]
IST.EDU Subject
Re: 1 480 MIP IFL CPU = ?
01/31/2007 12:13
PM
Please respond to
Linux on 390 Port
[EMAIL PROTECTED]
IST.EDU
Subject
Re: 1 480 MIP IFL CPU = ?
01/31/2007 12:13
PM
Please respond to
Linux on 390 Port
[EMAIL PROTECTED]
IST.EDU
Instructions and architecture matter. I'd put
@VM.MARIST.EDU
cc:
Subject:Re: 1 480 MIP IFL CPU = ?
It also doesn't account for the fact that the Intel processor is doing its
own I/O, where that is almost completely off-loaded on the z9 to other
processors. In all, it really is an apples to oranges comparison, and
whatever
2007-02-01 14:50
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP IFL CPU = ?
Hi Robert.
Don't you think, that, for example, SCSI controller plays the same role in
Intel world?
Thank you.
WBR, Sergey
RPN01 [EMAIL PROTECTED]
Sent by: Linux on 390 Port LINUX-390@VM.MARIST.EDU
:
Subject:Re: 1 480 MIP IFL CPU = ?
Not quite, Sergey
In the play computer world, the CPU drives most of the IO - on z, most IO
functions are offloaded to the SAP.
SCSI controller can better be compared to a z control unit.
Regards
From: Linux on 390 Port
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP IFL CPU = ?
Hi Bertus.
Can we tell that most of the IO is a significant part of overall cpu
power? You know, if it took 1-2% of the CPU, it wouldn't be so much.
Also, is your statement right if we compare Linux/390 with zfcp
To: LINUX-390@VM.MARIST.EDU
cc:
Subject:Re: 1 480 MIP IFL CPU = ?
Difficult to say, Sergey
Most z9 boxes running z/OS, z/VM an Linux get away with the standard 2
SAPs / MCM, so the percentage is probably lower than 20% on average. Heavy
IO OSs such as TPF often uses more
is not doing the I/O. On
chip cache helps, but it is still a problem.
-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] Behalf Of
Sergey Korzhevsky
Sent: Thursday, February 01, 2007 4:51 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP IFL CPU = ?
Hi Robert.
Don't you
On Thu, 1 Feb 2007 08:17:13 -0800
Fargusson.Alan [EMAIL PROTECTED] wrote:
The actual problem with I/O on Intel and most system other then zSeries is
not ware the actual I/O takes place. The problem is that the CPU and DMA are
on the same bus. This means that while I/O is taking place the
At some performance presentation I heard nice sentence
about this - ... and all those fast processors wait
at the same high speed
Marian
--- Alan [EMAIL PROTECTED] wrote:
periods of polling. That
can significantly impact CPU availability and since
its often time bound a
faster processor
-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On
Behalf Of Dave Jones
Sent: Wednesday, January 31, 2007 8:52 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP IFL CPU = ?
Well, according to this article Design methods for attaining IBM
System z9
To: LINUX-390@VM.MARIST.EDU
cc:
Subject:Re: 1 480 MIP IFL CPU = ?
-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On
Behalf Of Dave Jones
Sent: Wednesday, January 31, 2007 8:52 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP
No problem, buy lower capacity setting and then you
can overclock through web via resourcelink, no need
for screwdriver.
Warning - works only on CP, not on IFL.
--- McKown, John [EMAIL PROTECTED]
Hum, I wonder if we could overclock a z9 CPU?
Well, it works on Intel
CPUs and AMD CPUs. Why not
Instructions and architecture matter. I'd put a z9 closer to
2.5Ghz than 1.7 in comparing linux workloads A 2Ghz intel
will NEVER outperform a z9.
Well, according to this article Design methods for attaining IBM
System z9 processor cycle-time goals
I'd like to make sure that everyone understands that my original posting
comparing a 1.7 GHz. z9 against a 2.0 GHz. Pentium was done tongue in
cheek;-)
DJ
barton wrote:
Instructions and architecture matter. I'd put a z9 closer to
2.5Ghz than 1.7 in comparing linux workloads A 2Ghz
On 1/31/07, Dave Jones [EMAIL PROTECTED] wrote:
I'd like to make sure that everyone understands that my original posting
comparing a 1.7 GHz. z9 against a 2.0 GHz. Pentium was done tongue in
cheek;-)
When I got my hands on a z9 we ran some instruction sequences to
determine the cycle
... Religious argument at best! :-( but the better answer would be
go back through the Archives of this list, as many many opinions have
been offered on the subject that IMHO simply imply it depends, as it
does.
Craig Otway wrote:
Can anyone point me to a chart or share statistics that
On 1/30/07, Craig Otway [EMAIL PROTECTED] wrote:
Can anyone point me to a chart or share statistics that show that each 480
MIP IFL CPU on a 2096 equals:
x number of x.x GHz Intel CPUs or x number of pSeries CPUs?
Barton's number is 4.
So 480 MIPS would be around a 2 GHz Pentium.
Rob
On Jan 30, 2007, at 3:24 PM, Craig Otway wrote:
Can anyone point me to a chart or share statistics that show that
each 480
MIP IFL CPU on a 2096 equals:
x number of x.x GHz Intel CPUs or x number of pSeries CPUs?
Unfortunately, the answer varies by a *LOT* depending on your
workload.
Can anyone point me to a chart or share statistics that show that
each
480
MIP IFL CPU on a 2096 equals:
x number of x.x GHz Intel CPUs or x number of pSeries CPUs?
Swags or rules of thumb are ok too. Thanks!
All together now: IT DEPENDS. 8-)
It's not a useful comparison. That's like
der Heij
Sent: Tuesday, January 30, 2007 2:01 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: [LINUX-390] 1 480 MIP IFL CPU = ?
On 1/30/07, Craig Otway [EMAIL PROTECTED] wrote:
Can anyone point me to a chart or share statistics that show that
each 480 MIP IFL CPU on a 2096 equals:
x number
: Tuesday, January 30, 2007 2:14 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: [LINUX-390] 1 480 MIP IFL CPU = ?
Can anyone point me to a chart or share statistics that show that
each
480
MIP IFL CPU on a 2096 equals:
x number of x.x GHz Intel CPUs or x number of pSeries CPUs?
Swags or rules
.
Mark Post
-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of
Craig Otway
Sent: Tuesday, January 30, 2007 5:28 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: 1 480 MIP IFL CPU = ?
We are looking at a new Cognos DW project running Oracle on the z/VM
Linux side
23 matches
Mail list logo