SEO

Someone else is going to do it soon or later so I'll beat them to it.

http://www-306.ibm.com/software/data...ks/111704.html

And if you believe what you just read ... take a good look at the small
print at the bottom of the page.

DB2 was run on:
DB2 UDB on IBM eServer p595 (64-way Power5 1.9GHz):
DB2 UDB on IBM eServer p570 16P (16-way Power5 1.9GHz)

While Oracle was run on:
HP Integrity rx5670 Cluster 64P (16 x 4-way Intel Itanium2 6M 1.5GHz)
Unisys ES7000 Aries 420 Ent. Server (16-way Intel Itanium2 1.5GHz)

Wow ... I guess IBM is saying a Power5 at 1.9GHz is equivalent to an
Itanium2 at 1.5GHz: I don't think so.

Only one citation is for identical hardware and there isn't enough
information to make a comparison meaningful. I'm thinking IBM is afraid
of a true head-to-head comparison or they wouldn't have had to stoop so
low.
--
Daniel A. Morgan
University of Washington
damorgan@x.washington.edu
(replace 'x' with 'u' to respond)


----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= East/West-Coast Server Farms - Total Privacy via Encryption =---

DA Morgan apparently said,on my timestamp of 22/01/2005 2:14 PM:
> Only one citation is for identical hardware and there isn't enough
> information to make a comparison meaningful. I'm thinking IBM is afraid
> of a true head-to-head comparison or they wouldn't have had to stoop so
> low.

Low, IBM? Narh!.....
I wonder why one of their sites here gave up
on the db2 mainframe implementation and is now
looking at other options outside of IBM. I guess
they didn't read this little marketing pearl.


--
Cheers
Nuno Souto
in sunny Sydney, Australia
wizofoz2k@yahoo.com.au.nospam

DA Morgan wrote:
> Someone else is going to do it soon or later so I'll beat them to it.
>
> http://www-306.ibm.com/software/data...ks/111704.html
>
> And if you believe what you just read ... take a good look at the
small
> print at the bottom of the page.

Why wouldn't you believe what was just read? They are industry
standard benchmarks and the full disclosure of the result are available
at www.tpc.org

>
> DB2 was run on:
> DB2 UDB on IBM eServer p595 (64-way Power5 1.9GHz):
> DB2 UDB on IBM eServer p570 16P (16-way Power5 1.9GHz)

There is a TPC-C result running on identical Power5 hardware (8 way p5
570). You can check it out at www.tpc.org. It shows DB2 result is
~430k and Oracle is at ~370k tansactions /minute. DB2 significantly
faster.

>
> While Oracle was run on:
> HP Integrity rx5670 Cluster 64P (16 x 4-way Intel Itanium2 6M 1.5GHz)
> Unisys ES7000 Aries 420 Ent. Server (16-way Intel Itanium2 1.5GHz)
>
> Wow ... I guess IBM is saying a Power5 at 1.9GHz is equivalent to an
> Itanium2 at 1.5GHz: I don't think so.

Yep Power5 hardware is much much faster than Itanium 2s. If IBM were
to run Oracle on a 64 way Power5 box, the result would be much closer
(but still below) to the 3.2million results from DB2, although I doubt
they ever would.

And the results they quoted are the highest results ever recorded with
the Oracle database.

>
> Only one citation is for identical hardware and there isn't enough
> information to make a comparison meaningful. I'm thinking IBM is
afraid
> of a true head-to-head comparison or they wouldn't have had to stoop
so
> low.

I already cited a head to head result. DB2 was faster.

> --
> Daniel A. Morgan
> University of Washington
> damorgan@x.washington.edu
> (replace 'x' with 'u' to respond)
>
>
> ----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet
News==----
> http://www.newsfeeds.com The #1 Newsgroup Service in the World!
>100,000 Newsgroups
> ---= East/West-Coast Server Farms - Total Privacy via Encryption =---

thu.nnguyen@gmail.com apparently said,on my timestamp of 22/01/2005 9:30 PM:

>
> Why wouldn't you believe what was just read? They are industry
> standard benchmarks and the full disclosure of the result are available
> at www.tpc.org

Do you believe in Santa as well?

