[Bug] Compilation failed during make -j

[qluoluo]

python 3.10.19
CUDA 12.8
g++ (Ubuntu 11.4.0-1ubuntu1~22.04.2) 11.4.0
pytorch 2.5.0

I cloned the flashinfer during the command 'git clone --recurse-submodules https://github.com/mit-han-lab/quest', but I found Compilation failed during command 'make -j':

[ 1%] Generate git revision file for nvbench_git_revision
[ 3%] Building CXX object 3rdparty/googletest/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o
[ 3%] Built target nvbench_git_revision_compute_git_info
[ 6%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/axis_base.cxx.o
[ 7%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/axes_metadata.cxx.o
[ 7%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/benchmark_base.cxx.o
[ 9%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/benchmark_manager.cxx.o
[ 11%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/blocking_kernel.cu.o
[ 12%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/criterion_manager.cxx.o
[ 14%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/float64_axis.cxx.o
[ 15%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/cuda_call.cu.o
[ 17%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/csv_printer.cu.o
[ 20%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/runner.cxx.o
[ 20%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/device_manager.cu.o
[ 22%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/int64_axis.cxx.o
[ 23%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/device_info.cu.o
[ 26%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/internal/nvml.cxx.o
[ 26%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/type_strings.cxx.o
[ 28%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/stopping_criterion.cxx.o
[ 30%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/state.cxx.o
[ 33%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/cupti_profiler.cxx.o
[ 33%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/printer_base.cxx.o
[ 34%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/markdown_printer.cu.o
[ 36%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/printer_multiplex.cxx.o
[ 38%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/json_printer.cu.o
[ 42%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/string_axis.cxx.o
[ 42%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/measure_hot.cu.o
[ 42%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/state_generator.cxx.o
[ 44%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/entropy_criterion.cxx.o
[ 46%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/stdrel_criterion.cxx.o
[ 47%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/measure_cupti.cu.o
[ 49%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/named_values.cxx.o
[ 52%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/detail/measure_cold.cu.o
[ 53%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/option_parser.cu.o
[ 53%] Building CXX object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.dir/type_axis.cxx.o
[ 55%] Linking CXX shared library ../../../lib/libgtest.so
[ 55%] Built target gtest
[ 57%] Building CXX object 3rdparty/googletest/googlemock/CMakeFiles/gmock.dir/src/gmock-all.cc.o
[ 58%] Building CXX object 3rdparty/googletest/googletest/CMakeFiles/gtest_main.dir/src/gtest_main.cc.o
[ 60%] Linking CXX shared library ../../../lib/libgtest_main.so
[ 60%] Built target gtest_main
[ 61%] Building CUDA object CMakeFiles/test_batch_decode.dir/src/test/test_batch_decode.cu.o
[ 63%] Building CUDA object CMakeFiles/test_prefill.dir/src/test/test_prefill.cu.o
[ 65%] Building CUDA object CMakeFiles/test_max_possible.dir/src/test/test_max_possible.cu.o
[ 66%] Building CUDA object CMakeFiles/test_page.dir/src/test/test_page.cu.o
[ 68%] Linking CXX shared library ../../../lib/libnvbench.so
[ 68%] Built target nvbench
[ 69%] Building CUDA object 3rdparty/nvbench/exec/CMakeFiles/nvbench.ctl.dir/nvbench-ctl.cu.o
[ 71%] Building CUDA object 3rdparty/nvbench/nvbench/CMakeFiles/nvbench.main.dir/main.cu.o
[ 73%] Linking CXX shared library ../../../lib/libgmock.so
[ 73%] Built target gmock
[ 74%] Building CXX object 3rdparty/googletest/googlemock/CMakeFiles/gmock_main.dir/src/gmock_main.cc.o
[ 76%] Linking CUDA executable ../../../bin/nvbench-ctl
[ 76%] Built target nvbench.ctl
[ 77%] Linking CXX shared library ../../../lib/libgmock_main.so
[ 77%] Built target gmock_main
[ 77%] Built target nvbench.main
[ 79%] Building CUDA object CMakeFiles/bench_batch_decode.dir/src/bench/bench_batch_decode.cu.o
[ 80%] Building CUDA object CMakeFiles/bench_prefill.dir/src/bench/bench_prefill.cu.o
[ 82%] Building CUDA object CMakeFiles/bench_decode_select_k.dir/src/bench/bench_decode_select_k.cu.o
[ 84%] Building CUDA object CMakeFiles/bench_page.dir/src/bench/bench_page.cu.o
[ 85%] Building CUDA object CMakeFiles/bench_max_possible.dir/src/bench/bench_max_possible.cu.o
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h: In function 'std::vector<dtype_out> cpu_reference::selected_single_mha(const std::vector&, const std::vector&, const std::vector&, const std::vector&, const size_t&, size_t, size_t, size_t, size_t, size_t, size_t, bool, flashinfer::QKVLayout, flashinfer::RotaryMode, float, float)':
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:383: error: wrong number of template arguments (2, should be 3)
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/3rdparty/flashinfer/include/flashinfer/layout.cuh:55:8: note: provided for 'template<flashinfer::QKVLayout kv_layout, unsigned int group_size, unsigned int head_dim> struct flashinfer::tensor_info_t'
55 | struct tensor_info_t {
| ^~~~~~~~~~~~~
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:419: error: expression list treated as compound expression in initializer [-fpermissive]
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:473: error: request for member 'get_num_qo_heads' in 'info', which is of non-class type 'int'
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:1410: error: request for member 'get_qo_elem_offset' in 'info', which is of non-class type 'int'
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:1480: error: request for member 'get_kv_elem_offset' in 'info', which is of non-class type 'int'
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:963: error: request for member 'get_kv_elem_offset' in 'info', which is of non-class type 'int'
193 | SWITCH_GQA_GROUP_SIZE(
| ^
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:16: error: request for member 'get_qo_elem_offset' in 'info', which is of non-class type 'int'
193 | SWITCH_GQA_GROUP_SIZE(
| ^~~~~~~~~~~~~~~
/inspire/hdd/project/exploration-topic/liuzhigeng-253108120105/projects/ffa/quest/kernels/src/include/cpu_reference.h:193:280: error: wrong number of template arguments (2, should be 3)
193 | SWITCH_GQA_GROUP_SIZE(
| ,

Could you please specify the recommended version of flashinfer for this project or suggest a patch?
Thank you!