two processors...
Borkow
I have compaq proliant 850R and they said that include two PII 200, but I
dont know if its truth, becouse I cant see it inside :((
Borkow
Peter Köhlmann
> How to check if my server have two proccesors in command line?
>
Well, open a cmd line in your beloved Toys-R-Us OS
I am sure that the vastly superior windows help will tell you anything you
need to know
> I have compaq proliant 850R and they said that include two PII 200, but I
> dont know if its truth, becouse I cant see it inside :((
>
Too bad for you that windows is so shitty
--
Warning: 10 days have passed since your last Windows reinstall.
Howard J. Rogers
As root, type:
cat /proc/cpuinfo.
Regards
HJR
houghi
Run top and then press '1'. I get something like:
top - 01:34:43 up 12:49, 5 users, load average: 0.16, 0.11, 0.03
Tasks: 114 total, 3 running, 111 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.0% us, 0.3% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 1.3% us, 0.7% sy, 0.0% ni, 98.0% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 1035232k total, 976020k used, 59212k free, 115224k buffers
Swap: 1048784k total, 0k used, 1048784k free, 653856k cached
I actualy have only one processor. It is a Intel Pentium 3.04G
I would love to have the Intel Xeon, but they come a bit steep in
pricing. :-) (3970 EUR in one place I looked)
Till now I bought a new PC whenever the next one was about 10 times
faster. 40MHz, 350MHz, 3GHz. So the next one will be about 25GHz.
--
http://www.houghi.org | | _)
\ \ \ /\ \ \ /\ \ \ / \ _ \ | | _` | \ | _ \ _|_` |
\_/\_/ \_/\_/ \_/\_/_)_| _|\___/\_,_|\__, |_| _|_|_)\___/_|\__, |
My experience with SUSE. ____/ ____/
Howard J. Rogers
> Borkow wrote:
>> How to check if my server have two proccesors in command line?
>>
>> I have compaq proliant 850R and they said that include two PII 200, but I
>> dont know if its truth, becouse I cant see it inside :((
>
> Run top and then press '1'. I get something like:
> top - 01:34:43 up 12:49, 5 users, load average: 0.16, 0.11, 0.03
> Tasks: 114 total, 3 running, 111 sleeping, 0 stopped, 0 zombie
> Cpu0 : 0.0% us, 0.3% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0%
> si
> Cpu1 : 1.3% us, 0.7% sy, 0.0% ni, 98.0% id, 0.0% wa, 0.0% hi, 0.0%
> si
> Mem: 1035232k total, 976020k used, 59212k free, 115224k buffers
> Swap: 1048784k total, 0k used, 1048784k free, 653856k cached
>
> I actualy have only one processor. It is a Intel Pentium 3.04G
>
> I would love to have the Intel Xeon, but they come a bit steep in
> pricing. :-) (3970 EUR in one place I looked)
>
> Till now I bought a new PC whenever the next one was about 10 times
> faster. 40MHz, 350MHz, 3GHz. So the next one will be about 25GHz.
:-)
I stick to the doubling rule (400MHz, 800MHz, 1.7GB, 3GB. It's more
expensive, but more fun, I think!)
Incidentally... I didn't realise that top would list the CPUs like that, and
so just tried it on my hyperthreaded PIV 3GHz, and got this:
top - 10:25:13 up 12:20, 1 user, load average: 1.66, 1.34, 1.22
Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
Cpu(s): 97.3% us, 2.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.3% hi, 0.0% si
No multiple CPUs.
I'm pretty certain Suse 9.1 Pro has reported on 2 CPUs before now. I am,
after all, running the SMP kernel:
shostakovich:/home/howardjr # uname -a
Linux shostakovich 2.6.4-52-smp #1 SMP Wed Apr 7 02:11:20 UTC 2004 i686 i686
i386 GNU/Linux
What's happened to my 'other CPU', and how do I get it back?!
IE, slightly more technically, what has happened to hyperthreading support
in my installation, given I have not altered anything in the BIOS?
Any clues?
Regards
HJR
houghi
>
> As root, type:
>
> cat /proc/cpuinfo.
Not sure if this works, but you could try this if you only want to count
the amount:
clear;grep processor /proc/cpuinfo|tail -n 1|echo You have `awk {'print
$3'+1}` processors
All naturaly on one line.
Yeah, it is absolutely useless, I know. I should make a bash script from
it and put a check into it to see if there is only 1 processor and to
have that put processor instead of processors at the end (Or not to put
an s at the end.
Brian
[snips]
>
> Incidentally... I didn't realise that top would list the CPUs like that, and
> so just tried it on my hyperthreaded PIV 3GHz, and got this:
>
> top - 10:25:13 up 12:20, 1 user, load average: 1.66, 1.34, 1.22
> Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
> Cpu(s): 97.3% us, 2.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.3% hi, 0.0% si
>
> No multiple CPUs.
>
It won't, as installed - unless you tell it to with "1" (it toggles
on/off). If you find it does show 2 CPUs, save top's settings with "W"
for next time.
Start top and press "h" or "?" for the help screen.
> I'm pretty certain Suse 9.1 Pro has reported on 2 CPUs before now. I am,
> after all, running the SMP kernel:
>
> shostakovich:/home/howardjr # uname -a Linux shostakovich 2.6.4-52-smp
> #1 SMP Wed Apr 7 02:11:20 UTC 2004 i686 i686 i386 GNU/Linux
>
> What's happened to my 'other CPU', and how do I get it back?!
>
What did cat /proc/cpuinfo show?
B.
--
make: *** No rule to make target `whoopee'. Stop.
Randall Schulz
At Wednesday 01 September 2004 17:17 in alt.os.linux.suse houghi wrote:
> Howard J. Rogers wrote:
>>
>> As root, type:
>>
>> cat /proc/cpuinfo.
>
> Not sure if this works, but you could try this if you only want to count
> the amount:
>
> clear;grep processor /proc/cpuinfo|tail -n 1|echo You have `awk {'print
> $3'+1}` processors
You can get a little simpler than that.
echo "Your system's processor count is $(grep processor /proc/cpuinfo |wc
-l)"
If you're not using BASH, then this will work (it'll work in BASH, too,
but I'm fond of the $(interpolatedCommand) syntax):
echo "Your system's processor count is" `grep processor /proc/cpuinfo |wc
-l`
(I chose this wording to avoid having to deal with the singular vs. plural
suffix on "processor").
> All naturaly on one line.
Naturally.
> Yeah, it is absolutely useless, I know. I should make a bash script from
> it and put a check into it to see if there is only 1 processor and to
> have that put processor instead of processors at the end (Or not to put
> an s at the end.
Oh, things get *waaaaay* more useless than this. I mean, xeyes? What the
hell use is that?
Randall Schulz
res
Sorry to ask but I'm a newbie, I also have dual cpus , where do you put the
argument of 1 ?
Howard J. Rogers
> On Thu, 02 Sep 2004 10:28:16 +1000, Howard J. Rogers wrote:
>
> [snips]
>>
>> Incidentally... I didn't realise that top would list the CPUs like that,
>> and so just tried it on my hyperthreaded PIV 3GHz, and got this:
>>
>> top - 10:25:13 up 12:20, 1 user, load average: 1.66, 1.34, 1.22
>> Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
>> Cpu(s): 97.3% us, 2.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.3% hi, 0.0%
>> si
>>
>> No multiple CPUs.
>>
> It won't, as installed - unless you tell it to with "1" (it toggles
> on/off).
"Sorry. Only one CPU detected"
:-(
> If you find it does show 2 CPUs, save top's settings with "W"
> for next time.
> Start top and press "h" or "?" for the help screen.
>
>> I'm pretty certain Suse 9.1 Pro has reported on 2 CPUs before now. I am,
>> after all, running the SMP kernel:
>>
>> shostakovich:/home/howardjr # uname -a Linux shostakovich 2.6.4-52-smp
>> #1 SMP Wed Apr 7 02:11:20 UTC 2004 i686 i686 i386 GNU/Linux
>>
>> What's happened to my 'other CPU', and how do I get it back?!
>>
> What did cat /proc/cpuinfo show?
howardjr@shostakovich:~> cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Mobile Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3189.790
cache size : 512 KB
physical id : 0
siblings : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid
bogomips : 6324.22
I saw "2 siblings" there... I thought that was it?
