dell 4gb ram PAE - asi dlhsie/delsi

patrik rybar rybar na datalock.sk
Úterý Duben 8 09:25:59 CEST 2003


Milan Kerslager wrote:
> On Mon, Apr 07, 2003 at 02:41:11PM +0200, patrik rybar wrote:
> 
>>mam maly problem s ktorym sa krutim stale dokola medzi supportom Dell-u a Redhat-u
>>
>>o co ide :
>>
>>dell 6600 4gb ram redhat as 2.1 , vidim v systeme iba 3,6 gb ram
>>
>>je to sposobene tym, ze :
>>
>>Systems that support hot-plug, hot-swap, and hot-add PCI capabilities,
>>with 4GB of memory installed may not report or have access to all
>>physical memory. An indication of this is the operating system
>>reporting less available memory than the amount of memory physically
>>present in the system.
>>
>>Hot-plug, hot-swap and hot-add PCI capabilities require that a certain
>>amount of memory in the upper range of the 4GB address space be
>>reserved. PCI devices installed in a system require all of their
>>resources to be defined and mapped in the 32-bit address (4GB) space.
>>The Flash Memory and APIC controllers are located at the top of memory
>>at address 0xFEC00000 and above, other PCI devices are defined below
>>this address range. Physical memory, which exists in this high address
>>space, may be unusable by the operating system.
>>
>>ve woknech sa to riesi parametrom /PAE do boot.ini
> 
> 
> "enterprise" kernel se od ostatnich v RH lisi tim, ze ma zapnute PAE
> (tj. zpristupneni pameti nad hranici 4GB na x86 architekture a to pouze
> u jader optimalizovanych pro i686 nebo Athlon). Pouziti PAE znamena
> penalizaci vykonu cca 6% a take bylo kdesi linux-kernel uvedeno, ze u
> gigabitovych karet snizuje vykon az 3x (myslim, ze to bude z technickych
> principu porad stejne). Pokud tedy nemusite, enterprise kernel
> NEpouzivejte.
> 
> Krome toho, ze nektere chipsety neumi presunout rezervovanou pamet nad
> hranici 4GB a proste pamet "sezerou", musite tedy uvazit, zda ozelite
> 400MB pameti nebo se smirite s penalizaci vykonu.
> 
> Ciste teoreticky by melo byt lepsi jit do 64bitove architektury, kdyz uz
> posilhavate po takovem vykonu.
> 
> 
>>support redhat tvrdi, ze to musi automaticky slapat pri pouziti
>>entreprise kernel-u
> 
> 
> To maji pravdu. Pokud to tak neni, bude neco spatne v domluve jadra a
> pouziteho chipsetu.
> 
> 
>>The enterprise kernel is able to use up to 64GB pagecache without any
>>modifications.  "If more than 4 Gigabytes is used then answer "64GB"
>>here. This selection turns Intel PAE (Physical Address Extension) mode
>>on.  PAE implements 3-level paging on IA32 processors. PAE is fully
>>supported by Linux, PAE mode is implemented on all recent Intel
>>processors (Pentium Pro and better). NOTE: If you say "64GB" here,
>>then the kernel will not boot on CPUs that don't support PAE! "
>>
>>
>>dell ma odkazuje na vyrobcu OS
>>
>>vyrobca OS robi mrtveho chrobaka a dava rady typu "mate kabel v elektrike ?"
>>
>>ani po update na najnovsi kernel pre RH AS ziadna zmena
>>uname -a
>>Linux dbsm 2.4.9-e.16enterprise #1 SMP Mon Mar 17 16:37:14 EST
>>ree
>>          total     used        free     shared    buffers       cached
>>Mem:       3605424     186700     3418724     28      7332        77364
>>-/+ buffers/cache:     102004     3503420
>>Swap:       6192996        0     6192996
>>
>>riesil niekto tento problem ?
> 
> 

dakujem,

konecne nejaka rozumna odpoved, od ktorej sa mozem odpichnut nejakym smerom



Další informace o konferenci Linux