[parisc-linux] bogomips?

[parisc-linux] bogomips?

[Peter Lavender]

Hi everyone,

What's a 715/64 equate to in terms of intel arches?

Here's my p90 with 48meg of ram:
Calibrating delay loop... 35.84 BogoMIPS

and my 715/64 with 128 meg of ram:
Calibrating delay loop... 63.69 BogoMIPS

Thanks,

Pete

Re: [parisc-linux] bogomips?

[kenneth westelinck]

Hmm, don't know. 

My little 486DX4 100MHz, 12MB RAM:
/proc# cat ./cpuinfo 
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 4
model           : 8
model name      : 486 DX/4
stepping        : 0
fdiv_bug        : no
hlt_bug         : no
sep_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme
bogomips        : 49.66

/proc# 


regards,

Kenneth

----- Original Message ----- 
From: "Peter Lavender" <plaven@bigpond.net.au>
To: "parisc" <parisc-linux@parisc-linux.org>
Sent: Saturday, January 12, 2002 12:10 PM
Subject: [parisc-linux] bogomips?


> Hi everyone,
> 
> What's a 715/64 equate to in terms of intel arches?
> 
> Here's my p90 with 48meg of ram:
> Calibrating delay loop... 35.84 BogoMIPS
> 
> and my 715/64 with 128 meg of ram:
> Calibrating delay loop... 63.69 BogoMIPS
> 
> Thanks,
> 
> Pete
> 
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux
> 

Re: [parisc-linux] bogomips?

[Michael S.Zick]

Hmmm....
I don't know if this is the intended operation or not (probably a question 
for the Linux tree dudes), but...

Kenneth's 486 reports at about 1/2 the clock speed (also notice the lack of a 
cpu MHz line in the report).

This Celeron reports at about 2* the clock speed. (Kernel 2.4.17)

If memory (mine) serves, my P-II system reported at appox. bogomips == cpu 
clock.  I don't recall what my Pentium or 386 system reported and have never 
had a 486.

I think I noticed the change in the relationship between bogomips and cpu 
clock early in kernel 2.2.x (while I was running the P-II system) - perhaps 
something has crept in from the Linux tree that is having this effect.

Mike

[mszick@localhost /proc]$ cat ./cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 6
model name      : Celeron (Mendocino)
stepping        : 5
cpu MHz         : 467.737
cache size      : 128 KB
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 mmx fxsr
bogomips        : 930.61
 
[mszick@localhost /proc]$

On Saturday 12 January 2002 05:35 am, you wrote:
> Hmm, don't know.
>
> My little 486DX4 100MHz, 12MB RAM:
> /proc# cat ./cpuinfo
> processor       : 0
> vendor_id       : GenuineIntel
> cpu family      : 4
> model           : 8
> model name      : 486 DX/4
> stepping        : 0
> fdiv_bug        : no
> hlt_bug         : no
> sep_bug         : no
> f00f_bug        : no
> coma_bug        : no
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu vme
> bogomips        : 49.66
>
> /proc#
>
>
> regards,
>
> Kenneth
>
> ----- Original Message -----
> From: "Peter Lavender" <plaven@bigpond.net.au>
> To: "parisc" <parisc-linux@parisc-linux.org>
> Sent: Saturday, January 12, 2002 12:10 PM
> Subject: [parisc-linux] bogomips?
>
> > Hi everyone,
> >
> > What's a 715/64 equate to in terms of intel arches?
> >
> > Here's my p90 with 48meg of ram:
> > Calibrating delay loop... 35.84 BogoMIPS
> >
> > and my 715/64 with 128 meg of ram:
> > Calibrating delay loop... 63.69 BogoMIPS
> >
> > Thanks,
> >
> > Pete
> >
> >
> > _______________________________________________
> > parisc-linux mailing list
> > parisc-linux@lists.parisc-linux.org
> > http://lists.parisc-linux.org/mailman/listinfo/parisc-linux
>
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Re: [parisc-linux] bogomips?

