Earlier this month Intel compiler engineers began [1]adding AVX10.2 support into the GCC 15 open-source compiler . Now as we approach the end of August, another big batch of AVX10.2 enablement has landed for this next GNU Compiler Collection release.

The initial batch of AVX10.2 enablement earlier in the month was mostly focused on adding the infrastructure with the new "-max10.2-256" and "-mavx10.2-512" switches and similar. AVX10.2 is coming with future Intel processors and is notable in that it's the first AVX10 version that will be found on both Intel P cores and E cores. AVX10.2 adds new AVX10 BF16 instructions, compare scalar FP with enhanced eflags, new convert instructions, integer and FP16 VNNI media new instructions, new min/max instructions, and saturating convert instructions.

With today's GCC 15 Git code, more of those new AVX10.2 instructions are now implemented. The newly-merged GCC 15 code adds the AVX10.2 instructions for compare, vector copy, min max, convert, BF16 instructions, and new media instructions as well.

It's great seeing Intel's timely open-source compiler enablement of new features continues. The LLVM/Clang work around AVX10.2 also [2]continues .

GCC 15 will be introduced as stable with the GCC 15.1 release in the first few months of 2025. It's still not publicly clear what will be the first P and E cores sporting AVX10.2 but at least there will be a supported open-source compiler out in time.



[1] https://www.phoronix.com/news/GCC-15-Merges-Initial-AVX10.2

[2] https://github.com/search?q=repo%3Allvm%2Fllvm-project+AVX10.2&type=pullrequests