>>Wow ... I guess IBM is saying a Power5 at 1.9GHz is equivalent to an
>>Itanium2 at 1.5GHz: I don't think so.
>
>
> Yep Power5 hardware is much much faster than Itanium 2s. If IBM were

You really didn't read what Daniel wrote, did you?
<sigh>

> (but still below) to the 3.2million results from DB2, although I doubt
> they ever would.

Me2.


--
Cheers
Nuno Souto
in sunny Sydney, Australia
wizofoz2k@yahoo.com.au.nospam

DA Morgan wrote:
> Someone else is going to do it soon or later so I'll beat them to it.
>
> http://www-306.ibm.com/software/data...ks/111704.html
>
> And if you believe what you just read ... take a good look at the small
> print at the bottom of the page.
>
> DB2 was run on:
> DB2 UDB on IBM eServer p595 (64-way Power5 1.9GHz):
> DB2 UDB on IBM eServer p570 16P (16-way Power5 1.9GHz)
>
> While Oracle was run on:
> HP Integrity rx5670 Cluster 64P (16 x 4-way Intel Itanium2 6M 1.5GHz)
> Unisys ES7000 Aries 420 Ent. Server (16-way Intel Itanium2 1.5GHz)
>
> Wow ... I guess IBM is saying a Power5 at 1.9GHz is equivalent to an
> Itanium2 at 1.5GHz: I don't think so.
>
> Only one citation is for identical hardware and there isn't enough
> information to make a comparison meaningful. I'm thinking IBM is afraid
> of a true head-to-head comparison or they wouldn't have had to stoop so
> low.
Pre-emptive trolling.. way to go...
Oracle wants its customers to rip out their Unix boxes and replace them
with RAC, claiming that it's cheaper. Also Oracle claims it can scale
limitless.
Well...
Let's assume the 3.2M TpmC are not from DB2, but only from the hardware.
The current #7 and #8 results (p690 on Oracle and DB2) were run as
apples to apples as it ever gets. Let's assume Oracle scales at least as
well on SMP as DB2 (IBM claims 99% scalability for DB2 in the article).
Extrapolating from there Oracle could also be expexted to achieve the
same result as DB2 at comparable price using the same metrics.
So... where is RAC on Intel cheaper? Would a 48 node RAC x4way CPUs
achieve the same result? What would it cost? Will the interconnects hold up?

Cheers
Serge

PS: Anyone know why RAC logs so much?

--
Serge Rielau
DB2 SQL Compiler Development
IBM Toronto Lab

Note in-line

Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html
The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/seminar.html
Public Appearances - schedule updated Jan 22nd 2005




"Serge Rielau" <srielau@ca.ibm.com> wrote in message
news:35fc2gF4dmo62U1@individual.net...
>
> PS: Anyone know why RAC logs so much?
>

If you can describe a simple scenario where the
logging done on RAC signficantly exceeds the
logging done on a non-RAC installation - with
an indication of data change volume, cross-instance
calls, and quantities of log generated, I can probably
tell you what's causing the difference. (And version
of Oracle, of course).


> --
> Serge Rielau
> DB2 SQL Compiler Development
> IBM Toronto Lab

Jonathan Lewis wrote:
> Note in-line
>
> Regards
>
> Jonathan Lewis
>
> http://www.jlcomp.demon.co.uk/faq/ind_faq.html
> The Co-operative Oracle Users' FAQ
>
> http://www.jlcomp.demon.co.uk/seminar.html
> Public Appearances - schedule updated Jan 22nd 2005
>
>
>
>
> "Serge Rielau" <srielau@ca.ibm.com> wrote in message
> news:35fc2gF4dmo62U1@individual.net...
>
>>PS: Anyone know why RAC logs so much?
>>
>
>
> If you can describe a simple scenario where the
> logging done on RAC signficantly exceeds the
> logging done on a non-RAC installation - with
> an indication of data change volume, cross-instance
> calls, and quantities of log generated, I can probably
> tell you what's causing the difference. (And version
> of Oracle, of course).
>
>
>
>>--
>>Serge Rielau
>>DB2 SQL Compiler Development
>>IBM Toronto Lab
>
>
>
I'm looking at the 1M+ TpmC Linux RAC result using O10g vs the 1M TpmC
Result without RAC on HP - also O10g
http://www.tpc.org/results/FDR/TPCC/...%2064P_FDR.pdf
vs.
http://www.tpc.org/results/FDR/TPCC/...d_1mil_fdr.pdf

Apparently the Linux RAC required 24500 GB (24.5TB!) for 8 hours of logging.
The HP SMP result only 2,481.03 (2.5TB). That's a factor of 10.

