From 5c0742ed07c097a1b57e0726e67e2c3c3a9a1f62 Mon Sep 17 00:00:00 2001
From: mudler <2420543+mudler@users.noreply.github.com>
Date: Mon, 3 Aug 2026 22:31:26 +0000
Subject: [PATCH 1/3] :arrow_up: Update TheTom/llama-cpp-turboquant

Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
---
 backend/cpp/turboquant/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/backend/cpp/turboquant/Makefile b/backend/cpp/turboquant/Makefile
index 162d09fd1268..91e5831aa72a 100644
--- a/backend/cpp/turboquant/Makefile
+++ b/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?=8a891f4b566efdbd3cea92fafee3227a0a267683
+TURBOQUANT_VERSION?=d0e2a8b64de4d42568932886fec638ef2a7e1d10
 LLAMA_REPO?=https://github.com/TheTom/llama-cpp-turboquant
 
 CMAKE_ARGS?=

From 216c5541702992a9f9d9b4ac1fc5982b288aacf1 Mon Sep 17 00:00:00 2001
From: localai-org-maint-bot
 <306269227+localai-org-maint-bot@users.noreply.github.com>
Date: Tue, 4 Aug 2026 04:08:33 +0000
Subject: [PATCH 2/3] chore(turboquant): advance past post-rebase build fixes

The initial post-rebase snapshot is followed by arm64 CI fixes and restored Vulkan TurboQuant weight-type wiring. Pin the corrected branch head so the dependency bump includes those fixes.

Assisted-by: Codex:gpt-5 [Codex]
---
 backend/cpp/turboquant/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/backend/cpp/turboquant/Makefile b/backend/cpp/turboquant/Makefile
index 91e5831aa72a..1e76b2bb58d2 100644
--- a/backend/cpp/turboquant/Makefile
+++ b/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?=d0e2a8b64de4d42568932886fec638ef2a7e1d10
+TURBOQUANT_VERSION?=21eab71bc6c5f6c8194e6494006f25d979d8b6c4
 LLAMA_REPO?=https://github.com/TheTom/llama-cpp-turboquant
 
 CMAKE_ARGS?=

From f83a64393b8620c4023f5414c82c0e800f4f8339 Mon Sep 17 00:00:00 2001
From: localai-org-maint-bot
 <306269227+localai-org-maint-bot@users.noreply.github.com>
Date: Tue, 4 Aug 2026 15:06:26 +0000
Subject: [PATCH 3/3] chore(turboquant): advance past worker stack fix

Pin the direct follow-up that removes TurboQuant's 64 KiB stack-local rotation matrix, which overflows reduced worker-thread stacks during backend tests.

Assisted-by: Codex:GPT-5 [Codex CLI]
---
 backend/cpp/turboquant/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/backend/cpp/turboquant/Makefile b/backend/cpp/turboquant/Makefile
index 1e76b2bb58d2..d9f9c5bba17e 100644
--- a/backend/cpp/turboquant/Makefile
+++ b/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?=21eab71bc6c5f6c8194e6494006f25d979d8b6c4
+TURBOQUANT_VERSION?=0967f499714dd6018494b480b710b849ca45b156
 LLAMA_REPO?=https://github.com/TheTom/llama-cpp-turboquant
 
 CMAKE_ARGS?=