CPU util wouldn't go above 40% on numa node1

[montagetao]

Here is the command I'm using to start the memcached instance:

numactl -N 0 --membind=0 memcached -p 11211 -l 127.0.0.1 -t 128 -u nobody -m 1000000 -c 9999 -d

numactl --physcpubind=64-127 --membind=1
memtier_benchmark -p 11211 --protocol=memcache_text
--threads=64 --clients=24 --key-maximum=800000000
--requests=500000 --ratio=1:0 --data-size-range=128-2048
--key-pattern=P:P --data-size-pattern=R

numactl --physcpubind=64-127 --membind=1
memtier_benchmark -p 11211 --protocol=memcache_text
--threads=64 --clients=24 --key-maximum=800000000
--requests=50000 --ratio=1:10 --data-size-range=128-2048
--key-pattern=R:R --data-size-pattern=R

test config:
GNR CPU6767P *2 [NUMA node0 1024GB+NUMA node1 1024GB]
each CPU has 64cores and 128T
OS: Ubunut 24.04.LTS
Memcached: 1.6.24
Memtier_benchmark:2.1.4

Here’s a summary of the issues I encountered:
CPU Utilization: On the NUMA node1, during the DB loading phase, I observed that only 12 cores were actively utilized. When running the read/write tests, the CPU utilization increased slightly, with 18 cores being engaged. This was far from the 100% CPU（64 cores）utilization

could you give some comments on it, thanks

[montagetao]

i found when add the "-t" in /etc/memcached.conf , then restart the memcached service
i see the cpu bottlneck on NUMA node1 when load DB(R:W 0:1), i think its the root cause that why the cpu utilization keep low