oracle itanium vs. pa-risc performance

[Tony Drake_2]

Does anyone have some real numbers as far as the difference between performance on a PA RISC server and an Itanium server (ie anyone migrated from one to the other?)

We have some apps that are doing just that, and we need some preliminary idea of just how many Oracle instances we can stack on these consolidation servers. Since we just put in the systems, we don't really have any good benchmarks.

Thanks.
Tony

[Alzhy]

I am not aware of any published BMs but I suppose the disparity is wide. Look at the TPC web site.. virtually all of the top performers are no longer RISC based chips (with the exception of IBM's POWER of course) .. most are Itanium2 or Opteron.

As an FYI to everyone:

HP just announced that it will spend 3 Billion dollars in the next 3 years on Integrity Systems -- so I guess the roadmap will really be for an Itanium one.

[A. Clay Stephenson]

Will you be running native or RA-RISC code on the Itanium boxes? If native then you don't need to have any concerns. If PA-RISC code is being executed under the ARIES emulator then you can expect about a 3X or so performance hit depending upon which Itanium and which PA-RISC boxes you are comparing. However, this is not as bad as it sounds because this only applies to the CPU portion of the execution and most database applications are I/O rather than CPU bound so that even with a 3X hit (and that value is getting better all the time) the actual execution impact may be very small --- and may even be reduced to less than zero by the overall gains in I/O.

[Tony Drake_2]

We are running Oracle 9.2.x in Native Itanium mode.

[Jean-Luc Oudart]

Tony

you could check the intel site such as http://www.intel.com/products/benchmarks/server/index2.htm

but it does not say much.

I suppose the best would be to benchmark your own application on both platform.

Regards
Jean-Luc

[Tony Drake_2]

Thats interesting info Jean Luc.

I was hoping someone had some concrete tests that they had done already in the real world.

[Geoff Wild]

Attached is the "Oracle9i migration from PA-RISC to Itanium" doc.

Some stats in there.

Rgds...Geoff

[Geoff Wild]

And another whitepaper "Evolving your HP-UX environments from PA-RISC to the IntelR"

Rgds...Geoff

[Indira Aramandla]

Hi Tony,

Metalink Note:266220.1 â Migration from HP PA-RISC (64bit) to HP Itanium ia64â explains the migration steps and performance.

There is no special conversion being made when moving an Oracle database between PA-RISC and the Itanium architecture. The database structures and the database block layout are identical to PA-RISC, so when you move your PA-RISC database to the Itanium architecture, the database engine has no knowledge of the fact that it is operating on data that was created on another platform. There would be no noticeable performance benefit from going through the process of recreating your database natively on the Itanium-based system.

Attached is a document of laboratory performance comparisons Oracle 9.2 environments (PA-RISC vs.Itanium architecture).

In

[Tony Drake_2]

Let me clarify what I am asking about.

I am fully aware of the fact that the native databases and the migrated databases are similar in performance. What I want to know is what is the comparison between the database on the PA RISC vs. the database on Itanium Server.

The HP benchmarks use IBM PowerPC processors.

I know that Itanium is faster than PA_RISC.

What I want to know is how much faster.

So if I have a database query that takes 10 minutes on a PA RISC, would it take 5 minutes on an Itanium? I realize this is simplistic, but I would like to know where to set expectations.

That is the kind of thing I am looking for.

Thanks for all your posts.

[Julio Yamawaki]

Hi,

I installed a rx-2600 with 800 MHz Itanium with Oracle 9.2.0.2 and about 120 users.
I installed a rp-3410 with 800 MHz PA-8800 with Oracle 9.2.0.0 and about 100 users.
Some key issues:
1. Oracle installation: 15 minutes in Itanium, 23 minutes in PA-RISC
2. Import: 50 GB base in Itanium: 2,5 hours, 45 GB in PA-Risc: 3 hours
All disks area 15K rpm, 3 GB memory in each case.
So, my conclusion is that Oracle in Itanium has more performance than in PA-Risc.

Regards,

[Tony Drake_2]

Thanks. That's great.

If anyone else has any similar numbers I'd like to see them.

[Steven E. Protter]

Reality check:

Julio Yamawaki, conclusion should be "comparable"

The speed of the i/o on the bus on those two systems may be different, as the system boards are different.

Based on the fact that the data loads are not the same size, the test can not conclude that either platform is faster.

SEP

[Tony Drake_2]

I realize the differences Steven. Although the two systems are not completely identical, and there are differences between ITanium and PA RISC system boards, this does tell me that the two are just that, comparable.

I intend to do some benchmarking of our own, but I wanted to see if anyone had any data so I would know what to expect.

Thanks.
Tony

[Geoff Wild]

I was told by our account rep, that on comparable servers (PARISC to Itanium ), that you could half the cpus and get about the same performance.

IE: 4 X CPU RP about the same as 2 X CPU RX....

I have not seen any stats to validate this....

RGds...Geoff

[Hein van den Heuvel]

The 'best' PA is pretty close to a lowest IPF.
A recent box with (1.5 or 1.6 Ghz, 9Mb cache) Itanium can often double the performance when replacing a somewhat older PS system with the same mount of processors.

One proof point

http://www50.sap.com/benchmark/sd2tier.asp -->
All 8-way SMP, Oracle 9i, HPUX 11i. All tests in a few months time span, and by the same team so rather compareable.


Date-------SD Users---Model
01/12/2004 1240 rp4440, PA-8800 1.0 GHz
09/12/2003 1500 rx7620, Intel Itanium 2, 1.5 GHz
04/29/2004 1320 rx4640-8, 4 HP mx2 dual 1.1 GHz Itanium


http://www.tpc.org/tpcc/results/tpcc_results.asp?orderby=hardware

07/30/03 541,674 64* HP PA-RISC 8700 875MHz
11/04/03 1,008,144 64*Intel Itanium2 1.5GHz


hth,
Hein.