Regards
HJR
Randall Schulz
At Wednesday 01 September 2004 18:19 in alt.os.linux.suse res wrote:
> Brian wrote:
>
>>> Incidentally... I didn't realise that top would list the CPUs like
>>> that, and so just tried it on my hyperthreaded PIV 3GHz, and got this:
>>>
>>> top - 10:25:13 up 12:20, 1 user, load average: 1.66, 1.34, 1.22
>>> Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
>>> Cpu(s): 97.3% us, 2.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.3% hi,
>>> 0.0% si
>>>
>>> No multiple CPUs.
>>>
>> It won't, as installed - unless you tell it to with "1" (it toggles
>> on/off). If you find it does show 2 CPUs, save top's settings with "W"
>> for next time.
>> Start top and press "h" or "?" for the help screen.
>>
>>> ...
>
> Sorry to ask but I'm a newbie, I also have dual cpus , where do you put
> the argument of 1 ?
The "top" command responds to characters typed while it's running. I
doesn't necessarily echo them, it just responds immediately or else
prompts for further arguments. Type '?' for built-in, on-screen help.
E.g., typing '1' toggles the "accumulate CPU usage stats into a single
average" display mode. Type 'u{userName}<RETURN>' to filter the output to
include only those processes whose user ID is that of the named user
(leave out the {curly braces}, by the way).
There a lots of other options available. Just experiment, top can't do
much (except kill or renice running processes--watch out for the 'k' and
'r' commands).
Randall Schulz
Randall Schulz
At Wednesday 01 September 2004 18:21 in alt.os.linux.suse Howard J. Rogers
wrote:
> Brian wrote:
>
>>> Incidentally... I didn't realise that top would list the CPUs like
>>> that, and so just tried it on my hyperthreaded PIV 3GHz, and got this:
>>>
>>> top - 10:25:13 up 12:20, 1 user, load average: 1.66, 1.34, 1.22
>>> Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
>>> Cpu(s): 97.3% us, 2.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.3% hi,
>>> 0.0% si
>>>
>>> No multiple CPUs.
>>>
>> It won't, as installed - unless you tell it to with "1" (it toggles
>> on/off).
>
> "Sorry. Only one CPU detected"
>
> :-(
>
>> ...
>>
>> What did cat /proc/cpuinfo show?
>
> howardjr@shostakovich:~> cat /proc/cpuinfo
> processor : 0
> ...
> siblings : 2
> ...
> flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
> mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid
> bogomips : 6324.22
>
> I saw "2 siblings" there... I thought that was it?
What kernel are you running? In order to take advantage of the two (sort
of) CPUs in your Hyper-Threading Pentium 4, you must run the "-smp"
kernel variant (SMP stands for "symmetric multi-processing," though
technically Intel's Hyper-Threading is what's known as SMT--simultaneous
multi-threading--the Linux kernel associates its SMT support with the SMP
support, at least in the configuration options provided by SuSE).
So, what does this command show:
% uname -r
2.6.5-7.108-smp
If you don't see the "-smp" there, you won't be able to exploit your HT
CPU. You'll have to install the "kernel-smp" package (first via YaST
followed by an update via YOU).
> Regards
> HJR
Good luck.
Randall Schulz
Howard J. Rogers
I think you might have missed that when I posted it earlier. I am indeed
running the smp kernel.
Regards & thanks anyway
HJR
Randall Schulz
At Wednesday 01 September 2004 18:56 in alt.os.linux.suse Howard J. Rogers
wrote:
> Randall Schulz wrote:
>
>> ...
>>
>>
>> What kernel are you running?
>
> I think you might have missed that when I posted it earlier. I am indeed
> running the smp kernel.
>
> Regards & thanks anyway
> HJR
Sorry--I was in catch-up mode for most of this thread and indeed
overlooked that fact.
What kernel options are you using? The wrong combination there can inhibit
HT support.
Also, look in /var/log/boot.msg. If things are working as they should,
you'd see something like this:
<4>CPU1: Intel(R) Pentium(R) 4 CPU 3.00GHz stepping 03
<6>Total of 2 processors activated (11927.55 BogoMIPS).
Randall Schulz
Howard J. Rogers
>
> Sorry--I was in catch-up mode for most of this thread and indeed
> overlooked that fact.
>
> What kernel options are you using? The wrong combination there can inhibit
> HT support.
>
> Also, look in /var/log/boot.msg. If things are working as they should,
> you'd see something like this:
Well, it's like looking behind the back of the fridge, isn't it?!
This doesn't look too good:
<4>Dell Inspiron with broken BIOS detected. Refusing to enable the local
APIC.
And then we have:
<4>Built 1 zonelists
<4>Kernel command line: root=/dev/hdc2 vga=0x317 desktop resume=/dev/hdc1
splash=silent
<6>bootsplash: silent mode.
<6>Initializing CPU#0
<4>PID hash table entries: 4096 (order 12: 32768 bytes)
<4>CKRM Initialized
<4>Detected 3189.790 MHz processor.
Something's stuffed, because I guarantee I saw two CPUs once. Somewhere. Oh,
it all seems so long ago...
I can smell another installation coming on.
Regards
HJR
Paul J Gans
>Borkow wrote:
>> How to check if my server have two proccesors in command line?
>>
>> I have compaq proliant 850R and they said that include two PII 200, but I
>> dont know if its truth, becouse I cant see it inside :((
>Run top and then press '1'. I get something like:
>top - 01:34:43 up 12:49, 5 users, load average: 0.16, 0.11, 0.03
>Tasks: 114 total, 3 running, 111 sleeping, 0 stopped, 0 zombie
> Cpu0 : 0.0% us, 0.3% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
> Cpu1 : 1.3% us, 0.7% sy, 0.0% ni, 98.0% id, 0.0% wa, 0.0% hi, 0.0% si
>Mem: 1035232k total, 976020k used, 59212k free, 115224k buffers
>Swap: 1048784k total, 0k used, 1048784k free, 653856k cached
>I actualy have only one processor. It is a Intel Pentium 3.04G
>I would love to have the Intel Xeon, but they come a bit steep in
>pricing. :-) (3970 EUR in one place I looked)
>Till now I bought a new PC whenever the next one was about 10 times
>faster. 40MHz, 350MHz, 3GHz. So the next one will be about 25GHz.
I'm profligate. I usually take the plunge when the "sweet
spot" CPU's are about three times the speed of the one I'm
running.
Right now I'm at 2.8 Ghz.
There are always a range of CPU speeds available. The slower
ones are cheaper, but there is usually a point where a slight
increase in speed has a large jump in price. The slower chip
is at the "sweet spot".
That point changes as production ramps up on fast CPUs and
prices drop. So the "sweet spot" often changes.
---- Paul J. Gans
Randall Schulz
At Wednesday 01 September 2004 19:16 in alt.os.linux.suse Howard J. Rogers
wrote:
> Randall Schulz wrote:
>
>> ...
>>
>> What kernel options are you using? The wrong combination there can
>> inhibit HT support.
>>
>> Also, look in /var/log/boot.msg. If things are working as they should,
>> you'd see something like this:
>>
>> <4>CPU1: Intel(R) Pentium(R) 4 CPU 3.00GHz stepping 03
>> <6>Total of 2 processors activated (11927.55 BogoMIPS).
>>
>
> Well, it's like looking behind the back of the fridge, isn't it?!
>
> This doesn't look too good:
>
> <4>Dell Inspiron with broken BIOS detected. Refusing to enable the local
> APIC.
D'Oh! I knew there was something else I was going to mention. Don't ask me
why, but for HyperThreading to work, it has to be both supported by and
enabled in the BIOS.
Judging from that message, this looks like a hardware-specific hack put
into the kernel for troublesome hardware. I grepped around in the kernel
sources and foud several references to Inspiron. There were several
references to sub-par BIOS implementations, including one place where it
stated that the 8/24/2000 release was "known bad (Dell Inspiron 5000e)"
while the 9/29/2000 releas was "known good (Dell Inspiron 5000)".
Perhaps you should check on updates to the BIOS and if there are any
available, flash your MB with the newest one. And if there is one (or
some) available, the release notes might disclose a pertinent fix.
I'm not sure what are the consequences of not having the APIC working
(Advanced Programmable Interrupt Controller), but I'm sure it's better if
it can be used. Here's the first thing I came across:
From <http://www.os.nctu.edu.tw/courseinfo/STOS_2003Fall/20030423.ppt>
(Google "APIC Linux Advantages" and view the HTML conversion of the PPT
file):
"Distributed Interrupt Handling
"APIC is introduced to replace conventional 8259 Programmable Interrupt
Controller.
