[Pw_forum] why k point parallelization -npool is so slow?

[Sitangshu Bhattacharya]

Dear Balabi,

I am having Intel E5 2640 v4 processors @2.4GHz arranged in 6 nodes (1
Master + 5 Slaves). Each node has 20 cores (2 physical CPUs 10 core each)
hyperthreaded to give 40 CPUs. The RAM is 323 GB per node. However, I am
having a weak switch connectivity ~ 1GBPS

As far as I see I have a good scaling in one node. However if I distribute
over several other nodes, it gets stuck, probably because of the weak
switch.

Regards,


On Wed, Sep 20, 2017 at 10:43 AM, balabi <balabi at qq.com> wrote:

> More information. I done all previous test on single node with
>
>
>    - CPU: 2*Intel Xeon E5-2682v4 (16 Cores, 40M Cache, 2.50 GHz, 9.6
>    GT/s QPI)
>    - RAM: 64 GByte, 4 Channels, DDR4-2400 ECC RDIMM, Bandwidth: 153.6 GB/s
>    - Network: Gigabit Ethernet, 100 Gbit/s Intel Omni-Path Fabric
>
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>



-- 
**********************************************
Sitangshu Bhattacharya (সিতাংশু ভট্টাচার্য), Ph.D
Assistant Professor,
Room No. 2221, CC-1,
Nanoscale Electro-Thermal Laboratory,
Department of Electrical and Communication Engineering,
Indian Institute of Information Technology-Allahabad
Uttar Pradesh 211 012
India
Telephone: 91-532-2922000 Extn.: 2131
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20170920/ba627093/attachment.html>