Now, if my understanding of RAC is correctly logging goes up for hot
pages. So I could imagine that the counter for the ORDERID may be nasty,
but not that nasty. Also only 10% of the orders are remote and the
cluster is partitioned.

Cheers
Serge
--
Serge Rielau
DB2 SQL Compiler Development
IBM Toronto Lab

I'm not sure I would call a TPCC disclosure
document with 163 or 206 pages a simple scenario,
but I've had a quick look through the documents.

The important points to note are that:
the log files are configured to 150 MB each

4 checkpoint occurred during the test

the checkpoints were triggered (apparently)
by explicit timeouts, rather than log switches.

I note that the figures you quote seem to be the
amount of space configured for log files, but
there is no explicit report of redo log generated -
only an upper limit implied by the checkpoint
count - and the document suggests that this
were forced on a timeout, not at logswitch.

In fact, the larger number you quote appears in
a block of numbers describing space allocation
for 60 days running, whereas the small number
appears in a section that comments on the need
for allocating logs for 8 hours.


--
Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html
The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/seminar.html
Public Appearances - schedule updated Dec 23rd 2004






"Serge Rielau" <srielau@ca.ibm.com> wrote in message
news:35fj7cF4ki6acU1@individual.net...
> Jonathan Lewis wrote:
>>
>> If you can describe a simple scenario where the
>> logging done on RAC signficantly exceeds the
>> logging done on a non-RAC installation - with
>> an indication of data change volume, cross-instance
>> calls, and quantities of log generated, I can probably
>> tell you what's causing the difference. (And version
>> of Oracle, of course).
>>
>>
>
> I'm looking at the 1M+ TpmC Linux RAC result using O10g vs the 1M TpmC
> Result without RAC on HP - also O10g
> http://www.tpc.org/results/FDR/TPCC/...%2064P_FDR.pdf
> vs.
> http://www.tpc.org/results/FDR/TPCC/...d_1mil_fdr.pdf
>
> Apparently the Linux RAC required 24500 GB (24.5TB!) for 8 hours of
> logging.
> The HP SMP result only 2,481.03 (2.5TB). That's a factor of 10.
>
> Now, if my understanding of RAC is correctly logging goes up for hot
> pages. So I could imagine that the counter for the ORDERID may be nasty,
> but not that nasty. Also only 10% of the orders are remote and the cluster
> is partitioned.
>
> Cheers
> Serge
>

thu.nnguyen@gmail.com wrote:

>
>
> There is a TPC-C result running on identical Power5 hardware (8 way p5
> 570). You can check it out at www.tpc.org. It shows DB2 result is
> ~430k and Oracle is at ~370k tansactions /minute. DB2 significantly
> faster.
>


You may want to look at the differences in the disk environment between
the two 'apples to apples' configurations. The IBM result used at least
20% more spindles than the Oracle result. It also had at least 30%
more clients driving the backend. This may explain why the IBM result
was 15% faster.

Jonathan Lewis wrote:
> I'm not sure I would call a TPCC disclosure
> document with 163 or 206 pages a simple scenario,
> but I've had a quick look through the documents.
>
> The important points to note are that:
> the log files are configured to 150 MB each
>
> 4 checkpoint occurred during the test
>
> the checkpoints were triggered (apparently)
> by explicit timeouts, rather than log switches.
>
> I note that the figures you quote seem to be the
> amount of space configured for log files, but
> there is no explicit report of redo log generated -
> only an upper limit implied by the checkpoint
> count - and the document suggests that this
> were forced on a timeout, not at logswitch.
>
> In fact, the larger number you quote appears in
> a block of numbers describing space allocation
> for 60 days running, whereas the small number
> appears in a section that comments on the need
> for allocating logs for 8 hours.
>
I was puzzling about that. The intro of the table of both 60 days and 8 hrs.
But if I apply 60 days to the log the difference is even stranger (just
the other way around). Obviously the form of the documents is rather
freestyle in this area which doesn't help grasping it.
Note that the table distinguished between space required and space
configured.
Thanks for the info.

Cheers
Serge

PS: Simple scenario equals less than 10 tables.. Imagine there would be
a real - real world benchmark ;-)

--
Serge Rielau
DB2 SQL Compiler Development
IBM Toronto Lab