"Intel IO APIC have 24 input, others may have as many as 64.
"Benefits of APIC
"Interrupt can be delivered to specific Local APIC or to the processor
which is running lowest priority task.
"Another important feature of the APIC allows CPUs to generate
inter-processor interrupts.
8259 is slow, while APIC is fast."
So it seems SMP- (and possibly SMT-) related and may have something to do
with why you're not getting your HT activated.
> And then we have:
>
> <4>Built 1 zonelists
> <4>Kernel command line: root=/dev/hdc2 vga=0x317 desktop
> resume=/dev/hdc1 splash=silent
> <6>bootsplash: silent mode.
> <6>Initializing CPU#0
> <4>PID hash table entries: 4096 (order 12: 32768 bytes)
> <4>CKRM Initialized
> <4>Detected 3189.790 MHz processor.
>
> Something's stuffed, because I guarantee I saw two CPUs once. Somewhere.
> Oh, it all seems so long ago...
>
> I can smell another installation coming on.
I wouldn't do that, at least not yet. A proper diagnosis of the problem is
always best. This isn't Microsoft Tech support, you know. Most of the
time reinstallation (of applications or operating systems) is a
superstitious and vastly overblown response to a problem.
> Regards
> HJR
RRS
Howard J. Rogers
> Howard,
>
> At Wednesday 01 September 2004 19:16 in alt.os.linux.suse Howard J. Rogers
> wrote:
[snip]
> D'Oh! I knew there was something else I was going to mention. Don't ask me
> why, but for HyperThreading to work, it has to be both supported by and
> enabled in the BIOS.
It is. I checked that one, thinking I may have had a fit of lunacy earlier
and switched it off for some reason. I hadn't.
> Judging from that message, this looks like a hardware-specific hack put
> into the kernel for troublesome hardware. I grepped around in the kernel
> sources and foud several references to Inspiron. There were several
> references to sub-par BIOS implementations, including one place where it
> stated that the 8/24/2000 release was "known bad (Dell Inspiron 5000e)"
> while the 9/29/2000 releas was "known good (Dell Inspiron 5000)".
>
> Perhaps you should check on updates to the BIOS and if there are any
> available, flash your MB with the newest one. And if there is one (or
> some) available, the release notes might disclose a pertinent fix.
Sad thing is, I have the A36 (I think it was) release, which is as new as it
gets. I will pay Dell a visit and have another poke around. Dell is,
however, the company that plasters "We recommend XP" all over their
marketing material, so hopes are not exactly high.
:-)
Well, I did it anyway... and guess what? It fixed nothing!
Yup, still single CPU (I'm beginning to wish the subject had never come
up!).
The curious thing for me is that the new installation has again installed
the SMP kernel:
howardjr@shostakovich:~> uname -r
2.6.4-52-smp
Now, on consulting my notebook of installs, I discover that I never checked
hyperthreading in Suse, but that it was working just fine in Fedora Core 2.
I'm going to obliterate Suse just one more time, put FC2 back on here, and
see for certain what I'm getting.
This is a bummer, frankly. It's not like I ever really noticed it wasn't
working (as you can tell from this thread!). So I suppose I can live
without it. But it's not a good feeling to know that some of your hardware
is not being used.
I appreciate the effort you've put into grepping around for things relevant.
If you could magic up a solution, I would be extremely happy of course. In
the meantime, is there no-one else with Suse 9.1 on an Inspiron 5150,
perhaps with HT working?
Regards
HJR
Zurab Davitiani
> As root, type:
>
> cat /proc/cpuinfo.
Usually, you don't have to be root to do this. The file should be readable
by everyone.
Randall Schulz
At Wednesday 01 September 2004 21:42 in alt.os.linux.suse Howard J. Rogers
wrote:
> Randall Schulz wrote:
>
>> ...
>>>
>>> I can smell another installation coming on.
>>
>> I wouldn't do that, at least not yet. A proper diagnosis of the problem
>> is always best. This isn't Microsoft Tech support, you know. Most of
>> the time reinstallation (of applications or operating systems) is a
>> superstitious and vastly overblown response to a problem.
>
> :-)
>
> Well, I did it anyway... and guess what? It fixed nothing!
>
> Yup, still single CPU (I'm beginning to wish the subject had never come
> up!).
>
> The curious thing for me is that the new installation has again
> installed the SMP kernel:
>
> howardjr@shostakovich:~> uname -r
> 2.6.4-52-smp
You should probably update to the latest kernel. SuSE has put out
2.6.5-7.108 to fix the audio CD recording problems.
However, using Linux Google (<http://www.google.com/linux>) to search for
"Inspiron 5150 HyperThreading" the first hit is relevant. You (well, I
anyway) must use the Google cached page, but it directly addresses this
issue:
"Does HyperThreading (HT) work on Inspiron 5150 with Linux?
"Hi,
"I am about to purchase a Inspiron 5150 laptop and I am going install
RedHat 9.0 on that machine. I have read this forum and there have been
one or two mentions that HT is NOT working on their machine. Is this a
general problem or have these problems been individual cases? Has anyone
been able to run Linux on Inspiron 5150 model with HT activated? I know
that there seems to be certain ACPI problems with Insipiron 5150, but are
there any other ones that I should be aware of when running Linux on the
machine.
"Re: Does HyperThreading (HT) work on Inspiron 5150 with Linux?
" Yes, hyperthreading will work but unless you're going to use Fedora
Core...You will need to re-compile the kernel, same with ACPI..you need
to build a custom kernel typically if you want to use bleeding edge
features..."
At least that's consistent with your experience. By the way, was
HyperThreading "bleeding edge" in November of 2003?
> ...
>
> I appreciate the effort you've put into grepping around for things
> relevant. If you could magic up a solution, I would be extremely happy
> of course. In the meantime, is there no-one else with Suse 9.1 on an
> Inspiron 5150, perhaps with HT working?
I'm a latter-day St. Jude.
> Regards
> HJR
RRS
David Wright
<snip>
>
> Oh, things get *waaaaay* more useless than this. I mean, xeyes? What the
> hell use is that?
>
>
> Randall Schulz
Killing the corporate mini and network backbone?
Way-back-when, when we got out first X-Terminal, some bright spark thought
it would be great fun to open up hundreds of tiny X-Eyes sessions on his
workstation, all in neat little row, then when the screen was full, he
waved the mouse around a manic speed...
An hour later, after re-starting the VAX 4000, the system manager removed
X-Eyes from the system and the techie that did the eyes experiment got a
rollicking... Think they also put in some sort of user process limit after
that as well ;-)
Dave
David Wright
Well, as they say Hyperthreading is one of the big selling points on their
Intel based CPU's and then flog them with XP Home - which won't boot on
multi-CPU or machines with Hyperthreading - you have to wonder if they know
anything about Intel CPU's... (As I've never bought a HT Pentium, I am only
going by what the magazine reviews at the time of XP's launch said, that XP
Home didn't support multi-processor or multi-threading, if Dell have a
botched together version of XP Home which does support HT, then I
apologise.)
It always struck me as strange to make a big thing of having hyperthreading
and then selling the machine with it switched off and an OS that doesn't
support it...
Could explain why there are comments to broken Dell BIOS's in the kernel
code as referenced earlier.
Dave
Juhan Leemet
I don't know about hyperthreading. Don't have that myself.
I do have SuSE 9.1 running on a dual 1GHz SMP Dell PowerEdge 1400SC. At
one point it didn't seem to recognize 2 CPUs. Turned out that I had both
the smp and default kernels installed, with the default installed last.
The softlink from /boot/vmlinuz pointed to the default. OK, makes sense. I
reinstalled the smp kernel (updated softlinks) and got my 2 CPUs back.
What does "more /proc/cpuinfo" show? On an SMP machine it shows multiple
entries: one for each CPU. Dunno what you see there for hyperthreading?
Another indicator (but not command line) for multiple CPUs is the xosview
panel. Cute moving coloured bars indicating dynamic % used. It has a
separate bar for each CPU. When I was running the default (mono) kernel,
it showed only 1 CPU bar. Now it shows both: one for each CPU.
BTW, I'm surprised to see (two sets of) interrupts (0-31). Huh? I thought
there were only 16 interrupts, but in SuSE-8.2 that grew to 24, and now on
9.1 it's 32. This must be some kind of multiple interrupt mapping?
