fix(turboquant): restore the stable pin

The new fork revision breaks CPU, ARM64, CUDA, SYCL, Vulkan, and gRPC builds. Restore the last compatible revision.

Assisted-by: Codex:gpt-5.6 [Codex]

Author: localai-org-maint-bot

backend/cpp/turboquant/Makefile

@@ -1,7 +1,7 @@
 
 # Pinned to the HEAD of feature/turboquant-kv-cache on https://github.com/TheTom/llama-cpp-turboquant.
 # Auto-bumped nightly by .github/workflows/bump_deps.yaml.
-TURBOQUANT_VERSION?=fca3093c9e6544476bbb2a139a25e17dd63627e1
+TURBOQUANT_VERSION?=8a891f4b566efdbd3cea92fafee3227a0a267683
 LLAMA_REPO?=https://github.com/TheTom/llama-cpp-turboquant
 
 CMAKE_ARGS?=