[llvm] [VPlan] Use VPWidenIntrinsicRecipe to support binary and unary operations with EVL-vectorization (PR #114205)

[via llvm-commits]

LiqinWeng wrote:

> DCMAKE_CXX_FLAGS
This is the command I used to run the spec 2017 test：
1. cmake .. -GNinja -DCMAKE_C_COMPILER=/home/wengliqin/WorkSpace/llvm-project/build/bin/clang -DCMAKE_CXX_COMPILER=/home/wenn/clang++ -DTEST_SUITE_SUBDIRS=External -DTEST_SUITE_BENCHMARKING_ONLY=ON  -DTEST_SUITE_RUN_SPEC_CFP=OFF  -DTEST_SUITE_RUN_TYPE=train -DTEST_SUITE_SPEC2017_ROOT=/home/wengliqin/WorkSpace/speccpu2017 -DTEST_SUITE_REMOTE_HOST=X60 -DTEST_SUITE_DIAGNOSE_FLAGS="--target=riscv64-unknown-linux-gnu -march=rv64imav -O3 -mllvm -force-tail-folding-style=data-with-evl -mllvm -prefer-predicate-over-epilogue=predicate-dont-vectorize --sysroot=/home/wengliqin/WorkSpace/tool/gcc-linux/sysroot --gcc-toolchain=/home/wengliqin/WorkSpace/tool/gcc-linux/"
2. ninja -j12 
3. ninja rsync
4. llvm-lit  -j8 -v .

I have some questions to ask（use spacemit-x60 verified）:
1. Are you using the train test ?
2. How can I print the stack for failed test cases? (Currently, my local binary packages of clang are all debug versions. In theory, if there is a problem during runing, the stack will be printed.)

The test that is currently hanging does not obtain more accurate information. Is it because there is no option added or for some other reason? Do you know?
![image](https://github.com/user-attachments/assets/b8a9e245-8180-410f-93c7-b8b2a2a6163c)
![image](https://github.com/user-attachments/assets/f921199b-1cc1-4267-af50-f10730c0c794)



https://github.com/llvm/llvm-project/pull/114205