update bert config (tinygrad#9555)

BEAM 4->5 for green, 2% faster
use AMD driver instead of AM for red, 5% faster

Author: chenyuxyz

examples/mlperf/training_submission_v5.0/tinycorp/benchmarks/bert/implementations/tinybox_green/run_and_time.sh

@@ -5,7 +5,7 @@ export MODEL="bert"
 export SUBMISSION_PLATFORM="tinybox_green"
 export DEFAULT_FLOAT="HALF" SUM_DTYPE="HALF" GPUS=6 BS=96 EVAL_BS=96
 
-export BEAM=4 BEAM_UOPS_MAX=3000 BEAM_UPCAST_MAX=256 BEAM_LOCAL_MAX=1024 BEAM_MIN_PROGRESS=5
+export BEAM=5 BEAM_UOPS_MAX=3000 BEAM_UPCAST_MAX=256 BEAM_LOCAL_MAX=1024 BEAM_MIN_PROGRESS=5
 export IGNORE_JIT_FIRST_BEAM=1
 export BASEDIR="/raid/datasets/wiki"
 

examples/mlperf/training_submission_v5.0/tinycorp/benchmarks/bert/implementations/tinybox_red/run_and_time.sh

@@ -17,7 +17,11 @@ DATETIME=$(date "+%m%d%H%M")
 LOGFILE="bert_red_${DATETIME}_${SEED}.log"
 
 # init
+sudo rmmod amdgpu || true
 BENCHMARK=10 INITMLPERF=1 RESET_STEP=1 BEAM_LOG_SURPASS_MAX=1 python3 examples/mlperf/model_train.py | tee $LOGFILE
 
 # run
+# TODO: AMD driver hangs during init, but is 5% faster per step in real run.
+sudo modprobe amdgpu
 PARALLEL=0 RUNMLPERF=1 python3 examples/mlperf/model_train.py | tee -a $LOGFILE
+sudo rmmod amdgpu || true