--
Juhan Leemet
Logicognosis, Inc.
Peter Köhlmann
< snip >
> BTW, I'm surprised to see (two sets of) interrupts (0-31). Huh? I thought
> there were only 16 interrupts, but in SuSE-8.2 that grew to 24, and now on
> 9.1 it's 32. This must be some kind of multiple interrupt mapping?
>
No. It is the APIC. This interrupt controller is present in every X86 based
SMP machine
--
Computers are like air conditioners -
they stop working properly when you open Windows
Howard J. Rogers
>> The curious thing for me is that the new installation has again
>> installed the SMP kernel:
>>
>> howardjr@shostakovich:~> uname -r
>> 2.6.4-52-smp
>
> You should probably update to the latest kernel. SuSE has put out
> 2.6.5-7.108 to fix the audio CD recording problems.
The full story (after a hasty round of CD shuffling) is:
Suse 9.0 with kernel 2.4.21-178-smp4G shows 2 processors
Fedora Core 2 with kernel 2.6.5-1.358smp shows 2 processors
Out-of-the-box Suse 9.1 Pro with 2.6.4-52-smp shows 1 processor
And a patched Suse 9.1 Pro with 2.6.5-7.104-smp shows 1 processor.
You'll notice that Fedora Core 2 has an *older* kernel than the patched Suse
9.1, and yet FC2 sees two kernels where Suse 9.1 Pro doesn't.
I wish someone would explain that to me!
> However, using Linux Google (<http://www.google.com/linux>) to search for
> "Inspiron 5150 HyperThreading" the first hit is relevant. You (well, I
> anyway) must use the Google cached page, but it directly addresses this
> issue:
>
> "Does HyperThreading (HT) work on Inspiron 5150 with Linux?
>
> "Hi,
>
> "I am about to purchase a Inspiron 5150 laptop and I am going install
> RedHat 9.0 on that machine. I have read this forum and there have been
> one or two mentions that HT is NOT working on their machine. Is this a
> general problem or have these problems been individual cases? Has anyone
> been able to run Linux on Inspiron 5150 model with HT activated? I know
> that there seems to be certain ACPI problems with Insipiron 5150, but are
> there any other ones that I should be aware of when running Linux on the
> machine.
>
>
> "Re: Does HyperThreading (HT) work on Inspiron 5150 with Linux?
>
> " Yes, hyperthreading will work but unless you're going to use Fedora
> Core...You will need to re-compile the kernel, same with ACPI..you need
> to build a custom kernel typically if you want to use bleeding edge
> features..."
>
>
> At least that's consistent with your experience. By the way, was
> HyperThreading "bleeding edge" in November of 2003?
I had seen the first of those articles a while back (actually I bought the
laptop in December 2003, and Red Hat 9 was a candidate O/S... I think I
remember reading that exact thread then). The second one you quote sounds a
bit more like it... I'll have a read.
I don't want to sound like a moaning minnie on the subject, but if FC2 can
do it, why can't Suse? I *don't* mean that to be a featureless whinge
actually -I really would like to understand what differentiates two distros
in this sort of regard, since I thought the kernel was kinda sacrosanct,
and it was merely the packaging that was different. Evidently not.
On another note: if I take the plunge and learn how to compile a custom
kernel, I fear that's just a can of worms (see my post on VMware, for
example). Besides which, I've never done it before, and don't have a clue.
> I'm a latter-day St. Jude.
Oh well... that's cheered me up no end, hasn't it!! :-)
Regards
HJR
Howard J. Rogers
[snip]
>> I appreciate the effort you've put into grepping around for things
>> relevant. If you could magic up a solution, I would be extremely happy of
>> course. In the meantime, is there no-one else with Suse 9.1 on an
>> Inspiron 5150, perhaps with HT working?
>>
>> Regards
>> HJR
>
> Well, as they say Hyperthreading is one of the big selling points on their
> Intel based CPU's and then flog them with XP Home
Mine came with XP Pro, actually.
You'll be able to put the garlic away shortly, but if I was installing XP
these days, I wouldn't even dream of mucking around with Home.
> - which won't boot on
> multi-CPU or machines with Hyperthreading - you have to wonder if they
> know anything about Intel CPU's... (As I've never bought a HT Pentium, I
> am only going by what the magazine reviews at the time of XP's launch
> said, that XP Home didn't support multi-processor or multi-threading, if
> Dell have a botched together version of XP Home which does support HT,
> then I apologise.)
>
> It always struck me as strange to make a big thing of having
> hyperthreading and then selling the machine with it switched off and an OS
> that doesn't support it...
Which, sadly, is what I appear to be currently stuck with courtesy of Suse
9.1.
OK, you can put the garlic away now. I'll be nice about Linux and Suse
again.
HJR
Howard J. Rogers
> On Thu, 02 Sep 2004 14:42:04 +1000, Howard J. Rogers wrote:
>
>> howardjr@shostakovich:~> uname -r
>> 2.6.4-52-smp
>
> Is there more than one smp kernel to choose from?
>
> On my SuSE 9.0 system I have a choice of:
>
> k_smp - Kernel with multiprocessor support which supports a maximum of 64G
> of RAM
>
> and
>
> k_smp4G - Kernel with multiprocessor support which supports a maximum of
> 4G of RAM
>
> I'm using k_smp4G and my hyperthreaded system shows two CPUs.
>
> Oh, and stop that reinstalling business!
>
> Karen
Hi Karen... OK. I've nearly stopped!
I don't get a choice of kernels. I could say that drives me nuts, but I
won't because that's just a poor joke.
I've posted elsewhere this arvo that Suse 9.0 did install a 2.4...smp4G
kernel. Fedora also provides me with a choice, an SMP and a non-SMP. But
9.1 just does the plain old smp one.
At least, I *think* that's true. Here's what I have in /boot:
erntypes-2.6.5-7.104-smp
System.map-2.6.5-7.104-smp
backup_mbr
boot -> .
config-2.6.5-7.104-smp
grub
initrd-2.6.5-7.104-smp
initrd-2.6.5-7.104-smp
message
symvers-2.6.5-7.104-i386-smp.gz
vmlinux-2.6.5-7.104-smp.gz
vmlinuz-2.6.5-7.104-smp
vmlinuz-2.6.5-7.104-smp
(Slightly edited, but only slightly).
Dare I ask two questions following on from yours, then? If Suse's
installation process doesn't automatically want to install different
flavours of kernel, does that mean it would be foolish to somehow force it
to do so? And, second, assuming it isn't beeing completely foolish, how
exactly would one go about doing that (during an installation, I mean. I
don't want to even think about retrofitting one to this current system,
since it was re-installed (ooops) in some haste)?
Thanks for trying to help, anyway.
Best regards
HJR
David Wright
> David Wright wrote:
>
> [snip]
>
>>> I appreciate the effort you've put into grepping around for things
>>> relevant. If you could magic up a solution, I would be extremely happy
>>> of course. In the meantime, is there no-one else with Suse 9.1 on an
>>> Inspiron 5150, perhaps with HT working?
>>>
>>> Regards
>>> HJR
>>
>> Well, as they say Hyperthreading is one of the big selling points on
>> their Intel based CPU's and then flog them with XP Home
>
> Mine came with XP Pro, actually.
>
> You'll be able to put the garlic away shortly, but if I was installing XP
> these days, I wouldn't even dream of mucking around with Home.
>
Yes, the business models come with Pro, but the home and generally
advertised models claim HyperThreading, and also XP Home, although nowhere
in the advert does it say that the two are mutually exclusive...
All my Windows machines have XP Pro now, although they were mostly delivered
with XP Home.
Most manufacturers believe that is you have decent hardware, then you won't
need to put it into a business network, and if you want to connect to a
business network, then you don't want decent hardware... (fast processor,
nVidia GeForce video and Augidy sound? Can't be doing with that
Professional nonsense, you want Home mate!)
I actually ended up being bounced around by Evesham.com. I wanted one of
their top of the line machines (about 2 years ago), but asked for XP
Professional. They said I needed to talk to the business division. The
business division said yes, I could have XP, but the hardware I wanted
wasn't within their remit, I could have a slower machine without an GeForce
card and without an SB Audigy soundcard.
