CANN 9.1.1 Release Notes

Version Download Address

https://www.hiascend.com/en/cann/download

Version Mapping

CANN Software Version Mapping Table

Version Compatibility

CANN and Ascend HDK Version Compatibility

Internal Matching Relationships of CANN Software Packages

CANN Combination Compatibility

CANN 9.1.1 contains three combined packages: Toolkit, ops (operator package), and NNAL (acceleration library). The following table lists the matching relationship between the CANN combined packages and their subpackages. You can upgrade six subpackages independently as needed.

CANN ops and Toolkit Compatibility

The CANN ops package has been decoupled from Toolkit and supports independent upgrade. You can install it as needed.

Compatibility of Independently Upgradable Subpackages in the CANN ops Package

For the specific operations of independent subpackage upgrade, see Independent Subpackage Upgrade.

Removed and Deprecated Features

NOTE

• Deleted features are those that have been removed in the current version.
• Deprecated features are those marked as deprecated in the current version and will be removed in future versions.
• Replace ${install_path} in this section with the CANN software installation path.

Runtime

The following directories, files, and APIs are marked as deprecated in CANN 9.1.0 and are planned for removal in versions after June 30, 2027.

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and are planned for removal in versions after December 30, 2026.

Graph Engine

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and are planned for removal in versions after December 30, 2026.

Starting from Ascend 950, the aclblas and aclop APIs are no longer recommended. They will be gradually deprecated in later versions. Migrate to the corresponding aclnn operator APIs.

Operator Libraries

• Incompatibility changes: Control operators Identity, IdentityN, Rank, Shape, and ShapeN are migrated from ops-nn to ops-math. These operators are migrated from the ops_proto_nn.h and es_nn directories to the ops_proto_math.h and es_math directories. You need to update the related header file references (!2333, !3995).

The following directories, files, and APIs are marked as deprecated in CANN 9.0.0 and will be removed in later versions after March 30, 2027.

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and are planned for removal in versions after December 30, 2026.

Operators are split into component packages. Each independent component package of different operators contains multiple libraries and header files. In the following table, ${ops_project} represents the actual cv, math, nn, transformer, and legacy.

Acceleration Libraries

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and will be removed in versions after December 30, 2026.

The following operator IRs are deprecated for the DVPP acceleration feature: Crop, CropAndResize, DecodeAndCropJpeg, DecodeJpeg, PadV3D, Resize, ResizeV2, ResizeBicubic, ResizeBilinearV2, ResizeNearestNeighborV2, ReverseV2, WarpPerspective , WarpAffineV2, AdjustContrast, AdjustHue, AdjustSaturation.

Debugging and Analysis Tools

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and will be removed in versions after December 30, 2026.

Diagnostics

The following directories, files, and APIs are marked as deprecated in CANN 8.5.0 and will be removed in versions after December 30, 2026.

Model Compression Toolkit

The following features of the Model Compression Toolkit are marked as deprecated in CANN 9.0.0 and have been removed in this version.

• Non-uniform quantization
• Automatic mixed precision
• Approximate calibration
• int4 quantization-aware training
• All amct_mindspore features

Known Issues

Symptom: When the value of the input parameter deqScale of the operator aclnnMatmulCompressDequant is inf/-inf, inf or nan is displayed in its precision details. [Introduced in] CANN 8.5.0 [Impact] Triggered by specially constructed test cases, not introduced by a new operator, and in actual quantization tasks there is no case where deqScale is -inf or inf. The impact is controllable in the current version. [Workaround] In actual quantization tasks, there is no case where deqScale is -inf or inf.

Fixed Issues

The following issues are fixed in version 9.1.1:

• Fixed the issue where the HF32 attribute was not correctly passed to the bottom layer, causing performance degradation, when the MM operators on Atlas A2 and Atlas A3 products and Ascend 950 NPUs were executed through the GE aclnn fallback callback path.
• Fixed the precision error issue of the RmsNorm operator on Atlas A2 and Atlas A3 products when the input shape exceeds the int32 maximum value.
• Fixed the issue where the _deterministic_level attribute was not inherited when the Matmul operator fusion path on Atlas A2 and Atlas A3 products was converted to Matmul v2, causing the attributes of subsequent nodes to be lost.
• Fixed the issue where the fine-grained FSDP communication and computation pipeline was blocked under host-side performance degradation of the Matmul operators on Atlas A2 and Atlas A3 products, causing overall model performance degradation.
• Fixed the issue where the FAG operator occasionally failed to fully clear the pad in EOD use cases.
• Fixed the issue where the orderStream graph rule validation failed when aclgraph AICPU expansion used the same stream for the HCCL main stream and the capture stream on Atlas A3 products.
• Fixed the issue where the AllGatherRingConcurrentDirect algorithm deadlocked when dispatching primary/standby stream tasks in large-scale clusters on Atlas A3 products.
• Fixed the issue where the RingMLAOperation API demo failed to run when importing environment variables from a non-mainline version on Atlas A2 and Atlas A3 products.
• Fixed the issue where torch_npu.npu_format_cast fell back from interception to aclop when the format and shape did not match on Atlas A2 and Atlas A3 products.
• (Incompatible change) Fixed the issue that the default values of pri_mask_mode and cmp_ori_win_right for the SparseFlashMLA operator's torch API in Atlas A2 and A3 products.

The following issues are fixed in version 9.1.0:

• Fixed the issue where the QuantBatchMatmulV3 operator on Atlas inference products caused read/write conflicts when AUTO_SYNC=false was specified during compilation.
• For Atlas A2 products and Atlas A3 products, fixed the issue where the MatmulReduceScatterV2 operator cleared the flag bit too quickly in cases with M less than 512, causing other devices to fail to detect it and eventually time out.
• Fixed the issue where, with MoE EP on Atlas A2 products, multi-bit faults injected into the on-chip memory of a D node prevented entry into the quick recovery process.
• Fixed the issue where communication timeout occurred probabilistically at random step positions after 500+ steps in a 4k-device training task.
• Fixed the refill backlog issue caused by frequent task interruption and restart of decode in PD disaggregation.
• Fixed the issue where the IR prototype name conflict between user-defined custom operators and built-in operators in the open-source repository caused the IR prototype name not to take effect.
• Fixed the issue where the aclnnAddmv operator occasionally triggered core dump during load testing.
• Fixed the issue where an exception was reported when a 10,000-device task was launched.
• Fixed the issue where MatMulV3 performance was worse than MatMulV2 for specific shapes in aclgraph use cases.
• Fixed the precision issues for fused_quant_mat_mul.
• Fixed the precision issues of BatchMatmulV2 with specific bias.
• Fixed the precision issues of the FAG operator with short-sequence GQA and NoMask.

Fixed Vulnerabilities

For details about the fixed vulnerabilities for open-source and third-party software in this version, see Fixed Vulnerabilities.