[Randolph Chung]

> What's a 715/64 equate to in terms of intel arches?
> 
> Here's my p90 with 48meg of ram:
> Calibrating delay loop... 35.84 BogoMIPS
> 
> and my 715/64 with 128 meg of ram:
> Calibrating delay loop... 63.69 BogoMIPS

bogomips is not a good measure of this..... try looking at the SPEC
website (www.spec.org) for the CPU95/SPECint95/SPECfp95 results.

randolph
-- 
   @..@                                         http://www.TauSq.org/
  (----)
 ( >__< )
 ^^ ~~ ^^

Re: [parisc-linux] bogomips?

[Tom]

On Sat, 12 Jan 2002, Peter Lavender wrote:

> Hi everyone,
>
> What's a 715/64 equate to in terms of intel arches?
>
> Here's my p90 with 48meg of ram:
> Calibrating delay loop... 35.84 BogoMIPS
>
> and my 715/64 with 128 meg of ram:
> Calibrating delay loop... 63.69 BogoMIPS

As a rough guess, my 715/100 feels like a Pentium chip of roughly twice
the speed, while my C180 feels like a P2 of roughly twice the speed. Don't
know if the bogomips back that up, but that's the way they feel.

Re: [parisc-linux] bogomips?

[Grant Grundler]

Tom wrote:
> As a rough guess, my 715/100 feels like a Pentium chip of roughly twice
> the speed, while my C180 feels like a P2 of roughly twice the speed. Don't
> know if the bogomips back that up, but that's the way they feel.

I've heard that from a bunch of people and seems like a good rule of thumb.

Someone once told me: "bogomips is just a dick waving contest".
(apologies to any female readers who are offended).
But that statement accurately  sums up how useful bogomips
is for comparing perf.

grant

Re: [parisc-linux] bogomips?

[kenneth westelinck]

Than my server has a rather small dick.
But they always claim, size doesn't count ;)

----- Original Message -----
From: "Grant Grundler" <grundler@dsl2.external.hp.com>
To: "Tom" <palinux@alaskatech.org>
Cc: "Peter Lavender" <plaven@bigpond.net.au>; "parisc"
<parisc-linux@parisc-linux.org>
Sent: Saturday, January 12, 2002 11:43 PM
Subject: Re: [parisc-linux] bogomips?


> Tom wrote:
> > As a rough guess, my 715/100 feels like a Pentium chip of roughly twice
> > the speed, while my C180 feels like a P2 of roughly twice the speed.
Don't
> > know if the bogomips back that up, but that's the way they feel.
>
> I've heard that from a bunch of people and seems like a good rule of
thumb.
>
> Someone once told me: "bogomips is just a dick waving contest".
> (apologies to any female readers who are offended).
> But that statement accurately  sums up how useful bogomips
> is for comparing perf.
>
> grant
>
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux
>

Re: [parisc-linux] bogomips?

[Christian Suder]

http://www.linuxdoc.org/HOWTO/mini/BogoMips.html

For those who believe BogoMips matter..... ;-)

	Christian 


On Sat, 12 Jan 2002 15:43:06 -0700
"Grant Grundler" <grundler@dsl2.external.hp.com> wrote:

> Tom wrote:
> > As a rough guess, my 715/100 feels like a Pentium chip of roughly
twice> > the speed, while my C180 feels like a P2 of roughly twice the
speed. Don't> > know if the bogomips back that up, but that's the way they
feel.> 
> I've heard that from a bunch of people and seems like a good rule of
thumb.> 
> Someone once told me: "bogomips is just a dick waving contest".
> (apologies to any female readers who are offended).
> But that statement accurately  sums up how useful bogomips
> is for comparing perf.
> 
> grant
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Re: [parisc-linux] bogomips?

