Search results

ACL: INT8 3x3 slower than F32 on small-channel conv (Neoverse-V2, SVE2)

Status: Open.

#1225 In ARM-software/ComputeLibrary;

· allnes opened on Dec 5, 2025

SME1 kernel support for AI workloads using ArmNN

Status: Open.

Feature

#1220 In ARM-software/ComputeLibrary;

· hpc-qti opened on Nov 20, 2025

NEConvolutionLayer - performance issue with <code>i8</code> input / <code>f32</code> output

Status: Open.

#1213 In ARM-software/ComputeLibrary;

· alvoron opened on Nov 13, 2025

ARM-software/ComputeLibrary conflicts with kleidiai

Status: Open.

#1178 In ARM-software/ComputeLibrary;

· BillyONeal opened on Sep 6, 2025

<code>NEQuantizationLayer</code> does not support <code>QSYMM8_PER_CHANNEL</code> as output

Feature Request

Status: Open.

Feature

#1168 In ARM-software/ComputeLibrary;

· alvoron opened on Apr 29, 2025

·  

v52.3.0

Dynamic support for gemmlowp

Feature Request

Question

Status: Open.

#1162 In ARM-software/ComputeLibrary;

· mason20240920 opened on Mar 10, 2025