From: Andi Kleen <ak@suse.de>
To: clameter@sgi.com
Cc: akpm@linux-foundation.org, travis@sgi.com,
Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>,
linux-kernel@vger.kernel.org
Subject: Re: [rfc 08/45] cpu alloc: x86 support
Date: Tue, 20 Nov 2007 04:16:32 +0100 [thread overview]
Message-ID: <200711200416.32954.ak@suse.de> (raw)
In-Reply-To: <20071120011333.619991903@sgi.com>
> 4k cpu configurations with 1k nodes:
>
> 4096 * 16MB = 64TB of virtual space.
>
> Maximum theoretical configuration 16384 processors 1k nodes:
>
> 16384 * 16MB = 256TB of virtual space.
>
> Both fit within the established limits established.
I might be pointing out the obvious, but on x86-64 there is definitely not
256TB of VM available for this.
Not even 64TB, as long as you want to have any
other mappings in kernel (total kernel memory 128TB, but it is split in
half for the direct mapping)
BTW if you allocate any VM you should also update
Documentation/x86_64/mm.txt which describes the mapping
> Index: linux-2.6/include/asm-x86/pgtable_64.h
> ===================================================================
> --- linux-2.6.orig/include/asm-x86/pgtable_64.h 2007-11-19
> 15:45:07.638390147 -0800 +++
> linux-2.6/include/asm-x86/pgtable_64.h 2007-11-19 15:55:53.165640248 -0800
> @@ -138,6 +138,7 @@ static inline pte_t ptep_get_and_clear_f
> #define VMALLOC_START _AC(0xffffc20000000000, UL)
> #define VMALLOC_END _AC(0xffffe1ffffffffff, UL)
> #define VMEMMAP_START _AC(0xffffe20000000000, UL)
> +#define CPU_AREA_BASE _AC(0xffffffff84000000, UL)
That's slightly less than 1GB before you bump into the maximum.
But you'll bump into the module mapping even before that.
For 16MB/CPU and the full 1GB that's ~123 CPUs if my calculations are correct.
Even for a not Altix that's quite tight.
I suppose 16MB/CPU are too large.
-Andi
next prev parent reply other threads:[~2007-11-20 3:20 UTC|newest]
Thread overview: 120+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-11-20 1:11 [rfc 00/45] [RFC] CPU ops and a rework of per cpu data handling on x86_64 clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 01/45] ACPI: Avoid references to impossible processors clameter, Christoph Lameter
2007-11-20 12:47 ` Mathieu Desnoyers
2007-11-20 20:16 ` Christoph Lameter
2007-11-20 15:29 ` Andi Kleen
2007-11-20 20:18 ` Christoph Lameter
2007-11-20 1:11 ` [rfc 02/45] cpu alloc: Simple version of the allocator (static allocations) clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 03/45] Generic CPU operations: Core piece clameter, Christoph Lameter
2007-11-20 3:17 ` Mathieu Desnoyers
2007-11-20 3:30 ` Christoph Lameter
2007-11-20 4:07 ` Mathieu Desnoyers
2007-11-20 20:36 ` Christoph Lameter
2007-11-20 1:11 ` [rfc 04/45] cpu alloc: Use in SLUB clameter, Christoph Lameter
2007-11-20 12:42 ` Mathieu Desnoyers
2007-11-20 20:44 ` Christoph Lameter
2007-11-20 21:23 ` Mathieu Desnoyers
2007-11-20 21:36 ` Christoph Lameter
2007-11-20 21:43 ` Mathieu Desnoyers
2007-11-20 1:11 ` [rfc 05/45] cpu alloc: Remove SLUB fields clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 06/45] cpu alloc: page allocator conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 07/45] cpu_alloc: Implement dynamically extendable cpu areas clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 08/45] cpu alloc: x86 support clameter, Christoph Lameter
2007-11-20 1:35 ` H. Peter Anvin
2007-11-20 2:02 ` Christoph Lameter
2007-11-20 2:18 ` H. Peter Anvin
2007-11-20 3:37 ` Nick Piggin
2007-11-20 3:59 ` Nick Piggin
2007-11-20 12:05 ` Andi Kleen
2007-11-20 3:16 ` Andi Kleen [this message]
2007-11-20 3:50 ` Christoph Lameter
2007-11-20 12:01 ` Andi Kleen
2007-11-20 20:35 ` Christoph Lameter
2007-11-20 20:59 ` Andi Kleen
2007-11-20 21:33 ` Christoph Lameter
2007-11-21 0:10 ` Christoph Lameter
2007-11-21 1:16 ` Christoph Lameter
2007-11-21 1:36 ` Andi Kleen
2007-11-21 2:08 ` Christoph Lameter
2007-11-21 13:08 ` Andi Kleen
2007-11-21 19:01 ` Christoph Lameter
2007-11-20 20:43 ` H. Peter Anvin
2007-11-20 20:51 ` Andi Kleen
2007-11-20 20:58 ` Christoph Lameter
2007-11-20 21:06 ` H. Peter Anvin
2007-11-20 21:34 ` Christoph Lameter
2007-11-20 21:01 ` H. Peter Anvin
2007-11-27 4:12 ` John Richard Moser
2007-11-20 1:11 ` [rfc 09/45] cpu alloc: IA64 support clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 10/45] cpu_alloc: Sparc64 support clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 11/45] cpu alloc: percpu_counter conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 12/45] cpu alloc: crash_notes conversion clameter, Christoph Lameter
2007-11-20 13:03 ` Mathieu Desnoyers
2007-11-20 20:50 ` Christoph Lameter
