[FA] Remove TRITON_INTEL_ENABLE_INSTR_SCHED (#2578)

A 5% geomean improvement is observed on FA advanced path with
`TRITON_INTEL_ENABLE_INSTR_SCHED` removed.

CI:
https://github.com/intel/intel-xpu-backend-for-triton/actions/runs/11541276782,
https://github.com/intel/intel-xpu-backend-for-triton/actions/runs/11540829777

---------

Signed-off-by: Whitney Tsang <[email protected]>

Author: whitneywhtsang

.github/workflows/triton-benchmarks.yml

@@ -253,7 +253,6 @@ jobs:
         run: |
           cd benchmarks/triton_kernels_benchmark
           TRITON_INTEL_ADVANCED_PATH=1 \
-          TRITON_INTEL_ENABLE_INSTR_SCHED=1 \
           TRITON_INTEL_ENABLE_ADDRESS_PAYLOAD_OPT=1 \
           IGC_VISAOptions=" -enableBCR" \
           python flash_attention_fwd_benchmark.py --reports $REPORTS

scripts/test-triton.sh

@@ -285,7 +285,6 @@ run_benchmark_attention() {
   echo "Advanced path:"
   TRITON_INTEL_ADVANCED_PATH=1 \
     TRITON_INTEL_ENABLE_ADDRESS_PAYLOAD_OPT=1 \
-    TRITON_INTEL_ENABLE_INSTR_SCHED=1 \
     IGC_VISAOptions=" -enableBCR" \
     python $TRITON_PROJ/benchmarks/triton_kernels_benchmark/flash_attention_fwd_benchmark.py
 }