diff --git a/.github/container/Dockerfile.triton b/.github/container/Dockerfile.triton
index ee0cbc4d9..1a7e90d58 100644
--- a/.github/container/Dockerfile.triton
+++ b/.github/container/Dockerfile.triton
@@ -65,6 +65,8 @@ sed -i -e 's|-Werror||g' CMakeLists.txt
 sed -i -e 's|\(LLVMAMDGPU.*\)|# \1|g' CMakeLists.txt
 # Do not build tests
 sed -i -e 's|^\s*add_subdirectory(unittest)|# unit tests disabled|' CMakeLists.txt
+# Avoid building GPUHello.cpp, which does not compile with cl693214397 of OpenXLA Triton
+sed -i -e 's|^add_subdirectory(Instrumentation)|# Instrumentation test disabled|' test/lib/CMakeLists.txt
 # Avoid error due to forward declaration of Module
 sed -i -e '/#include "llvm\/IR\/IRBuilder.h"/a #include "llvm/IR/Module.h"' test/lib/Instrumentation/GPUHello.cpp
 # Google has patches that mess with include paths in source files
diff --git a/.github/container/build-jax.sh b/.github/container/build-jax.sh
index db01bccff..95cf5246b 100755
--- a/.github/container/build-jax.sh
+++ b/.github/container/build-jax.sh
@@ -274,25 +274,25 @@ if [[ ! -e "/usr/local/cuda/lib" ]]; then
 fi
 
 if ! grep 'try-import %workspace%/.local_cuda.bazelrc' "${SRC_PATH_JAX}/.bazelrc"; then
-  echo -e '\ntry-import %workspace%/.local_cuda.bazelrc' >> "${SRC_PATH_JAX}/.bazelrc"
+    echo -e '\ntry-import %workspace%/.local_cuda.bazelrc' >> "${SRC_PATH_JAX}/.bazelrc"
 fi
 cat > "${SRC_PATH_JAX}/.local_cuda.bazelrc" << EOF
-build:cuda --repo_env=LOCAL_CUDA_PATH="/usr/local/cuda"
-build:cuda --repo_env=LOCAL_CUDNN_PATH="/opt/nvidia/cudnn"
-build:cuda --repo_env=LOCAL_NCCL_PATH="/opt/nvidia/nccl"
+build --repo_env=LOCAL_CUDA_PATH="/usr/local/cuda"
+build --repo_env=LOCAL_CUDNN_PATH="/opt/nvidia/cudnn"
+build --repo_env=LOCAL_NCCL_PATH="/opt/nvidia/nccl"
 EOF
-time python "${SRC_PATH_JAX}/build/build.py" \
+
+pushd ${SRC_PATH_JAX}
+time python "${SRC_PATH_JAX}/build/build.py" build \
     --editable \
     --use_clang \
-    --enable_cuda \
-    --build_gpu_plugin \
-    --gpu_plugin_cuda_version=$TF_CUDA_MAJOR_VERSION \
+    --wheels=jaxlib,jax-cuda-plugin,jax-cuda-pjrt \
     --cuda_compute_capabilities=$TF_CUDA_COMPUTE_CAPABILITIES \
-    --enable_nccl=true \
     --bazel_options=--linkopt=-fuse-ld=lld \
-    --bazel_options=--override_repository=xla=$SRC_PATH_XLA \
+    --local_xla_path=$SRC_PATH_XLA \
     --output_path=${BUILD_PATH_JAXLIB} \
     $BUILD_PARAM
+popd
 
 # Make sure that JAX depends on the local jaxlib installation
 # https://jax.readthedocs.io/en/latest/developer.html#specifying-dependencies-on-local-wheels
@@ -300,8 +300,8 @@ line="jaxlib @ file://${BUILD_PATH_JAXLIB}/jaxlib"
 if ! grep -xF "${line}" "${SRC_PATH_JAX}/build/requirements.in"; then
     pushd "${SRC_PATH_JAX}"
     echo "${line}" >> build/requirements.in
-    echo "jax-cuda${TF_CUDA_MAJOR_VERSION}-pjrt @ file://${BUILD_PATH_JAXLIB}/jax_gpu_pjrt" >> build/requirements.in
-    echo "jax-cuda${TF_CUDA_MAJOR_VERSION}-plugin @ file://${BUILD_PATH_JAXLIB}/jax_gpu_plugin" >> build/requirements.in
+    echo "jax-cuda${TF_CUDA_MAJOR_VERSION}-pjrt @ file://${BUILD_PATH_JAXLIB}/jax-cuda-pjrt" >> build/requirements.in
+    echo "jax-cuda${TF_CUDA_MAJOR_VERSION}-plugin @ file://${BUILD_PATH_JAXLIB}/jax-cuda-plugin" >> build/requirements.in
     PYTHON_VERSION=$(python -c 'import sys; print("{}.{}".format(*sys.version_info[:2]))')
     bazel run --verbose_failures=true //build:requirements.update --repo_env=HERMETIC_PYTHON_VERSION="${PYTHON_VERSION}"
     popd
@@ -316,13 +316,13 @@ else
 fi
 
 # install jax and jaxlib
-pip --disable-pip-version-check install -e ${BUILD_PATH_JAXLIB}/jaxlib -e ${BUILD_PATH_JAXLIB}/jax_gpu_pjrt -e ${BUILD_PATH_JAXLIB}/jax_gpu_plugin -e "${SRC_PATH_JAX}"
+pip --disable-pip-version-check install -e ${BUILD_PATH_JAXLIB}/jaxlib -e ${BUILD_PATH_JAXLIB}/jax-cuda-pjrt -e ${BUILD_PATH_JAXLIB}/jax-cuda-plugin -e "${SRC_PATH_JAX}"
 
-# after installation (example)
-#  jax                     0.4.32.dev20240808+9c2caedab /opt/jax
-#  jax-cuda12-pjrt         0.4.32.dev20240808           /opt/jaxlibs/jax_gpu_pjrt
-#  jax-cuda12-plugin       0.4.32.dev20240808           /opt/jaxlibs/jax_gpu_plugin
-#  jaxlib                  0.4.32.dev20240808           /opt/jaxlibs/jaxlib
+## after installation (example)
+# jax               0.4.36.dev20241125+f828f2d7d /opt/jax
+# jax-cuda12-pjrt   0.4.36.dev20241125           /opt/jaxlibs/jax-cuda-pjrt
+# jax-cuda12-plugin 0.4.36.dev20241125           /opt/jaxlibs/jax-cuda-plugin
+# jaxlib            0.4.36.dev20241125           /opt/jaxlibs/jaxlib
 pip list | grep jax
 
 # Ensure directories are readable by all for non-root users
diff --git a/.github/container/install-nsight.sh b/.github/container/install-nsight.sh
index f3e4e0715..4aa001cf1 100755
--- a/.github/container/install-nsight.sh
+++ b/.github/container/install-nsight.sh
@@ -12,7 +12,7 @@ export DEBIAN_FRONTEND=noninteractive
 export TZ=America/Los_Angeles
 
 apt-get update
-apt-get install -y nsight-compute nsight-systems-cli
+apt-get install -y nsight-compute nsight-systems-cli-2024.6.1
 apt-get clean
 