As I was working away a lot and didn't have time to build a machine myself,
I settled on the power option and upgraded the OS, but nowadays, I build
them myself again. At least then I get exactly what I want... The good(?)
news nowadays is that with most of the machines going over to Linux, if I
do need a test machine with Windows on it for some reason, I have more than
enough spare licences kicking around... ;-)
>> - which won't boot on
>> multi-CPU or machines with Hyperthreading - you have to wonder if they
>> know anything about Intel CPU's... (As I've never bought a HT Pentium, I
>> am only going by what the magazine reviews at the time of XP's launch
>> said, that XP Home didn't support multi-processor or multi-threading, if
>> Dell have a botched together version of XP Home which does support HT,
>> then I apologise.)
>>
>> It always struck me as strange to make a big thing of having
>> hyperthreading and then selling the machine with it switched off and an
>> OS that doesn't support it...
>
> Which, sadly, is what I appear to be currently stuck with courtesy of Suse
> 9.1.
>
> OK, you can put the garlic away now. I'll be nice about Linux and Suse
> again.
>
> HJR
Not a problem ;-) Although is it a Linux/SuSE problem or a poorly written
BIOS problem with the Dell? I must admit, when I worked for a client that
used Dells as their desktop standard, I used to peruse the Dell newsgroup
and most people didn't have nice things to say about their hardware,
especially where it often seemed to deviate from accepted standards, or
used butchered OEM versions of hardware. I don't know if they've changed
their practices in the last couple of years....
Dave
David Wright
<snip>
>
> Not a problem ;-) Although is it a Linux/SuSE problem or a poorly written
> BIOS problem with the Dell? I must admit, when I worked for a client that
> used Dells as their desktop standard, I used to peruse the Dell newsgroup
> and most people didn't have nice things to say about their hardware,
> especially where it often seemed to deviate from accepted standards, or
> used butchered OEM versions of hardware. I don't know if they've changed
> their practices in the last couple of years....
>
> Dave
ok, going over your reply to Karen, it seems to be a 9.1 Kernel 2.6 issue -
with the Dell BIOS specifically or with Intel HP in general we can't tell
at this point... ;-)
Dave
Howard J. Rogers
> Yes, the business models come with Pro, but the home and generally
> advertised models claim HyperThreading, and also XP Home, although nowhere
> in the advert does it say that the two are mutually exclusive...
>
> All my Windows machines have XP Pro now, although they were mostly
> delivered with XP Home.
>
> Most manufacturers believe that is you have decent hardware, then you
> won't need to put it into a business network, and if you want to connect
> to a business network, then you don't want decent hardware... (fast
> processor, nVidia GeForce video and Augidy sound? Can't be doing with that
> Professional nonsense, you want Home mate!)
That is actually very true, though I'd not thought about it till now!
[snip]
> Not a problem ;-) Although is it a Linux/SuSE problem or a poorly written
> BIOS problem with the Dell?
Don't know. But if Fedora and Suse 9.0 can give me HT, I think it might be
called a 9.1 issue!!
And things were going along so well...
Cheers
HJR
Howard J. Rogers
> In article <4136c8ab$0$16145$afc3...@news.optusnet.com.au>,
> Howard J. Rogers <h...@dizwell.com> wrote:
>>
>> The full story (after a hasty round of CD shuffling) is:
>>
>> Suse 9.0 with kernel 2.4.21-178-smp4G shows 2 processors
>> Fedora Core 2 with kernel 2.6.5-1.358smp shows 2 processors
>>
>> Out-of-the-box Suse 9.1 Pro with 2.6.4-52-smp shows 1 processor
>> And a patched Suse 9.1 Pro with 2.6.5-7.104-smp shows 1 processor.
>
> I'd vote for whatever version of the kernel SuSE 9.1 started with (or the
> modifications they made) didn't support HT. Try downgrading or upgrading,
> or (horrors) compiling your own kernel.
>
>> You'll notice that Fedora Core 2 has an *older* kernel than the patched
>> Suse 9.1, and yet FC2 sees two kernels where Suse 9.1 Pro doesn't.
>
> More support for the theory that it's SuSE 9.1's source kernel (2.6.4?)
> that's incompatible.
>
>> On another note: if I take the plunge and learn how to compile a custom
>> kernel, I fear that's just a can of worms (see my post on VMware, for
>> example). Besides which, I've never done it before, and don't have a
>> clue.
>
> Eh, it's no biggie. Start with a working config file. I can supply mine
> (SuSE 9.1, 2.6.8, Thinkpad T40 [1.4 MHz P4-M, Radeon], ACPI) if you can't
> find one. Read http://www.digitalhermit.com/linux/Kernel-Build-HOWTO.html
Thanks for the link!
I really don't feel up to doing this though!!
Call me a scaredy cat.
Regards
HJMiaow.
houghi
> You can get a little simpler than that.
>
> echo "Your system's processor count is $(grep processor /proc/cpuinfo |wc
> -l)"
Nice one.
--
http://www.houghi.org | | _)
\ \ \ /\ \ \ /\ \ \ / \ _ \ | | _` | \ | _ \ _|_` |
\_/\_/ \_/\_/ \_/\_/_)_| _|\___/\_,_|\__, |_| _|_|_)\___/_|\__, |
My experience with SUSE. ____/ ____/
Paul J Gans
>[snip]
>:-)
I have linux on an Inspiron 5150. I'll check tonight when I get
home.
---- Paul J. Gans
Paul J Gans
<huge snip>
>Well, as they say Hyperthreading is one of the big selling points on their
>Intel based CPU's and then flog them with XP Home - which won't boot on
>multi-CPU or machines with Hyperthreading - you have to wonder if they know
>anything about Intel CPU's... (As I've never bought a HT Pentium, I am only
>going by what the magazine reviews at the time of XP's launch said, that XP
>Home didn't support multi-processor or multi-threading, if Dell have a
>botched together version of XP Home which does support HT, then I
>apologise.)
I thought hyperthreading was fairly new. Didn't it come out last
year and then only in some CPU's?
>It always struck me as strange to make a big thing of having hyperthreading
>and then selling the machine with it switched off and an OS that doesn't
>support it...
Ah, marketing... Isn't it wonderful?
>Could explain why there are comments to broken Dell BIOS's in the kernel
>code as referenced earlier.
And oh yes, let's not forget cost control... ;-)
---- Paul J. Gans
Paul J Gans
>In article <4136c8ab$0$16145$afc3...@news.optusnet.com.au>,
>>
>> Suse 9.0 with kernel 2.4.21-178-smp4G shows 2 processors
>> Fedora Core 2 with kernel 2.6.5-1.358smp shows 2 processors
>>
>> Out-of-the-box Suse 9.1 Pro with 2.6.4-52-smp shows 1 processor
>> And a patched Suse 9.1 Pro with 2.6.5-7.104-smp shows 1 processor.
>I'd vote for whatever version of the kernel SuSE 9.1 started with (or the
>modifications they made) didn't support HT. Try downgrading or upgrading,
>or (horrors) compiling your own kernel.
>> You'll notice that Fedora Core 2 has an *older* kernel than the patched Suse
>> 9.1, and yet FC2 sees two kernels where Suse 9.1 Pro doesn't.
>More support for the theory that it's SuSE 9.1's source kernel (2.6.4?)
>that's incompatible.
>> On another note: if I take the plunge and learn how to compile a custom
>> kernel, I fear that's just a can of worms (see my post on VMware, for
>> example). Besides which, I've never done it before, and don't have a clue.
>Eh, it's no biggie. Start with a working config file. I can supply mine
>(SuSE 9.1, 2.6.8, Thinkpad T40 [1.4 MHz P4-M, Radeon], ACPI) if you can't
>find one. Read http://www.digitalhermit.com/linux/Kernel-Build-HOWTO.html .
Slightly off-topic: Is the config file used to compile the
SuSE installed kernel available anywhere? I don't think
that .oldconfig that comes with the kernel source install
is it, though I may be wrong.
----- Paul J. Gans
Paul J Gans
It has to be the Inspiron BIOS. I have an smp kernel running on my
home brew box at home (dual boot with Windows) and it reports 2
processors on my 2.8 Ghz CPU. I no longer recall if it is a
Prescott or the other version. I know I bought the older one
because I was leery of heat problems reported with the new chip.
---- Paul J. Gans
Brian
[snips]
>
> I'd vote for whatever version of the kernel SuSE 9.1 started with (or the
> modifications they made) didn't support HT. Try downgrading or upgrading,
> or (horrors) compiling your own kernel.
