"uname -p" on a AMD hardware shows i386 why?

Hi,

We have Solaris 10 installed on an AMD hardware.

Why does the "uname -p" command, which is supposed to display the processor type return i386?

I should have displayed x86_64, isn't it?

# uname -p

i386

# psrinfo -pv

The physical processor has 1 virtual processor (0)

x86 (AuthenticAMD family 15 model 5 step 10 clock 2389 MHz)

AMD Opteron(tm) Processor 850

The physical processor has 1 virtual processor (1)

x86 (AuthenticAMD family 15 model 5 step 10 clock 2389 MHz)

AMD Opteron(tm) Processor 850

The physical processor has 1 virtual processor (2)

x86 (AuthenticAMD family 15 model 5 step 10 clock 2389 MHz)

AMD Opteron(tm) Processor 850

The physical processor has 1 virtual processor (3)

x86 (AuthenticAMD family 15 model 5 step 10 clock 2389 MHz)

AMD Opteron(tm) Processor 850

# isainfo -kv

64-bit amd64 kernel modules

Please suggest,

Thanks & Regards,

[1007 byte] By [gshekar] at [2007-11-26 11:35:44]
# 1

> Hi,

>

> We have Solaris 10 installed on an AMD hardware.

> Why does the "uname -p" command, which is supposed to

> display the processor type return i386?

> I should have displayed x86_64, isn't it?

> # uname -p

> i386

Depends on what is intended by 'type', I suppose. Even on SPARC, the individual processor model is not shown, just 'sparc'. Both of them form processor families which have upward binary compatibility for executables.

Use other commands like you've done (such as psrinfo and isainfo) to get more specific processor information.

I guess the question is, why do you care about the specific processor model? What depends on that information? Given that, we might be able to suggest alternatives.

--

Darren

Darren_Dunham at 2007-7-7 11:33:05 > top of Java-index,General,Maintenance...
# 2

Hi,

I am looking for a way in which end users / developers at my location can determine the hardware architecture by themselves, without needing to be root user.

They determine the hardware architecure of a linux server themselves by using the "uname -m" command as shown below:

[root@bangpcrh214 ~]# uname -m

i686

[root@bangpcia3 ~]# uname -m

ia64

[root@bgpcem15 ~]# uname -m

x86_64

[root@bangpcplin1 ~]# uname -m

ppc64

However on a solaris server irrespective of i686 or x86_64 (EM64T/AMD) the "uname -m" command shows x86pc.

bash-2.05# uname -m

i86pc

below output is from a AMD server:

# uname -m

i86pc

# uname -m

sun4u

Please suggest,

Thanks & Regards

gshekar at 2007-7-7 11:33:05 > top of Java-index,General,Maintenance...
# 3

> Hi,

>

> I am looking for a way in which end users /

> developers at my location can determine the hardware

> architecture by themselves, without needing to be

> root user.

I don't have access to an opteron system at the moment. I assume that 'uname -k' also is identical though.

psrinfo -v (which you showed earlier) does not require root access.

I'm not sure if 'prtdiag' works properly on the original release of Solaris 10, but it should be available on later versions. It will mention the processor type.

--

Darren

Darren_Dunham at 2007-7-7 11:33:05 > top of Java-index,General,Maintenance...
# 4
bash-2.05# which psrinfo/usr/sbin/psrinfoYou must be root to run psrinfo.Also -k option is invalid for uname.bash-2.05# uname -kuname: illegal option -- kusage: uname [-snrvmapiX]uname [-S system_name]
gshekar at 2007-7-7 11:33:05 > top of Java-index,General,Maintenance...
# 5

> bash-2.05# which psrinfo

> /usr/sbin/psrinfo

>

> You must be root to run psrinfo.

I have no idea where that is coming from.

$ id

uid=1004(ddunham) gid=10(staff)

$ /usr/sbin/psrinfo

0on-linesince 09/11/06 16:51:14

1on-linesince 09/11/06 16:51:15

In addition, I don't see that error message within psrinfo...

# strings /usr/sbin/psrinfo | grep root

#

> Also -k option is invalid for uname.

>

> bash-2.05# uname -k

> uname: illegal option -- k

> usage: uname [-snrvmapiX]

> uname [-S system_name]

Sorry. Was thinking 'arch -k'. But I'm afraid that'll give the same information as 'uname -m'.

--

Darren

Darren_Dunham at 2007-7-7 11:33:05 > top of Java-index,General,Maintenance...