[Alzhy]

Interesting SAP benchmarks Hein.

It looks like an 8-way PA8800 1.0Ghz performed at par with an 8-way Intanium2 boxen when doing the SAP benchmarks:

http://www50.sap.com/benchmark/pdf/cert0104.pdf
http://www50.sap.com/benchmark/pdf/cert3704.pdf

[Geoff Wild]

Actually, Nelson's compares Windoze to HP-UX - different classes RX76xx to RP44xx

This is better:

4way Itanium, HPUX, RX4640
http://www50.sap.com/benchmark/pdf/cert3004.pdf

to a

8way PARISC, HPUX, RP4440
http://www50.sap.com/benchmark/pdf/cert0104.pdf

Performance very close...

Rgds...Geoff

[Steven E. Protter]

As I also forgot to mention.

800 Megahertz on PA-RISC is not the same as 800 Megahertz on Itanium.

RISC has different(reduced) Instruction set.

SEP

[Florian Heigl (new acc)]

Steven - I think the Itanium is a risc cpu, are You sure it isn't?

(I don't like it anyway. Currently at our site it's ~500/4 hppa/ipf and it's good that way.)

[Tony Drake_2]

ITanium and PA RISC are different architectures. ITanium is not really RISC, nor is it really CISC...they call it EPIC (Explicit Paralell Instruction...) Since Itanium is "Intel", I would have to think it probably more akin to a CISC instruction set, because Intel has never been big on RISC. They have tried to do quasi-risc architectures with some of the pentium chips where they take CISC operations and decompose them, but they never had a RISC chip Par-Se, at least in the PC realm.

There is no other chip that does things quite like this if I read the papers right...I am not pro or Con mind you...but it has the potential in certain types of software (I am thinking Scientific stuff like I did when in college) it has real potential to be blindingly fast. If programmed correctly, other software has the potential to also be faster than an equivalent non-itanium CISC (like a pentium) or RISC (like PA_RISC).

The thing is that all this becomes a moot point in most situations. I don't deal with engineers or scientists or programmers. I deal with managers and DBA's who want to know how much and how fast.

So regardless of the "correctness", it is easiest for me to say "an 800 Mhz itanium costs X dollars and a PA RISC costs Y, and they are equivalent, or half as fast, or twice as fast"...etc.
Thanks for all the replies. Keep them coming.

[A. Clay Stephenson]

This is essentially an impossible question to answer with anything other than "it depends". The vast majority of Oracle applications are i/o bound, then memory bound, and finally CPU bound. You could very well find that even if you were comparing a relatively slow PA-RISC based box to an Itanium box that as long as the I/O subsystems and memory were comparable, the actual performance differences might be very small. I suggest that you use Glance to the performance tool of your choice to see how much CPU time your "typical" Oracle processes are using and apply a rather safe 1.5X improvement. If the total fraction of the process time is 15% then the actual performance gains are quite small.

Obviously, if this were a CPU bound application (e.g. Finite Element Analysis) then you would see huge differences and the comparisons are easy but in applications where I/O is the bottleneck the CPU's effect may be difficult to measure --- and harder to justify.

The very last thing you want to do is "prove" that the Itanium box is twice as fast based on benchmarks only to find out that 2X is really only 1.05x because the CPU is such a small component. Many managers (and unfortunately many IT staffer's) have bought into the PC mentality that the only thing that matters is that "I have more GHz than you".

[Julio Yamawaki]

Hi,

Let me correct some data:

rx2600 box was powered by a 1,3 GHz Itanium, so, I was checking that the power im TPM-C is about the same as a rp3440 800 MHz: 29000 tpm.
Also, I put rx2600 in production in february, 2004 and rp3410 in november, 2004.
So, Itanium needs more GHz to deploy the same TPM.
In my opinion, when executing long running queries, rx2600 has a better response time than rp3410, of course, with Oracle 9i database.
Also, I'm starting to upgrade some database to 10g and in a few months I will have some real world data about this databases in production in rx and rp box and I will send this for you.

Regards,

[Tony Drake_2]

Julio: thanks. I look forward to it.

Clay: good points. I did not realize that CPU for Oracle was such a minimal component. Most of my servers tend to be pretty heavily loaded.

I will keep that in mind as I size servers and start deploying applications.

[Jeff Schussele · Accepted solution]

Hi Tony,

I have to whole-heartedly agree with Clay here. We have dozens & dozens & dozens of Oracle servers here and they're all different in the demands they present. But we've found that THE common bottlenecks - in order - are I/O then memory & finally CPU.
Our biggest pig DB was not "tamed" until we threw an rp8400 with 8 - yes EIGHT - 2GB fibre channel cards for "normal" I/O & 2 more for the hotbackup - yes we *had* to move that traffic off the "standard" fibre channel cards.
Now the battles we fight are getting the DBAs to optimize table/index layout & more importantly tighten up their sloppy SQL code.
Initially, when we ran 32-bit Oracle, we constantly fought the inherent memory constraints. But now that we're all 9i & living fat, dumb & happy in 64bitville we don't have that problem near as much. But then again when extremely sloppy SQL code rears it's ugly head we can see "some" CPU pressure...BUT....we also see monitoring routines kicking in at the same time. Gotta love those developer-types - their response to application degradation is to fire up *more* processes to "monitor" the situation.
Heh, I guess it's a job security kinda thing for them.
But the Oracle moral is regardless of CPU architecture put *more* money into I/O than CPU & you have a better chance of meeting SLAs my friend.

My 2 cents,
Jeff