>
Hmm, it's an odd one Ok - but just to chip in with my findings of an
out-the-box 9.1 install.
Didn't pay much attention to what it thought about my HT-enabled P4 CPU,
(I was more interested in the end result!) but one of the last things it
did on completing the install was to automatically fire up YOU and install
the latest updates available (this was shortly after 9.1 was released).
On booting this freshly installed & updated system, there were 2 CPUs
purring away nicely. This was all auto-detected and dealt with by the
SuSE installer - I'd only tweaked the usual user-based options -
partition, GUI choice, etc. etc.
There have been a few kernel updates since 9.1 was released and so far,
none have clobbered smp.
B.
--
Outgoings certified 100% Microsoft-free.
Howard J. Rogers
> I have linux on an Inspiron 5150. I'll check tonight when I get
> home.
>
> ---- Paul J. Gans
Hi Paul:
Did you have a look? I saw you refer to a 'home-brew' box with HT enabled,
but home-brew doesn't sound quite like an inspiron...
If it was the inspiron, do you know what BIOS you're using? I'm definitely
using the latest A36 one.
Appreciate it.
Regards
HJR
Howard J. Rogers
> On Thu, 02 Sep 2004 10:49:28 -0700, Howard J. Rogers wrote:
>
>> Dare I ask two questions following on from yours, then? If Suse's
>> installation process doesn't automatically want to install different
>> flavours of kernel, does that mean it would be foolish to somehow force
>> it to do so? And, second, assuming it isn't beeing completely foolish,
>> how exactly would one go about doing that (during an installation, I
>> mean. I don't want to even think about retrofitting one to this current
>> system, since it was re-installed (ooops) in some haste)?
>
> Oh dear, I didn't mean that I had a choice at boot, I mean I have a choice
> of installing. When I go to Yast, install software, search for kernel, I
> find several choices.
>
> Since you have a new installation now, fetch your update list and see what
> the choices are for new kernels. I don't do automatic updates, I look over
> the list and pick and choose what I want. It seems that I always have a
> choice of several when a kernel update is offered.
>
> There have been posts made here before about maintaining multiple kernels
> and I'll have to refer you to those notes.
>
> Karen
Message received and understood on the installation business. I'll give that
a go shortly. Very much appreciate your efforts...
Regards
HJR
re...@verizon.net
wrote:
>Res,
>
>At Wednesday 01 September 2004 18:19 in alt.os.linux.suse res wrote:
>
/snip/
>> Sorry to ask but I'm a newbie, I also have dual cpus , where do you put
>> the argument of 1 ?
>
>The "top" command responds to characters typed while it's running. I
>doesn't necessarily echo them, it just responds immediately or else
>prompts for further arguments. Type '?' for built-in, on-screen help.
>
>E.g., typing '1' toggles the "accumulate CPU usage stats into a single
>average" display mode. Type 'u{userName}<RETURN>' to filter the output to
>include only those processes whose user ID is that of the named user
>(leave out the {curly braces}, by the way).
>
>There a lots of other options available. Just experiment, top can't do
>much (except kill or renice running processes--watch out for the 'k' and
>'r' commands).
>
>Randall Schulz
thanks
Howard J. Rogers
> Eh, it's no biggie. Start with a working config file. I can supply mine
> (SuSE 9.1, 2.6.8, Thinkpad T40 [1.4 MHz P4-M, Radeon], ACPI) if you can't
> find one. Read http://www.digitalhermit.com/linux/Kernel-Build-HOWTO.html
> .
It's been a long thread, and I'm grateful to everyone that's contributed to
it. Apologies to Hactar for not noticing this offer of a 'working config
file'. Until about 40 minutes ago, I had no idea what a config file was!
Anyway, I followed Hactar's other advice on the HOWTO, drank a foolhardy
amount of alcohol, and downloaded the 2.6.8.1 kernel and patch and got
compiling. After a lot of faffing around, and silly mishaps, I can report
this:
top - 10:07:25 up 7 min, 1 user, load average: 0.02, 0.37, 0.28
Tasks: 82 total, 1 running, 81 sleeping, 0 stopped, 0 zombie
Cpu0 : 1.7% us, 0.3% sy, 0.0% ni, 98.0% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 1.3% us, 0.3% sy, 0.0% ni, 98.3% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 1034780k total, 288832k used, 745948k free, 26864k buffers
Swap: 2040244k total, 0k used, 2040244k free, 166392k cached
You might notice that there are two CPUs there. Joy!
The only thing I changed in xconfig, IIRC, is to specify I was using a
Pentium 4 M (I think it defaulted to using an Intel x86 or something
equally generic). Of course, I also upped my kernel version several
notches, which may have helped!
I don't know what else I've busted doing this, so I'm going to wipe the
machine yet again, and do it all over from scratch, and try and make sure
it's all beautifully clean and perfect. Confidence is not high, but we
shall see.
I have just two further questions. The uname -r command now reports as
follows:
shostakovich:~ # uname -r
2.6.8.1
No SMP extension there, despite Top's output. Should I be concerned? Is the
kernel name reported there at all configurable during the compilation
process? Is it just a cosmetic thing, in other words?
Second, I am currently using the usb-audio generic driver, but have
downloaded the specific Extigy driver. I can compile that and install it as
a module, but I would like to build it into the kernel instead since I'm in
that line of business these days. How would one go about making a driver
which you have obtained elsewhere appear in the list of options in xconfig
to be built-in?
Many thanks to everyone again,
Regards
HJR
Paul J Gans
>>[snip]
>>:-)
I've now done that. I'm an idiot, as will become obvious.
First, Suse 9.1 only detects a single processor on the Inspiron
5150. Second, the Inspiron runs a Mobile Intel chip and as far
as I know, those are NOT hyperthreaded.
So the OS isn't wrong. And I should have known that right off.
---- Paul J. Gans
Paul J Gans
>In article <ch7h50$lp3$3...@reader1.panix.com>,
>>
>> >Eh, it's no biggie. Start with a working config file.
>> Slightly off-topic: Is the config file used to compile the
>> SuSE installed kernel available anywhere? I don't think
>> that .oldconfig that comes with the kernel source install
>> is it, though I may be wrong.
>My first try was /proc/config.gz, but that didn't work for me (turns out _a_
>problem was the malformed partition table on my laptop and it may have
>worked given a valid definition for /usr). I'd try that first. If you're
>not going to use an initrd, remember to make various things not modules.
Thanks. I've built kernels before. I've just never found
the config file used by the maker of the distribution I
was using.
So I usually had to reconfigure from scratch. And of course
things didn't work, so I'd configure again. And usually after
four or five compiles I'd get it right. This is a major bother
even though CPU speeds have gone up faster than the amount of
code to be compiled. I sadly remember 45 minute compiles on
what was then a fast machine...
----- Paul J. Gans
Howard J. Rogers
> First, Suse 9.1 only detects a single processor on the Inspiron
> 5150.
Ah Ha!
> Second, the Inspiron runs a Mobile Intel chip and as far
> as I know, those are NOT hyperthreaded.
But mine is a "Mobile Intel(R) Pentium(R) 4" too. And my Dell has a little
"Intel Inside" stick on it that includes two tiny little 'lozenges' on it,
reading 'H' and 'T'... so guess what, it *is* hyperthreaded. Well, mine is
anyway. Maybe you bought an earlier model? Check your BIOS, anyway, because
there should be a thing there where you can actually enable or disable the
BIOS.
Proof? Well, I've already posted it, but (feeling terribly clever with
myself for having recompiled my kernel for the first time ever) here goes
again for nothing:
top - 10:08:13 up 8 min, 1 user, load average: 0.54, 0.50, 0.33
Tasks: 82 total, 1 running, 81 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.7% us, 0.3% sy, 0.0% ni, 96.0% id, 0.0% wa, 3.0% hi, 0.0% si
Cpu1 : 0.7% us, 0.0% sy, 0.0% ni, 99.3% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 1034780k total, 289080k used, 745700k free, 27124k buffers
Swap: 2040244k total, 0k used, 2040244k free, 166336k cached
As I also mentioned elsewhere, the trick was to explicitly set the processor
type in xconfig when compiling the new kernel. I suppose this, and other
contributions, suggests that something in the Dell BIOS prevents accurate
recognition of the processor first time around.
> So the OS isn't wrong. And I should have known that right off.
