python - 控制 dask 中的核心/线程数量

Question

我有一个具有以下规范的工作站:

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
Address sizes:       46 bits physical, 48 bits virtual
CPU(s):              16
On-line CPU(s) list: 0-15
Thread(s) per core:  2
Core(s) per socket:  8
Socket(s):           1
NUMA node(s):        1
Vendor ID:           GenuineIntel
CPU family:          6
Model:               79
Model name:          Intel(R) Xeon(R) CPU E5-1660 v4 @ 3.20GHz
Stepping:            1
CPU MHz:             1200.049
CPU max MHz:         3800.0000
CPU min MHz:         1200.0000
BogoMIPS:            6400.08
Virtualization:      VT-x
L1d cache:           32K
L1i cache:           32K
L2 cache:            256K
L3 cache:            20480K
NUMA node0 CPU(s):   0-15
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm rdt_a rdseed adx smap intel_pt xsaveopt cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts flush_l1d

我已经实现了 dask 来分发一些计算，并且我以这种方式设置 Client():

if __name__ == '__main__':
    cluster = LocalCluster()
    client = Client(cluster, asyncronous=True, n_workers=8,
                    threads_per_worker=2)
    train()

当我使用 dask.compute(*computations, Scheduler='distributed') 调用我的延迟函数时，dask 似乎正在使用所有资源。仪表板如下所示:

现在，如果我继续将 Client() 更改为:

if __name__ == '__main__':
    cluster = LocalCluster()
    client = Client(cluster, asyncronous=True, n_workers=4,
                    threads_per_worker=2)
    train()

我预计会使用一半的资源，但正如您在我的仪表板上看到的那样，情况并非如此。

为什么 dask Client() 仍在使用所有资源？我将不胜感激任何对此的意见。

Answer 1

如果您尚未指定集群，Client 类将为您创建一个集群。这些关键字仅在不传递现有集群实例时才有效。您应该将它们放入对 LocalCluster 的调用中:

cluster = LocalCluster(n_workers=4, threads_per_worker=2)
client = Client(cluster, asynchronous=True)

或者您可以直接跳过创建集群

client = Client(asynchronous=True, n_workers=4, threads_per_worker=2)