ispc Files

ISPC release featuring Intel AMX (Advanced Matrix Extensions) support in the standard library, performance improvements for backward memory access patterns, and bug fixes. Based on a patched LLVM 21.1.8.

Standard Library:

• Intel AMX (Advanced Matrix Extensions) support has been added to the standard library. AMX provides hardware acceleration for matrix operations, particularly useful for machine learning workloads. The new <amx.isph> header provides functions for tile configuration, data loading/storing, and matrix dot products for INT8, BF16, and FP16 data types. AMX is supported on avx512spr, avx512gnr, and avx10.2dmr targets.

Language Changes:

• Integral type aliases (size_t, ptrdiff_t, intptr_t, uintptr_t) can now be used as non-type template parameters.

Performance:

• Optimized backward memory access patterns (e.g., dst[size-1-i]) to use contiguous load/store operations with vector shuffle, providing 5-10x speedup compared to scatter/gather operations.

Deprecated Targets:

• The sse2-i32x4 and sse2-i32x8 targets are no longer deprecated. Based on customer feedback indicating active use, we have decided to retain these targets and removed the deprecation warning.

Bug Fixes:

• Fixed integral type aliases not being accepted as non-type template parameters.

• Fixed varying control flow regression on NEON targets introduced in v1.26.0.

• Fixed performance regression on Apple Silicon (and other ARM platforms) introduced in v1.26.0, which caused up to 30% slowdown in some workloads.

• Fixed sub-optimal code generation when using extract() that caused unnecessary stack spills.

Build System:

• Updated default LLVM version to 21.1.8.
• Added support for building with LLVM 22.0 and LLVM 23.0.

Recommended versions of Runtime Dependencies when targeting GPU:

Linux:

• Intel(R) Graphics Compute Runtime https://github.com/intel/compute-runtime/releases/tag/25.13.33276.16
• Level Zero Loader https://github.com/oneapi-src/level-zero/releases/tag/v1.20.2
• Threading Building Blocks (TBB)

Alternatively, you can use a validated gfx driver stack supporting Intel Arc(TM) available at https://dgpu-docs.intel.com/driver/installation.html

Windows:

• Intel(R) Graphics Windows(R) DCH Drivers 32.0.101.8250 https://www.intel.com/content/www/us/en/download/785597/869290/intel-arc-graphics-windows.html
• Level Zero Loader https://github.com/oneapi-src/level-zero/releases/tag/v1.20.2
• OpenCL(TM) Offline Compiler (OCLOC) https://www.intel.com/content/www/us/en/developer/articles/tool/oneapi-standalone-components.html (this is needed for AoT compilation on Windows only)
• Supported GPU platforms: Intel(R) Arc Graphics, 11th-13th Gen Intel(R) Core processor graphics

Components revisions used in GPU-enabled build:

• https://github.com/KhronosGroup/SPIRV-LLVM-Translator/commit/cce2baf
• https://github.com/intel/vc-intrinsics/commit/ddfa825
• https://github.com/oneapi-src/level-zero/commit/v1.20.2
• https://github.com/llvm/llvm-project/commit/2078da4 (llvmorg-21.1.8) + patches from llvm_patches folder