2007-11-20 1:11 ` [rfc 13/45] cpu alloc: workqueue conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 14/45] cpu alloc: ACPI cstate handling conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 15/45] cpu alloc: genhd statistics conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 16/45] cpu alloc: blktrace conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 17/45] cpu alloc: SRCU clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 18/45] cpu alloc: XFS counters clameter, Christoph Lameter
2007-11-20 8:12 ` Christoph Hellwig
2007-11-20 20:38 ` Christoph Lameter
2007-11-21 4:47 ` David Chinner
2007-11-21 4:50 ` Christoph Lameter
2007-11-20 1:11 ` [rfc 19/45] cpu alloc: NFS statistics clameter, Christoph Lameter
2007-11-20 13:02 ` Mathieu Desnoyers
2007-11-20 20:49 ` Christoph Lameter
2007-11-20 20:56 ` Trond Myklebust
2007-11-20 21:28 ` Mathieu Desnoyers
2007-11-20 21:48 ` Trond Myklebust
2007-11-20 21:50 ` Mathieu Desnoyers
2007-11-20 22:46 ` Trond Myklebust
2007-11-21 0:53 ` Mathieu Desnoyers
2007-11-20 21:26 ` Mathieu Desnoyers
2007-11-20 1:11 ` [rfc 20/45] cpu alloc: neigbour statistics clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 21/45] cpu alloc: tcp statistics clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 22/45] cpu alloc: convert scatches clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 23/45] cpu alloc: dmaengine conversion clameter, Christoph Lameter
2007-11-20 12:50 ` Mathieu Desnoyers
2007-11-20 20:46 ` Christoph Lameter
2007-11-20 1:11 ` [rfc 24/45] cpu alloc: convert loopback statistics clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 25/45] cpu alloc: veth conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 26/45] cpu alloc: Chelsio statistics conversion clameter, Christoph Lameter
2007-11-20 1:11 ` [rfc 27/45] cpu alloc: convert mib handling to cpu alloc clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 28/45] cpu_alloc: convert network sockets clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 29/45] cpu alloc: Use for infiniband clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 30/45] cpu alloc: Use in the crypto subsystem clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 31/45] cpu alloc: Remove the allocpercpu functionality clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 32/45] Module handling: Use CPU_xx ops to dynamically allocate counters clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 33/45] x86_64: Use CPU ops for nmi alert counter clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 34/45] x86_64: Fold percpu area into the cpu area clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 35/45] X86_64: Declare pda as per cpu data thereby moving it " clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 36/45] X86_64: Place pda first in " clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 37/45] x86_64: Support for fast per cpu operations clameter, Christoph Lameter
2007-11-20 2:00 ` H. Peter Anvin
2007-11-20 2:03 ` Christoph Lameter
2007-11-20 2:15 ` H. Peter Anvin
2007-11-20 2:17 ` David Miller
2007-11-20 2:19 ` H. Peter Anvin
2007-11-20 3:23 ` Andi Kleen
2007-11-20 2:45 ` Paul Mackerras
2007-11-20 1:12 ` [rfc 38/45] x86_64: Remove obsolete per_cpu offset calculations clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 39/45] x86_64: Remove the data_offset field from the pda clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 40/45] x86_64: Provide per_cpu_var definition clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 41/45] VM statistics: Use CPU ops clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 43/45] x86_64: Add a CPU_OR to support or_pda() clameter, Christoph Lameter
2007-11-20 1:12 ` [rfc 44/45] Remove local_t support clameter, Christoph Lameter
2007-11-20 12:59 ` Mathieu Desnoyers
2007-11-20 20:48 ` Christoph Lameter
2007-11-20 1:12 ` [rfc 45/45] Modules: Hack to handle symbols that have a zero value clameter, Christoph Lameter
2007-11-20 2:20 ` Mathieu Desnoyers
2007-11-20 2:49 ` Christoph Lameter
2007-11-20 3:29 ` Mathieu Desnoyers
2007-11-20 1:18 ` [rfc 00/45] [RFC] CPU ops and a rework of per cpu data handling on x86_64 Christoph Lameter
2007-11-20 1:51 ` David Miller
2007-11-20 1:59 ` Christoph Lameter
2007-11-20 2:10 ` David Miller
2007-11-20 2:12 ` Christoph Lameter
2007-11-20 3:25 ` Andi Kleen
2007-11-20 3:33 ` Christoph Lameter
2007-11-20 4:04 ` David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200711200416.32954.ak@suse.de \
--to=ak@suse.de \
--cc=akpm@linux-foundation.org \
--cc=clameter@sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@polymtl.ca \
--cc=travis@sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox