back to article Ellison whips out his Sparc TPC-C test

Larry Ellison, Oracle's chief executive officer, has ants in his pants. Or something. Instead of waiting until October 14 to roll out the official benchmark tests on Sparc-based servers that show a Sparc cluster can scale farther and cost less than a big Power 595 SMP server from IBM - as the company had been promising in ads …

COMMENTS

This topic is closed for new posts.

Page:

FAIL

I expected better from Larry and company

Scootter's top ten list for the new benchmark

1) Woops not double digit....I guess that is why we are not supposed to guesstimate performance

2) The picture in the WSJ only showed one rack...woops need two for 12 systems

3) 384 cores with 3,072 threads...wow what are the odds an application will scale like that

4) T2+ price drop to .5 ( Sorry but I thought the "Oracle core factor table" was to equal performance per license? Why not drop all T processors?

5) 3 year term licenses who buys them (The Oracle price should really be double which is what perpetual licenses cost vs. 3 year) would the real cost per $3.60?

6) Doesn't flash have reliability issues?

7) Does anyone actually buy the T2 systems for database? Flash does not change their reliance on round robin of light threads.

8) Duh..Yes less energy but that is true of solid state vs. disk drives.

9) IBM software costs $2.3M while Oracle is $7.8M for three year term but for perpetual licenses its $15.6M

10) Sun will never do another bogus benchmark...because Sun is history :-(

0
0
Happy

Ahhh... TPM... did Oracle poo on your IBM?

I would be much more interested in a SMP solution, but the cost difference is not to be sniffed at. If IBM decides to cluster together multiple 595's then the price will escalate to the extreme. If IBM want's to compete with this solution they need a smaller box with the same scalability of the CMT boxes that Sun sells.

0
0
WTF?

Congratulations Oracle/Sun

384 Sparc T2+ cores are indeed faster (by whole 26%) than 64 Power6 cores released 2 years ago. Oracle and Sun should be proud of this.

If you calculate Oracle licenses, even with new 0.5 multiply factor for T2+ cores, it's still 192 licenses (+RAC option) versus 64. Good luck!

0
0
Silver badge
FAIL

Someone needs to talk to the right people at IBM...

"But in practice, IBM has also announced its own clustering technologies for online transaction processing--that would be the DB2 PureScale extensions for DB2 running on AIX atop Power-based servers announced last week and shipping in December, and a similar AIX-Power cluster called the Smart Analytics System announced in late July that is tuned for data warehousing and analytics on mountains of data."

Hmmm.

When IBM bought Informix, they bought 3 database platforms.... IDS 7.x , IDS 9.x, and XPS (8.x).

XPS was the data warehouse platform which was a shared nothing distributed RDBMS model.

XPS is still available to customers and rumor has it that there are still customers running XPS.

Why IBM refuses to recognize that they own Informix Dynamic Server (IDS) and that its for sale, is beyond me. There's a certain retailer based in Arxxxxs that runs IDS in all of their stores and they are controlled by a small team. (Rumor has it at 8 man team of DBAs for 15K database servers.)

I guess they want to keep it a competitive sekret.

Rate this a fail, not just because Oracle is yet again playing with the numbers, but also IBM for not recognizing IDS and XPS.

0
0
Big Brother

Just an expensive lease.

As the First AC poster pointed out the licenses are 3 year limited licenses. So you don't buy the software you license it, for 3 years. If you were to buy the software it would cost you 192 licenses times 82.000 USD=15.744.000 USD+10391040 USD=26.135.040 USD List price. - Discount.

And notice there is not any normal Oracle Support offering on this

http://www.oracle.com/corporate/pricing/technology-price-list.pdf

On page 8 it clearly states that on 'term licenses' you still pay 22% for your normal software license update and support.

Oracle Incident Server Support Package is basically metalink access.

http://www.oracle.com/support/collateral/oracle-technical-support-policies.pdf

So basically on the software side on this you get to lease the software for 3 years with Metalink as your support.

I mean that is just laughable.

And putting up 384 cores and only getting 7.7M tpmc is not impressive.

And the per thread throughput is simply awful, I mean 384 cores x 8 =3072 threads, that gives you a staggering 2512 tpmc per thread. That less throughput per thread than results using 200Mhz PentiumPro's from 10+ years ago.

I am not impressed.

// Jesper

0
0
Bronze badge

RAC Scaling

And maybe time for a new OLTP benchmark...

Personally, I tend to think the SAP benchmark is a better guide to real world application performance than TPC. I suppose TPC'C' can be used as an alternativate to iometer if you want to thrash your I/O system, but this sort of configuration truly bears no resmblence to any real system I know of. We have very large systems but barely anything with 10% of that amount of disk space on a single application (and it's not all DB storage either).

As far as RAC goes, it's great for data warehousing in my experience, it's good for availability, but if you want to scale a very large OLTP workload you are into some serious application tuning. Quite simply, an SMP is easier to use and if you have a few very hot single-threaded components to your app it will suffer really badly on the T2.

The flash story is really going to be imporant in the future. These benchamrks don't major enough on response times to my mind. TPC'C' is an old-fasjhioned type application that has fairly simple transactions. We struggle to minimise reponse times in call centres and the like on complex multi-phase applications, and it's not uncommon for database transactions to spend three quarters of their time on I/O, even with very extensive caching. As for reliability of flash - firstly no enterprise in its right mind will run large scale mission-critical systems without RAID support. mechanical disks are not exactl;y 100% reliable themselves. Secondly, for those that worry about write-endurance, large enterprises will have support contracts - within the useful lifetime of a large array (maybe 6 years or so), then replacements are covered by that.

There's a lot more required on flash - hardware compression, to optimise the space, automatic methods of data management and so on. But it is soming and it will eventually dominate the market for high thropughput OLTP database storage.

0
0
Silver badge

Clustered TPC-C is a silly benchmark

Even M$ have stopped posting results. All the top TPC-C scores used to be for clustered M$SQL systems. The way the test works should scale perfectly across a cluster so the result you score is just a function of how deep your pockets are to build the test system.

If Oracle are cutting the price of their licenses on SPARC they could well find that the merger is blocked by the competition authorities. Oracle have grow rich by being vendor neutral, A lot of Oracle licenses are sold to run on IBM kit. If Oracle buy Sun and then change their pricing structure to favour their own products they are breaching the anti-trust rules.

0
0
Silver badge
Happy

Laughing at the IBMers for a change.

Sorry, but it is funny. First off, congratulations where they are due, Snoreacle have managed to produce what is currently the number one TPC-C result. The fact that it is just as irrellevant as the previous record results is neither here or there, it is still the number one for now.

I am very amused at the posts grumbling at how the Sun kit is not a real World build (some sore losers in the IBM camp, maybe?), but then neither were any of the previous ten-odd top results. Sure, Larry and co have short-stroked the CPUs (only 2512 tpmc per thread across 384 cores is pretty pathetic), but how is that different to when IBM massively partitioned their database and short-stroked their disks on their benchmarks? Effectively, the whole TPC-C benchmark is the equivalent of drag-racing, with the dragsters used being special-purpose designs you're not going to see on the highstreet. So, it is a bit rich for the IBM pots to be grumbling about the Sun kettles. Sure, there is no way I could take the Sun design into the boardroom and expect not to get laughed out of the building, but then there was no way I could have taken IBM's previous TPC-C champion build in there either. From a practical perspective both represented very poor value for very, very large amounts of cash.

What should be worrying the Oracle supporters more is that this is again starting to show a pattern of what could be, at best, described as "unfair" marketing. Others might just call it downright dishonest. As noted above, Larry managed to forget the decimal point (hmmmm, I'm sure it just slipped his mind!), and their ad did imply they would be beating a single IBM system with a single Sun cluster, not two racks of servers. One of the factors in Sun's fall from grace was the over-hype and over-aggressive sales tactics, that led to many Sun customers being disappointed with what they actually got. Larry seems to have started his drive into hardware with a rather dishonest campaign complete with a public rebuke from TPC for - effectively - lying. That's not going to generate much customer confidence in those previously burnt by the Sunshine.

/Even more P&L!

0
1
Coat

As the old saying goes ..

There are lies, damn lies, and IT statistics

0
0
Megaphone

The real price of Oracle on 12 T5440 is $19M!!!!

12 Systems

384 Cores

192 Oracle Licenses

Oracle EE cost: $6,840,000

Oracle RAC cost: $3,312,000

Oracle Partitioning:$1,656,000

Three year maintenance: $7,793,280

Total 3 year Oracle cost: $19,601,280

0
0
Go

Some takeaways... Let the horse races begin!

I also liked this blog on the topic.

- http://blogs.sun.com/BestPerf/entry/tpc_c_world_record_sun

Things that caught my eye:

- Order of magnitude fewer racks to produce a faster solution

- Order of magnitude fewer watts per 1000 tpmC

- Sun's 36 sockets to IBM's 32 sockets

- 10 GigE & FC instead of InfiniBand

- Intel based OpenSolaris storage servers, instead of AMD "Thumper " based servers

Some thoughts:

- The order of magnitude improvements in space and power consumption was obviously more compelling to someone than shooting for an order of magnitude improvement in performance

- The performance could have been faster by adding more hosts to the RAC configuration, but the order of magnitude comparisons would be lost

- The cost savings for superior performing SPARC cluster is dramatic: fewer hardware components for maintenance , lower HVAC costs, lower UPS costs, lower generator costs, lower cabling costs, lower data center square footage costs

- The pricing per SPARC core (noted in the Reg article) is still to high for the T2 and T2+ processors, in comparison to the performance with competing sockets

- The hammering by a few El Reg readers about the CoolThreads processors not being capable of running large databases is finally put to rest

It would have been nice to see:

- a more scalable SMP solution, but this solution will expand better in an IBM horse race

- a full Sun QDR InfiniBand configuration

- a full end-to-end 10GigE configuration

- the T2 with embedded 10GigE clustered instead of the T2+ with the 10GigE card

Let the horse races begin!

0
0
Happy

@Anon: I expected better from Larry and company

Anonymous Coward Posted Monday 12th October 2009 20:20 GMT

~ 1) Woops not double digit....I guess that is why we are not supposed to guesstimate performance

Whole digit savings on power and rack floor space

~ 3) 384 cores with 3,072 threads...wow what are the odds an application will scale like that

Nearly the same number of sockets - who would have thought a Sun socket would out-perform an IBM socket

~ 4) T2+ price drop to .5 ( Sorry but I thought the "Oracle core factor table" was to equal performance per license? Why not drop all T processors?

Pricing on the T1 is pretty good, the T2 core licensing is overpriced by Oracle

~ 5) 3 year term licenses who buys them... 6) Doesn't flash have reliability issues?

Items 5 & 6 are probably related. Implementing ZFS mitigates flash reliability issues.

~ 7) Does anyone actually buy the T2 systems for database? Flash does not change their reliance on round robin of light threads.

If T2 out-performs IBM POWER and Intel Itanium, socket per socket, why wouldn't someone run T2, especially in an Oracle Standard license? Looks like "round robin of light threads" is a winning innovative combination. IBM and Intel couldn't figure it out, years ago!

0
0
Anonymous Coward

TPC-C and Flash SSD

Ok...Oracle's price/perf. number is merely a result of tweaking the per-core sortware license pricing. But since this is the first ever TPC-C result based on Flash-SSD, I thought maybe pulling out the software costs and examining just the hardware performance might still yield some useful information.

So, for 7.7 million transactions-per-minute we have:

- CPU: 192cores@1.6GHz = = 12.5K TPMc per core*GHz

- DRAM: 6TB = 1.3M TPMc per TByte

- DIsks: 4,800 SSDs = 1,600 TPMc per Disk

Based on historical TPC-C system trends, the CPU number about right for a cluster, considering the ~30% penalty from clustering overhead. On the SMP side, IBM gets around 19K per core*GHz on , and Oracle's last result on HP Superdome was almost identical at 20K.

(interesting to note that if you divide the 12.5K number by the 20K number, you get exactly the amount of Oracle's T2+ price cut last week).

Likewise, the DRAM number is right in line considering that clusters are also more DRAM hungry than SMP, the Sunacle cluster needs about 30% more DRAM per TPMc than either Oracle on HP Superdome or DB2 on Power 595.

On the other hand, the "disk" number -- using SSDs -- seems really bizarre. Oracle's most recent large scale results (indeed all TPC-C results at this scale) get around 600 TPMc per disk, using 15K RPM HDDs. At only 1,600 TPMc per 'disk' (and at $2,000 each), these units are only 2.6x faster than a $250 spinning disk.

Sun claims each of these units can do 18,000 IOPS, 85x faster than 15K HDD. How do we go from 85X faster to two-point-six times faster?

Since the Sun SSD costs $83,000 per TByte, I thought using 450 x 1TB SATA disks to cover the TPC 60-day space requirements was a clever workaround.

What I can't understand why the system STILL needs 384 x 15K RPM disks. I guess this would only make sense if the SSDs were actually slower than HDDs on writes, and the HDDs were needed for the redo logs.

0
0
FAIL

Re: @Anon: I expected better from Larry and company

> Implementing ZFS mitigates flash reliability issues.

Really? But according to Oracle, ZFS is not supported with Oracle RAC.

0
0
Anonymous Coward

T2+ is a "network facing" chip not for databases <=Sun quote

Whole digit savings on power and rack floor space.

===>Solid state vs. lots of disks to make sure the heads on on the right platter

Nearly the same number of sockets - who would have thought a Sun socket would out-perform an IBM socket.

===>Welcome to benchmarks with solid state drives. A p595 can easily beat 12 sparc systems with SSD

Pricing on the T1 is pretty good, the T2 core licensing is overpriced by Oracle

===> looks like even Oracle will

Oracle is not dropping core pricing on T2 to where it should be...sorry customers

Items 5 & 6 are probably related. Implementing ZFS mitigates flash reliability issues.

====> Flash has a very limited write capability so is not good for high transaction rate systems...unless it is a benchmark

If T2 out-performs IBM POWER and Intel Itanium, socket per socket, why wouldn't someone run T2, especially in an Oracle Standard license?

===>isn't this about EE? Standard only goes to four sockets and has very limited software which is available

0
0
Thumb Up

RE: T2+ is a "network facing" chip not for databases <=Sun quote

Anonymou 2 posts, "T2+ is a 'network facing' chip not for databases <=Sun quote"

Well, if the T2+ scores the highest TPC-C benchmark recorded in the world - I guess Sun was underestimating the capability of their own processor.

Refreshing to see a vendor who does not over-hype their product!

Anonymous 1 posts, "Nearly the same number of sockets - who would have thought a Sun socket would out-perform an IBM socket."

Anonymous 2 posts, "A p595 can easily beat 12 sparc systems with SSD"

I think it will be nice to see the IBM benchmark! I am more interested in seeing how IBM licks the write limitations of flash - I expect it will be different from the way Sun architected around the limitations of flash. Competition is good for all system customers!

Anonymous 1 posts, "Items 5 & 6 are probably related. Implementing ZFS mitigates flash reliability issues."

Anonymous 2 posts, "Flash has a very limited write capability so is not good for high transaction rate systems...unless it is a benchmark"

This is true, unless you have ZFS. ZFS architects worked around this issue. This is the beauty of ZFS, and why ZFS makes flash so disruptive, even in a high transaction rate system.

Anonymous 1 posts, "If T2 out-performs IBM POWER and Intel Itanium, socket per socket, why wouldn't someone run T2, especially in an Oracle Standard license?"

Anonymous 2 posts, "Isn't this about EE? Standard only goes to four sockets and has very limited software which is available"

Clustering is available using Standard licensing with up to 4 sockets. The T2+ with 4 sockets is able to cluster without EE. This means, you can get a substantial performance boost via clustering with T2+ without EE, for most standard Oracle applications.

http://netmgt.blogspot.com/2009/03/oracle-database-license-change-ibm.html

The limitation of the software products is not that substantial for common Oracle database users.

1) Few people use partitioning - some enterprise applications that I have used will actually do the equivalent of partitioning, at the application layer, so you don't have to buy EE for that option.

2) Compression can be done at the ZFS level, instead of the Oracle level (to save I/O, flash, and disk utilization).

3) Scalability can be done by boosting 1x4 socket servers to nx4 socket servers, etc. Under Solaris 10, Capped Zones can be legally used to limit the number of CPU's used by the Oracle database, offering other processors for applications on the same platform. Multiple capped zones can be used on the same SMP platform in a RAC cluster, to avoid EE pricing penalty, consolidate hardware, and add a level of RDBMS redundancy.

I have not been forced to run any applications that require EE in any enterprise under Solaris, unless there was some company restriction forcing people not to use capped zones.

0
0

Impressive!!!

Compare to the earlier top result.

IBM used 76 racks vs SUN used 9 racks!! That is mighty impressive. IBM used almost 10 times more racks to get a lesser result. And look at the Wattage consumption of IBM compared to SUN.

And regarding cost, only the P595 machine itself costed $17.1 million. Jesus. And it needed 32 CPUs with the "Uberfast 4.7GHz Power6+". Now we are talking about really big costs.

The SUN solution used 48 CPUs with the 1.4GHz Niagara to get a 25% better result. How can that be, if the Niagara is so slow, which FUDers say? If the Niagara would be so slow, it would be impossible for it to beat the three times higher clocked Power6? Someone is wrong on this.

Some web pages say SUN used 36 sockets. If there were 12 of the T5440 then it should be 48 sockets. So 36 sockets are wrong.

http://blogs.sun.com/BestPerf/entry/tpc_c_world_record_sun

.

Ive earlier told you FUDers that SUN could have easily won the TPC benches if SUN wanted too. But SUN thinks these TPC benches are meanless. Here is proof that I was right all the time. But still, I dont like TPC benches, as I feel they are artificial. Who can afford $17 million for the P595 machine? It used several thousand short stroked discs. Who uses that in production? These TPC monster machines are not really representative for real work. So why? I understand Oracle wants to show off the synergy effects of acquiring SUN, so that would be an acceptable reason.

Anyway. All your TPC results are belong to Oracle + SUN.

0
1
Anonymous Coward

David Halko and Kebabbert finally show....too bad they are impressed with what they hate

Too bad Oracle RAC still can't scale past 4 nodes for OLTP (unless of course its a highly tuned benchmark)

We have dumped RAC for everything we can.....failover is easier and with virtualization it does not waste resources.

Oracle gamed these results. Make the software stack perpetual and add real maintenance fees of 22% vs. just metalink access and the price is over $4 per transaction. This is significantly more than the p595 result.

A p595 might cost more than 12 4socket boxes but its a single SMP (easy) and has a lower TCO.

Add in the value of virtualization which most customers use vs. one TPC workload and its business value is stellar!!

0
0
Big Brother

RE:Kebabbert

You don't get it do you.

This system is actually quite expensive, when you realize what you are getting.

You pay 8 MUSD in software licenses to run a complex database on some entry level servers for 3 years. When those 3 years are gone you will pay another 8 MUSD, to run this system for another 3 years. Who is stupid enough to do that.

The real price of the software licenses if you were to actually buy them is almost 20MUSD in list price with 3 years of real support, (you only get Metalink access in the benchmark as your support)

So to answer your question "Who can afford $17 million for the P595 machine?"

One that is clever enough to buy a power 595 with a real software license and real support.

Again as it often is with you you forget to check your facts.

If you look at the reports for the the power 595 and the T5440 benchmark reports you will see that the price of the actual power 595 server + the server software is less than that of the 12xT5440+server software, even though you are buying time limited licenses on the T5440. If you bought normal licenses and software support the T5440 hardware + software licenses would be double that of a power 595 + software license.

So all this benchmark has proven is that it is in real life it would much more expensive to run your workload on RAC running on a bunch of T5440 compared to running it on a big tin SMP box.

Do the math

http://www.tpc.org/results/individual_results/Sun/Sun_T5440_TPC-C_Cluster_ES_101109.pdf

and here

http://www.tpc.org/results/individual_results/IBM/IBM_595_20080610_ES.pdf

// Jesper

0
0

@Jesper Frimann

What dont I get? That I dont like TPC benchmarks, as all such systems are pathological and no one can afford them? How many times have I said that? Who can afford $17.1 million for the IBM P595 machine? No one. SUNs hardware is also expensive, but so what. These benches are not realistic. IBM P595 used several thousand short stroked hard drives, who does that?

And for the sucky P595 which scales bad. Each CPU 4.7GHz Power6+ uses huge amount of power. Have you seen the power consumption in the link Ive posted above? Is 500-600 watt per Power6+ CPU realistic? The 1.4GHz T2+ uses ~110 Watt and it outperforms the 4.7GHz Power6. It is funny how bad characteristics the Power6 has. Uses 5-6 times(?) more power AND uses 3 times higher frequency, and still the Niagara is 6 times faster on Siebel v8 benches. And also faster on TPC-C.

.

http://www.c0t0d0s0.org/archives/5941-The-new-TPC-H-benchmark-from-IBM.html

"Almost 4 years ago IBM published a result of the p5-595. This system yielded with 1.9 Ghz Power5 CPU 100,512.30 QphH. 4 years of development, 256 GByte more memory and 2.6 times the frequency gives you just roundabout 50% more performance[154,115.8 QphH]. Nice, but not that impressive, especially given the effort put into this CPU in regard of cycles."

From 1.9GHz Power5 to 5GHz Power6 gives you 50% more performance. I call this BAD scaling. Who doesnt? The comments are also fun. He shows that the Power7 is slower per core, than the Power6.

.

Some people dont get it. Who is fastest right now? How can that be, if the Niagara is so slow? Facts show that some people here are lying. Say what you want about the Niagara, lie how much you want, but Niagara has the performance crown.

And for you FUDers who say that Niagara doesnt scale beyond 4 sockets and therefore produces inferior solutions: well you are WRONG as evidenced by this article. The point of scaling is to offer more performance. But what if you are the fastest on the planet? Then what is the point of complaining "it doesnt scale"? It is already the fastest! It doesnt need to scale beyond this! It is fastest on earth. And if you really need to scale, just add another T5440. That works good as we can see from this article.

And still, the T5440 is $76.000 and one cheapo IBM P570 is $413.000. In Siebel v8, IBM needs 6 of the P570 to match one T5440. That is bad in my opinion.

0
0
Alert

Some thoughts ...

Some thoughts to share with you guys ....

SUN benchmark:

7,717,510.6 tpmC

Bench date: October 11, 2009

Servers : 12 servers T5440 giving a total number of 384 cores -> that means 20098 tpmC/core

Clients: 24 clients using Intel Xeon E5540 giving a total of 96 cores (they used machines with one Quad core CPU per machine)

Storage:

4800 x 24GB SSD Flash Modules

-> this is a good choice because this is GREEN and you don't have the same actuators limit you may have with standard disks. If those disks can be used as a kind of cache with the most read data, this is cool. If used as a kind of write cache (to avoid the actuators limit), this is less cool because you may have to replace it (SSD write reliability is not that good). check this out: http://www.itbusinessedge.com/cm/blogs/cole/ssds-and-the-reliability-factor/?cs=36187

384 x 300GB 15K rpm 2.5” SAS

24 x 300GB 10K rpm 2.5” SAS

449 x 1TB 7.2K rpm 3.5” SATA

As you know, on storage level, in order to deliver enough I/O's per second you need either enough amount of disks (arms) or you need a kind of memory (being SSD) to avoid queuing on disk level and performance degradation.