'Tis too!! :-)
Seriously... check your BIOS. And if that mentions hyperthreading (on page
6, I believe from memory), then it's time to compile a kernel or two! If
not, you've been dudded by Dell!!
Regards
HJR
>
> ---- Paul J. Gans
Paul J Gans
>Hi Paul:
>Appreciate it.
>Regards
>HJR
I've responded to this already, but another won't hurt. My Inspiron
finds 1 cpu. But the 5150 runs a Mobile Pentium something or other
(I didn't write it down). I don't think *any* of the Mobile Pentiums
were hyperthreaded.
I didn't check the bios though.
My desktop (or better, deskside) box is home-brew with a 2.8 Ghz
Pentium chosen because it *was* hyperthreaded. It (I'm using it
right now) reports two CPUs.
---- Paul J. Gans
Howard J. Rogers
>>HJR
>
> I've responded to this already, but another won't hurt. My Inspiron
> finds 1 cpu. But the 5150 runs a Mobile Pentium something or other
> (I didn't write it down). I don't think *any* of the Mobile Pentiums
> were hyperthreaded.
Our posts are crossing, I think.
Yes, the Mobile Pentiums were HT'd. At lease, the 3.2GHz one was.
Regards
HJR
Howard J. Rogers
Oh, by the way, this from
http://www.intel.com/products/notebook/processors/index.htm?iid=ipp_note+proc&
Intel® Pentium® M Processor
The Intel® Pentium® M processor is a key component of Intel® Centrino™
mobile technology+. Based on an architecture designed specifically for
mobile computing, the Intel Pentium M processor delivers outstanding mobile
performance and low power enhancements for sleeker, lighter notebook
designs.
Mobile Intel® Pentium® 4 Processor with HT Technology†
Notebooks based on the Mobile Intel® Pentium® 4 processor with HT
Technology† deliver high-performance for users of larger notebook form
factors, with capabilities to allow users to take advantage of
multithreaded and processor-intensive multimedia applications all in a
portable form factor.
In other words, the Mobile P4 *is* HT'd. But the P4-M isn't. I think!!
It's one of the worst bits of product naming since Windows 95, I reckon.
Regards
HJR
Randall Schulz
At Thursday 02 September 2004 17:16 in alt.os.linux.suse Howard J. Rogers
wrote:
> ...
>
> Anyway, I followed Hactar's other advice on the HOWTO, drank a foolhardy
> amount of alcohol, and downloaded the 2.6.8.1 kernel and patch and got
> compiling. After a lot of faffing around, and silly mishaps, I can
> report this:
Way to go! For future reference, coffee is much recommended for most any
software-related purpose, at least on the creation side. Using the stuff
is another story, I suppose...
> Many thanks to everyone again,
> Regards
> HJR
RRS
Howard J. Rogers
> To modify the kernel options requires some sort of patch. Since you can
> make it as a module, why not go ahead and have that module loaded
> automatically?
>
Sotto voce... Because I don't really know how to do that yet!
Truth of the matter is that on first installation, Yast detects the internal
sound card, which is awful. I let that ride. Upon reboot after
installation, and with the Extigy now plugged in, it suddenly detects the
"usb-audio" device. The KaMix mixer window even defaults to the "usb-audio"
device.
If I manually load the proper Extigy drivers... er, I have no idea where
they are, whether they've been loaded, whether they're in use, or what.
KaMix is not noticeably different, for starters. And the sound doesn't
sound particularly better (or worse).
So my cunning plan was to add in the proper Extigy drivers as a static,
compiled option, and get rid of the "usb_audio" generic thing altogether.
Then there could be no ambiguity about it.
But I think it's sounding too much like a deep dark hole, and I might just
live with the generics after all. At least the music plays.
As you can gather, I'm still rather green on all this sort of stuff. And I
used to think installing Oracle was tricky!
(Incidentally, this is one of the major reasons why I ditched XP: it was so
damned boring, and you can't have as much fun as this with it!!)
Regards
HJR
Howard J. Rogers
> Eh, it's no biggie. Start with a working config file. I can supply mine
> (SuSE 9.1, 2.6.8, Thinkpad T40 [1.4 MHz P4-M, Radeon], ACPI) if you can't
> find one. Read http://www.digitalhermit.com/linux/Kernel-Build-HOWTO.html
Could I take you up on that offer, please?
The mail address is h...@dizwell.com
Appreciate it.
Regards
HJR
Paul J Gans
>Ah Ha!
>'Tis too!! :-)
>Regards
>HJR
I will look, but it isn't that serious to me. The machine
is plenty fast enough and I don't run any CPU intensive
computations on it.
There's another way to check. It's a dual boot machine.
Windows XP task manager will show two CPUs if it detects
them. I know that it does on my desktop.
---- Paul J. Gans
Paul J Gans
>In article <ch8e38$2vm$4...@reader1.panix.com>,
>You can relive those days again if you desire. Just remove (or disable)
>some of your RAM...
Hey, at my age *I* have that problem. I don't need to have
it in my computer too!.... ;-)
------ Paul J. Gans
Paul J Gans
>In article <4137b7f7$0$9064$afc3...@news.optusnet.com.au>,
>Howard J. Rogers <h...@dizwell.com> wrote:
<snip>
>> I have just two further questions. The uname -r command now reports as
>> follows:
>>
>> shostakovich:~ # uname -r
>> 2.6.8.1
>>
>> No SMP extension there, despite Top's output. Should I be concerned?
>No. SuSE (etc) appends "-smp" merely to differentiate the SMP-supporting
>kernel from the non-SMP-supporting kernel. "-smp" isn't automatically
>added.
Hmmm. uname -r on my desktop machine gives 2.6.4-54.5-smp.
<snip>
----- Paul J. Gans
Howard J. Rogers
> In article <4137db8b$0$28794$afc3...@news.optusnet.com.au>,
> Howard J. Rogers <h...@dizwell.com> wrote:
>> Hactar wrote:
>>
>> > To modify the kernel options requires some sort of patch. Since you
>> > can make it as a module, why not go ahead and have that module loaded
>> > automatically?
>> >
>>
>> Sotto voce... Because I don't really know how to do that yet!
>
> Edit /etc/sysconfig/kernel, look for "MODULES_LOADED_ON_BOOT". Don't put
> "foo.ko" there, but rather just "foo". (Look down a couple paragraphs if
> that's gibberish.)
>
>> Truth of the matter is that on first installation, Yast detects the
>> internal sound card, which is awful. I let that ride. Upon reboot after
>> installation, and with the Extigy now plugged in, it suddenly detects the
>> "usb-audio" device. The KaMix mixer window even defaults to the
>> "usb-audio" device.
>>
>> If I manually load the proper Extigy drivers... er, I have no idea where
>> they are, whether they've been loaded, whether they're in use, or what.
>
> You said they're kernel modules. Loaded modules can be listed by lsmod ,
> removed with rmmod and loaded (with dependencies) by modprobe or (without
> dependencies) by insmod. They generally live under /lib/modules/<kernel
> version>/ . Under kernel 2.6.x if you lsmod and see "foo", then it comes
> from a file named "foo.ko". Sometimes modules will announce their findings
> via syslog when they're loaded.
>
> Now if it's not a module and you were mistaken, I have no idea.
>
>> KaMix is not noticeably different, for starters. And the sound doesn't
>> sound particularly better (or worse).
>
> Loading the module may produce no change in sound, but rather different
> (or
> more) mixer channels, "new" hardware, etc. If it produces no visible
> change at all, don't bother with it.
>
>> So my cunning plan was to add in the proper Extigy drivers as a static,
>> compiled option, and get rid of the "usb_audio" generic thing altogether.
>> Then there could be no ambiguity about it.
>
> Well, I generally oppose non-modularized drivers. You can't unload them
> without rebooting. (Say your Extigy hardware wedges. If it's a module,
> rmmod and modprobe and you're on your way. If it's compiled in, tough;
> gotta reboot.)
>
>> As you can gather, I'm still rather green on all this sort of stuff. And
>> I used to think installing Oracle was tricky!
>
> Well, I've never used Oracle and I recompile kernels and modules all the
> time. Up to your perspective, I suppose.
>
>> (Incidentally, this is one of the major reasons why I ditched XP: it was
>> so damned boring, and you can't have as much fun as this with it!!)
>
> Some relative said I spent too much time fiddling with my computer. As I
> told her, some of us _like_ this.
