Re: [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization

To: David Rientjes <rientjes@xxxxxxxxxx>
Subject: Re: [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization
From: Gang Li <gang.li@xxxxxxxxx>
Date: Thu, 25 Jan 2024 11:47:04 +0800
Cc: Linux-MM <linux-mm@xxxxxxxxx>, LKML <linux-kernel@xxxxxxxxxxxxxxx>, ligang.bdlg@xxxxxxxxxxxxx, David Hildenbrand <david@xxxxxxxxxx>, Mike Kravetz <mike.kravetz@xxxxxxxxxx>, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>, Muchun Song <muchun.song@xxxxxxxxx>, Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>
In-reply-to: <15E775EC-8E4B-4145-8931-067DF60E55F1@linux.dev>
References: <20240118123911.88833-1-gang.li@linux.dev> <20240118123911.88833-8-gang.li@linux.dev> <da1258e3-f828-4bbc-a2c2-8fe1ef808c9a@linux.dev> <ef5b09e9-8cfb-4e74-89d4-5ebde12361cf@linux.dev> <15E775EC-8E4B-4145-8931-067DF60E55F1@linux.dev>

On 2024/1/25 10:48, Muchun Song wrote:>>>> 12t 1G77000 18300 76.23%


I am curious how many NUMA nodes does this system have? I suspect
it should not be one.


Hi David Rientjes,

I'm also curious. Can you show us some details about this machine?
Such as `lscpu`.

Thanks!

References:
- [RESEND PATCH v4 0/7] hugetlb: parallelize hugetlb page init on boot
  - From: Gang Li
- [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization
  - From: Gang Li
- Re: [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization
  - From: Muchun Song
- Re: [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization
  - From: Gang Li
- Re: [PATCH v4 7/7] hugetlb: parallelize 1G hugetlb initialization
  - From: Muchun Song