Simpoint checkpoint reload problem

[badgate503]

Hey guys. I use the following script to generate Simpoint checkpoints:

build/${ARCH}/gem5.${METHOD} --outdir=${SE_OUT_DIR_CHECKPOINT} \
configs/deprecated/example/se.py \
--take-simpoint-checkpoint=\
${SE_simpoint_file_path},${SE_weight_file_path},\
${interval_length},${warmup_length} \
--cpu-type=AtomicSimpleCPU \
${CACHE_PRAM} \
${MEM_PRAM} \
-c ${SE_ELF_ROUTE} \
-o "${SE_INPUT_ROUTE}${OPTIONS}"

the interval_length and warmup_length are 50M and 250M respectively
And it just so happens that one of the checkpoints given by Simpoint start from Interval 0, making the warmup phase completely truncated, logically speaking, the checkpoint should still run one interval. but when I reload the checkpoints with following script:

build/${ARCH}/gem5.${METHOD} --outdir=${SE_OUT_DIR_RELOAD}/$1 \
configs/deprecated/example/se.py \
${CPU_PARM} \
${CACHE_PRAM} \
${MEM_PRAM} \
-c ${SE_ELF_ROUTE} \
-o "${SE_INPUT_ROUTE}${OPTIONS}" \
--restore-simpoint-checkpoint -r $1 --checkpoint-dir ${SE_OUT_DIR_CHECKPOINT} 

the checkpoint which start from interval 0 just terminates after 10000 ticks, instead of running one interval. while other checkpoints ran as expected

info: Standard input is not a terminal, disabling listeners.
Resuming from m5out/bfs/bfs_checkpoint/cpt.simpoint_02_inst_0_weight_0.009901_interval_50000000_warmup_0
Resuming from SimPoint #2, start_inst:0, weight:0.009901, interval:50000000, warmup:0
Switch at curTick count:10000
Switched CPUS @ tick 500
switching cpus
Exiting @ tick 10000 because simulate() limit reached

I really can't figure out this on myself. and I searched online but found nothing helpful