SSRegression在两方情况下，每一方有多个物理计算节点，如何提高运行速度

[kmoker8s]

Issue Type

Bug

Source

binary

Secretflow Version

1.5.0.dev20240321

OS Platform and Distribution

CentOS Linux release 7.9.2009 (Core)

Python version

3.9

Bazel version

none

GCC/Compiler version

none

What happend and What you expected to happen.

在两方参与计算时，每一方都拥有多个物理计算节点。具体的用的secretflow的SSRegression来做逻辑回归，在这种配置下，希望提供配置或解决方案提高整个系统的运行速度

Reproduction code to reproduce the issue.

sf.shutdown()
sf.init(parties=['alice', 'bob'],address=alice_ip+':9394')
#sf.init(address='192.168.207.221:9394', cluster_config=cluster_config)
alice = sf.PYU('alice')
bob = sf.PYU('bob')
#carol = sf.PYU('carol')

spu_config = {
            'nodes': [
                {'party': 'alice', 'id': 'local:0', 'address': alice_ip + ':12945'},
                {'party': 'bob', 'id': 'local:1', 'address': bob_ip + ':12946'},
                # {'party': 'carol', 'id': 'local:2', 'address': '127.0.0.1:12347'},
            ],
            'runtime_config': {
                # SEMI2K support 2/3 PC, ABY3 only support 3PC, CHEETAH only support 2PC.
                # pls pay attention to size of nodes above. nodes size need match to PC setting.
                'protocol': spu.spu_pb2.SEMI2K,
                'field': spu.spu_pb2.FM128,
            },
        }
# SPU settings

my_spu = sf.SPU(spu_config)


# your code to run.
# init log

#logging.basicConfig(stream=sys.stdout, level=logging.INFO)    
#start = time.time()
timecollect.start()

train_vdf = v_read_csv(
    {alice: train_alice_path, bob: train_bob_path},
    keys="row_num",
    drop_keys="row_num",
    spu=my_spu,
    psi_protocl="ECDH_PSI_2PC"
)

test_vdf = v_read_csv(
    {alice: test_alice_path, bob: test_bob_path},
    keys="row_num",
    drop_keys="row_num",
    spu=my_spu,
    psi_protocl="ECDH_PSI_2PC"
)
# 初始化模型
lr_model = SSRegression(my_spu)

# 训练模型
lr_model.fit(
    x=X_train,
    y=y_train,
    epochs=epoch,
    learning_rate=learning_rate,
    batch_size=batch_size,
    sig_type='t3',
    reg_type='logistic',
    penalty='l2',
    l2_norm=l2_r,
    eps=0.0001
)

[da-niao-dan]

SS Regression 的计算瓶颈在于网络而不是计算资源。8C单台机器的CPU也利用不满。
没有直接配置物理机机群而加速SS Regression的方法。