动并行架构中新增对FSDP等策略的支持,并增强了动态Shape流水线并行功能;面向科学计算场景,实现了对高阶导数的支持,并拓展了相关算子切分推导规则的覆盖范围,进一步提升了自动并行架构的通用性与易用性。此外,围绕推理场景,优化了AI编译器CINN及动转静SOT功能,显著提升易用性与执行调度性能。

新特性

功能增强

  • FlexCheckpoint优化AOA宏展开功能,为fuse类宏支持传入axis属性。#75282
  • FlexCheckpoint AOA解析功能增强,支持优化器状态与模型状态共用一套AOA标记,切分信息传递、模型加载与存储共用一套AOA标记。#75613, #76013, #76437
  • FlexCheckpoint支持ShardingStage2和ShardingStage3策略。#76309, #76538
  • 优化FlexCheckpoint报错信息。 #76813, #77266
  • 自动并行中层 API 流水线 Hook 支持处理元组对象。#75081
  • 自动并行流水线并行策略支持动态shape。#75724
  • 支持自动并行场景下的FlexCheckpoint机制及优化器状态字典的分片处理。#76240, #76305
  • 升级 DLPack 到v1.2 版本,全面支持 TVM FFI,支持 C 函数交换协议、DataType 交换协议、Device 交换协议等新特性。#75193, #75205, #75650, #75854, #75973, #76828, #76673
  • 将ComparePriority 函数返回类型从 bool升级为int,以满足 std::sort 的严格弱序要求,并引入 SortComparePriority 包装函数确保排序算法正确性。#76027
  • 动转静支持禁用编译超时自动回退功能。#76386
  • 统一EqualAllOpInferSymbolicShape 函数与InferMeta 逻辑,提升编译器与框架动态图调度结果的一致性。#76477
  • 为all_reduce,c_allreduce_sum,c_concat,c_identity,flash_attn_unpadded,mp_allreduce_sum等算子添加 infer_symbolic_shape 接口,支持符号化形状推导。#76783, #76836

... (truncated)

Commits

Dependabot compatibility score

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.


Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

  • @dependabot rebase will rebase this PR
  • @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
  • @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
  • @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

Bumps [paddlepaddle](https://github.com/paddlepaddle/paddle) from 2.6.0 to 3.3.0.
- [Release notes](https://github.com/paddlepaddle/paddle/releases)
- [Changelog](https://github.com/PaddlePaddle/Paddle/blob/develop/RELEASE.md)
- [Commits](PaddlePaddle/Paddle@v2.6.0...v3.3.0)

---
updated-dependencies:
- dependency-name: paddlepaddle
  dependency-version: 3.3.0
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
@dependabot dependabot bot added dependencies Pull requests that update a dependency file python Pull requests that update Python code labels Mar 12, 2026
@dependabot @github
Copy link
Author

dependabot bot commented on behalf of github Mar 12, 2026

Assignees

The following users could not be added as assignees: akuporos, jiwaszki, p-wysocki, rkazants. Either they do not exist or they do not have the correct permissions to be added as an assignee.

Please fix the above issues or remove invalid values from dependabot.yml.

@dependabot dependabot bot added dependencies Pull requests that update a dependency file python Pull requests that update Python code labels Mar 12, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

category: dependency_changes dependencies Pull requests that update a dependency file python Pull requests that update Python code

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants