Merge branch 'user/yibinl/fix_build' into 'release/1.1.1-25.12-NIM'

Chery-pick: [https://nvbugs/5655885][fix] fix invalid instruction error in 2shot ar kernel on Ampere (#9394)

See merge request ftp/tekit!9896

Signed-off-by: Yibin Li <yibinl@nvidia.com>

Author: tijyojwad

cpp/tensorrt_llm/kernels/communicationKernels/allReduceFusionKernels.cu

@@ -134,11 +134,17 @@ public:
             // corresponding CTA has not been launched.
             for (int flag_idx = blockIdx.x; flag_idx < kBarrierFlagCount; flag_idx += gridDim.x)
             {
+#if (defined(__CUDA_ARCH__) && (__CUDA_ARCH__ >= 900))
                 asm volatile(
                     "st.global.relaxed.sys.b32 [%1], %0;" ::"r"(m_flag_value), "l"(m_target_flag + flag_idx * NRanks));
+#else
+                st_flag(m_target_flag + flag_idx * NRanks, m_flag_value);
+#endif
             }
+#if (defined(__CUDA_ARCH__) && (__CUDA_ARCH__ >= 900))
             // Single release fence
             asm volatile("fence.release.sys;");
+#endif
 
             while (ld_flag(m_current_flag) == prev_flag(m_flag_value))
             {

jenkins/Build.groovy

@@ -62,22 +62,22 @@ def BUILD_CONFIGS = [
   (CONFIG_LINUX_X86_64_VANILLA) : [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars ENABLE_MULTI_DEVICE=1 --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars NIXL_ROOT=/opt/nvidia/nvda_nixl --micro_benchmarks",
     (TARNAME) : "TensorRT-LLM.tar.gz",
-    (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
+    (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
   ],
   (CONFIG_LINUX_X86_64_PYBIND) : [
     (WHEEL_EXTRA_ARGS) : "--binding_type pybind --extra-cmake-vars ENABLE_MULTI_DEVICE=1 --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars NIXL_ROOT=/opt/nvidia/nvda_nixl --micro_benchmarks",
     (TARNAME) : "pybind-TensorRT-LLM.tar.gz",
-    (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
+    (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
   ],
   (CONFIG_LINUX_X86_64_SINGLE_DEVICE) : [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars ENABLE_MULTI_DEVICE=0 --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars ENABLE_UCX=0 --micro_benchmarks",
     (TARNAME) : "single-device-TensorRT-LLM.tar.gz",
-    (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
+    (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
   ],
   (CONFIG_LINUX_X86_64_LLVM) : [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars ENABLE_MULTI_DEVICE=1 --extra-cmake-vars WARNING_IS_ERROR=ON --micro_benchmarks -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_CUDA_HOST_COMPILER=clang -DCMAKE_LINKER_TYPE=LLD",
     (TARNAME) : "llvm-TensorRT-LLM.tar.gz",
-    (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
+    (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
   ],
   (CONFIG_LINUX_AARCH64): [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars NIXL_ROOT=/opt/nvidia/nvda_nixl",
@@ -504,7 +504,7 @@ def buildWheelInContainer(pipeline, libraries=[], triple=X86_64_TRIPLE, clean=fa
         if (triple == AARCH64_TRIPLE) {
             extra_args = "-a '90-real;100-real;103-real;120-real'"
         } else {
-            extra_args = "-a '90-real;100-real;103-real;120-real'"
+            extra_args = "-a '80-real;86-real;89-real;90-real;100-real;103-real;120-real'"
         }
     }
     if (pre_cxx11abi) {