[Jochen Friedrich]

On Sun, 13 Jan 2002, Christian Suder wrote:

> For those who believe BogoMips matter..... ;-)

The document lists a HP9000/720 running HP/UX with 48 BogoMIPS. This seems 
to match Linux (Linux is even a bit faster :-) ):

# cat /proc/cpuinfo
processor       : 0
cpu family      : PA-RISC 1.1a
cpu             : PA7000 (PCX-S)
cpu MHz         : 50.000000
model           : 9000/720
model name      : Cobra (720)
hversion        : 0x00002000
sversion        : 0x00000481
I-cache         : 128 KB
D-cache         : 256 KB (WB)
ITLB entries    : 96
DTLB entries    : 96
BTLB fixed      : max. 4096 pages, pagesize=4096 (16MB)
BTLB fix-entr.  : 4 instruction, 4 data (0 combined)
BTLB var-entr.  : 0 instruction, 0 data (0 combined)
bogomips        : 49.86
software id     : 86914152

Cheers,
Jochen

Re: [parisc-linux] bogomips?

[Michael S.Zick]

On Sunday 13 January 2002 04:12 am, Christian Suder wrote:
> http://www.linuxdoc.org/HOWTO/mini/BogoMips.html
>
> For those who believe BogoMips matter..... ;-)
>
> 	Christian
>
Well, I would never claim that BogoMips is, in its self, useful for anything.

But since its code is closely related to the udelay code, which affects 
scheduling and many drivers...

Once upon a time, both my Celeron and my P-II reported approximately 1.00 * 
clock (same as above reference), then with some change of kernel version, 
jumped to 2.00 * clock.

BUT, that might have been an intentional change in the routine, or...

No promises on this one: I sometimes keep old kernels, perhaps I can find the 
"before and after" kernels where I noticed the change.

Mike

Re: [parisc-linux] bogomips?

[Matthew Wilcox]

On Sun, Jan 13, 2002 at 09:48:20AM -0600, Michael S. Zick wrote:
> Once upon a time, both my Celeron and my P-II reported approximately 1.00 * 
> clock (same as above reference), then with some change of kernel version, 
> jumped to 2.00 * clock.

yeah, that was a change prumpf made.  i forget why, but there was an
important reason.

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] bogomips?

[Michael S.Zick]

Oops,
I am once again exposed as a fool.  Please ignore my earlier copy of this 
message - I didn't read the full reference.  The change from 1.0 * clock to 
2.0 * clock is described in the reference.
Sorry folks,
Mike


On Sunday 13 January 2002 04:12 am, Christian Suder wrote:
> http://www.linuxdoc.org/HOWTO/mini/BogoMips.html
>
> For those who believe BogoMips matter..... ;-)
>
> 	Christian
>
Well, I would never claim that BogoMips is, in its self, useful for anything.

But since its code is closely related to the udelay code, which affects 
scheduling and many drivers...

Once upon a time, both my Celeron and my P-II reported approximately 1.00 * 
clock (same as above reference), then with some change of kernel version, 
jumped to 2.00 * clock.

BUT, that might have been an intentional change in the routine, or...

No promises on this one: I sometimes keep old kernels, perhaps I can find the 
"before and after" kernels where I noticed the change.

Mike

Re: [parisc-linux] bogomips?

[Alan Cox]

> On Sun, Jan 13, 2002 at 09:48:20AM -0600, Michael S. Zick wrote:
> > Once upon a time, both my Celeron and my P-II reported approximately 1.00 * 
> > clock (same as above reference), then with some change of kernel version, 
> > jumped to 2.00 * clock.
> 
> yeah, that was a change prumpf made.  i forget why, but there was an
> important reason.

We changed the x86 bogomip computation because it overflowed on a 2GHz 
processor. (Yes Linux 2.0 does not work on 2GHz preventium IV 8))

In all cases bogomips serves only two other serious purposes. Comparison
of identical processors, and checking nobody sold you the wrong processor.

Even in the x86 world this is quite obvious, with a 1.5GHz AMD flattening
a 2GHz Intel on actual throughput.

Alan