>
Brilliant answer, so many thanks... (If only the man pages made as much
sense!). I will let you know how I go.
Regards
HJR
David Wright
<snip>
> It's one of the worst bits of product naming since Windows 95, I reckon.
>
> Regards
> HJR
Oh, I don't know, I always thought name Windows ME after Yuppy Flu was a big
mistake... At least it lived up to its name, slow, unreliable and sick as a
dog...
I still loved the bogus press release that went round a couple of years ago,
went something like: Microsoft announce their new version of Windows for
the marketplace. Incorporating the compactness and efficiency of Windows
CE; the availability of drivers under Windows ME; and the reliability and
stability of Windows NT: The all new Windows CEMENT.
Dave
David Wright
<snip>
> I will look, but it isn't that serious to me. The machine
> is plenty fast enough and I don't run any CPU intensive
> computations on it.
>
> There's another way to check. It's a dual boot machine.
> Windows XP task manager will show two CPUs if it detects
> them. I know that it does on my desktop.
>
> ---- Paul J. Gans
Depends Paul. If you are dual booting to Windows XP Home, then
Hyperthreading is disabled, because the Home version will refuse to boot on
a HyperThreaded system... This was my complaint earlier in the thread, the
standard configuration for Dell computers (mass market, not business
versions) is that they come with Windows XP Home installed and
HyperThreading disabled, even though HT is an advertised bonus of buying a
Dell system...
Dave
Mark Burgo
Why not just do a cloneconfig and make prepare-all that gives you the same
config as your system is using.
--
Mark B.
Linux 2.6.5-7.104-default
7:10am up 12 days 10:42, 9 users, load average: 0.13, 0.25, 0.56
Paul J Gans
Hey, what's with the uptime bits? Here's mine on this machine:
Linux 2.4.20-4GB
12:17:25 up 308 days, 18:12, 4 users, load average: 0.07, 0.04, 0.13
---- Paul J. Gans
Mark Burgo
Just part of the sig. It was better when my uptime was over 300 days.
--
Mark B.
Linux 2.6.5-7.104-default
12:41pm up 12 days 16:13, 9 users, load average: 0.13, 0.14, 0.26
houghi
>>Linux 2.6.5-7.104-default
>> 7:10am up 12 days 10:42, 9 users, load average: 0.13, 0.25, 0.56
>
> Hey, what's with the uptime bits? Here's mine on this machine:
>
> Linux 2.4.20-4GB
> 12:17:25 up 308 days, 18:12, 4 users, load average: 0.07, 0.04, 0.13
houghi@penne : uptime
6:40pm up 10:00, 4 users, load average: 0.06, 0.05, 0.01
So yiu mean you have not updated your kernel with the needed security
patches? What was your IP again? ;-)
--
http://www.houghi.org | | _)
\ \ \ /\ \ \ /\ \ \ / \ _ \ | | _` | \ | _ \ _|_` |
\_/\_/ \_/\_/ \_/\_/_)_| _|\___/\_,_|\__, |_| _|_|_)\___/_|\__, |
My experience with SUSE. ____/ ____/
Arthur Hagen
> Paul J Gans wrote:
>>
>>
>> Linux 2.4.20-4GB
>> 12:17:25 up 308 days, 18:12, 4 users, load average: 0.07, 0.04,
>> 0.13
>
> houghi@penne : uptime
> 6:40pm up 10:00, 4 users, load average: 0.06, 0.05, 0.01
>
> So yiu mean you have not updated your kernel with the needed security
> patches? What was your IP again? ;-)
(I know *you're* kidding, but someone might take this seriously despite of
your smiley, so...)
It is NOT necessary to install all "critical" security updates as they
appear. Only those that are exploitable on your system. If there's a
critical patch for SATA support, there's no need to update the kernel unless
you use SATA *and* lets other people access your system. A kernel NFS patch
isn't critical if your machine is behind a firewall that blocks NFS. A
machine that has been up and running for 3 years can be completely safe,
despite not having been patched.
Regards,
--
*Art
George Peatty
<david_c...@hotmail.com> wrote:
>The all new Windows CEMENT.
ROFLMAO!!
Harold Stevens
[Snip...]
> It is NOT necessary to install all "critical" security updates as they
> appear. Only those that are exploitable on your system.
FWIW, SuSE YOU seems to relatively smart about this, IME. I've upgraded
several systems in a day, including laptops, and noticed that YOU seems
to tailor security patches to the local YaST install profile.
For example, PCMCIA security patches usually aren't applied, except on
laptops (obviously, PCMCIA isn't useful/installed on most desktops).
--
Regards, Weird (Harold Stevens) * IMPORTANT EMAIL INFO FOLLOWS *
Pardon any bogus email addresses (wookie) in place for spambots.
Really, it's (wyrd) at airmail, dotted with net. DO NOT SPAM IT.
Kids jumping ship? Looking to hire an old-school type? Email me.
David Wright
nnnI thank you, I'm here all week ;-)
Dave
Paul J Gans
>> Paul J Gans wrote:
>>
>>
>>>>HJR
>>>
>>> I've responded to this already, but another won't hurt. My Inspiron
>>> finds 1 cpu. But the 5150 runs a Mobile Pentium something or other
>>> (I didn't write it down). I don't think *any* of the Mobile Pentiums
>>> were hyperthreaded.
>>
>> Our posts are crossing, I think.
>>
>> Yes, the Mobile Pentiums were HT'd. At lease, the 3.2GHz one was.
>>
>> Regards
>> HJR
>Oh, by the way, this from
>http://www.intel.com/products/notebook/processors/index.htm?iid=ipp_note+proc&
>Intel? Pentium? M Processor
>The Intel? Pentium? M processor is a key component of Intel? Centrino?
>mobile technology+. Based on an architecture designed specifically for
>mobile computing, the Intel Pentium M processor delivers outstanding mobile
>performance and low power enhancements for sleeker, lighter notebook
>designs.
>Mobile Intel? Pentium? 4 Processor with HT Technology?
>Notebooks based on the Mobile Intel? Pentium? 4 processor with HT
>Technology? deliver high-performance for users of larger notebook form
>factors, with capabilities to allow users to take advantage of
>multithreaded and processor-intensive multimedia applications all in a
>portable form factor.
>In other words, the Mobile P4 *is* HT'd. But the P4-M isn't. I think!!
>It's one of the worst bits of product naming since Windows 95, I reckon.
Mine is NOT hyperthreaded. It is a Mobile P4 2.66 Ghz CPU with
speedstep. I'm almost certain that it came out *before* hyperthreading
came in. The chip is NOT a Centrino.
I have no bios setting for enabling hyperthreading either.
I chose the chip because of the resulting price for the computer.
I could have chosen a faster (and hence newer) one.
--- Paul J. Gans
Paul J Gans
>Regards,
>--
>*Art
Yup. I didn't take it seriously.
The machine is coming down soon for a brain transplant. It will
move to a new box (my old home machine) and, I hope, SuSE 9.2.
---- Paul J. Gans
Arthur Hagen
> speedstep. I'm almost certain that it came out *before*
> hyperthreading came in. The chip is NOT a Centrino.
Centrino isn't a chip. It's a marketing label that manufacturers can put on
the case if they use the right combination of Intel CPUs, Intel chipsets and
Intel WiFi, and pay a certification fee to Intel. As such, I see it as
worse than useless, as it only adds to the price without giving anything in
return.
The Pentium M is *much* faster than the Pentium 4M and Pentium 4 compared to
the clock speed. The Pentium M is really much like a Pentium III Tualatin-S
with twice the cache and SSE2 added. A PIII-S tully easily outperforms a P4
at 1.5 times the clock speed, and the Pentium M takes it even farther.
As such, I would love to get a Pentium M for an ATX motherboard -- it would
be a worthy replacment for the Tualatin-S.
Unfortunately, the Pentium-M is rather expensive. However, if buying a
laptop, look whether you can find a PIII-S equipped one -- a 1.13S will
easily outperform all the mobile Celerons and many of the P4's, despite the
slower SDRAM speed.
> I have no bios setting for enabling hyperthreading either.
>
> I chose the chip because of the resulting price for the computer.
> I could have chosen a faster (and hence newer) one.
You don't usually want HyperThreading in a notebook. The temperature
increases significantly when running with HT, and that's not a good thing
for a laptop.
Regards,
--
*Art