 rm -rf /var/lib/apt/lists/*
diff --git a/.github/container/nsys-jax-combine b/.github/container/nsys-jax-combine
index 00e53efb6..36ef782af 100755
--- a/.github/container/nsys-jax-combine
+++ b/.github/container/nsys-jax-combine
@@ -192,8 +192,9 @@ with zipfile.ZipFile(args.output, "w") as ofile:
             )
             # Gather output files of the scrpt
             for path in working_dir.rglob("*"):
-                with open(working_dir / path, "rb") as src, ofile.open(
-                    str(path.relative_to(mirror_dir)), "w"
-                ) as dst:
+                with (
+                    open(working_dir / path, "rb") as src,
+                    ofile.open(str(path.relative_to(mirror_dir)), "w") as dst,
+                ):
                     # https://github.com/python/mypy/issues/15031 ?
                     shutil.copyfileobj(src, dst)  # type: ignore
diff --git a/.github/container/test-jax.sh b/.github/container/test-jax.sh
index d73ed6d7e..3a14c7a72 100755
--- a/.github/container/test-jax.sh
+++ b/.github/container/test-jax.sh
@@ -26,7 +26,7 @@ jax_source_dir() {
 
 query_tests() {
     cd `jax_source_dir`
-    python build/build.py --configure_only
+    python build/build.py build --wheels=jaxlib,jax-cuda-plugin,jax-cuda-pjrt --configure_only
     bazel query tests/... 2>&1 | grep -F '//tests:'
     exit
 }
@@ -113,8 +113,6 @@ for t in $*; do
 done
 
 TEST_TAG_FILTER_ARRAY=()
-TEST_TAG_FILTER_ARRAY+=('-multiaccelerator')
-
 COMMON_FLAGS=$(cat << EOF
 --@local_config_cuda//:enable_cuda
 --cache_test_results=${CACHE_TEST_RESULTS}
@@ -163,7 +161,10 @@ case "${BATTERY}" in
         ;;
 esac
 
-TEST_TAG_FILTERS=$(IFS=, ; echo "--test_tag_filters=${TEST_TAG_FILTER_ARRAY[*]}")
+TEST_TAG_FILTERS=""
+if [[ ${#TEST_TAG_FILTER_ARRAY[@]} > 0 ]]; then
+    TEST_TAG_FILTERS=$(IFS=, ; echo "--test_tag_filters=${TEST_TAG_FILTER_ARRAY[*]}")
+fi
 
 print_var NCPUS
 print_var NGPUS
@@ -190,5 +191,5 @@ pip install matplotlib
 ## Run tests
 
 cd `jax_source_dir`
-python build/build.py --configure_only
+python build/build.py build --wheels=jaxlib,jax-cuda-plugin,jax-cuda-pjrt --configure_only
 bazel test ${BAZEL_TARGET} ${TEST_TAG_FILTERS} ${COMMON_FLAGS} ${EXTRA_FLAGS}
diff --git a/.github/workflows/_ci.yaml b/.github/workflows/_ci.yaml
index 2f77c07f4..4816c709e 100644
--- a/.github/workflows/_ci.yaml
+++ b/.github/workflows/_ci.yaml
@@ -295,6 +295,7 @@ jobs:
         docker run -i --shm-size=1g --gpus all \
         ${{ needs.build-jax.outputs.DOCKER_TAG_FINAL }} \
         bash <<"EOF" |& tee tee test-gpu.log
+          nvidia-cuda-mps-control -d
           test-jax.sh -b gpu
         EOF
       STATISTICS_SCRIPT: |
@@ -463,8 +464,9 @@ jobs:
         docker run -i --shm-size=1g --gpus all --volume $PWD:/output \
         ${{ needs.build-triton.outputs.DOCKER_TAG_FINAL }} \
         bash <<"EOF" |& tee test-triton.log
-          # autotuner tests from jax-triton now hit a triton code path that uses utilities from pytorch...
-          pip install --no-deps torch --index-url https://download.pytorch.org/whl/cpu
+          # autotuner tests from jax-triton now hit a triton code path that uses utilities from pytorch; this relies on
+          # actually having a CUDA backend for pytoch
+          pip install --no-deps torch
           python /opt/jax-triton/tests/triton_call_test.py --xml_output_file /output/triton_test.xml
         EOF
       STATISTICS_SCRIPT: |