IBM benchmark:

6,085,166 tpmC

Bench date: 06/10/08

Servers : One server with 64 cores Power6-> that means 95080 tpmC/core

Clients: 128 client machines Intel Dual-core Xeon 2.0GHz with a total of 256 cores -> this is an old Intel technology because at the benchmark time, the new Intel Xeon (Nehalem) CPU where not available !

-> this is very important because today, with the new Intel processors, we could probably estimate the amount of client to 4 times less (according to the last Intel benchmarks). I'm giving you this information because if I red the bench well, this is part of the price/performance calculated price.

Storage:

8 x 146.8GB 15K RPM SCSI disks

10992 x 73.4GB 15K RPM 4Gb FC disks

-> this is NOT GREEN but reliable.

As a conclusion, I would say, if you compare the amount of server core/tpmC, for sure SUN is not good (1/4 of the IBM performances). To my opinion, that's why they are clustering the servers. Even Intel Xeon processor (nehalem) are much more faster than SPARC.

On storage level, SUN used the right config if you think GREEN but I'm not convinced this is the best config in terms of reliability.

There is one aspect they don't discuss in the tpc-C benchmarks in general: what about the management costs (firmware upgrade, OS patches, ...) ? (24 servers <> 1 server)

SUN made a better benchmark one year after IBM's one and gained about 25% more performances -> to me, this is logical due to the fact the technologies are better and better. I would not expect something else.

Concerning the comments I red here, I think you are right, nobody will buy such a non realistic config (whatever it's an IBM or a SUN one) but if you need to buy one small server giving you the best performances, for sure, you have to select Power technology (according to the tpmc/core) . There is no doubt about it. (Check the SAP benchmark or any other transaction based benchmark, you will get the same figures). If you don't like Power or IBM for one reason or an other you'd better buy Intel boxes than UltraSPARC T2 !

If you are buying a server for a non transactional workload, this story may be different but this needs to be checked.

The only reason to buy SUN (to me) is because you are addicted to Solaris :-) or because you are going to run a Web application (non transactionnal applicaton in general) -> in that case, you can use all the threads the processor is providing without being at the I/O system limit.

And concerning the price/performance, I'm not sure a one year ago price is still accurate today ;-)

Cheers.

0
0
FAIL

Anonymous: too bad they are impressed with what they hate

Anonymous Coward Posted, "Too bad Oracle RAC still can't scale past 4 nodes for OLTP (unless of course its a highly tuned benchmark)"

Seems like you missed the announcement of Exadata 2. They just bolt on a new node for OLTP.

http://www.ideationcloud.com/2009/09/17/exadata-2-oraclesun-announce-fastest-oltp-box-on-the-planet/

Anonymous Coward Posted, "Oracle gamed these results."

Oracle released platforms under 2 different reference Sun architectures. I guess Sun could do what you could not do with whatever architecture you were deploying. Too bad you couldn't figure out how to install Oracle RAC.

Who does one believe, an "Anonymous Coward" who couldn't get Oracle RAC to work or a vendor who released 2 different reference architectures and offers a 10 million US$ challenge?

http://dcsblog.burtongroup.com/data_center_strategies/2009/10/oracles-exadata-2-results-and-10m-challenge.html

I bet on Oracle over an Anonymous Coward who couldn't figure out how to install Oracle RAC.

0
0
Anonymous Coward

All this noise says nothing about Sun, Oracle or IBM

Does anybody here really think that IBM won't be able to manage a piddling 26% performance increase when they too plug in SSDs in place of disk?

The only thing we've learned from any of this is:

1) SSDs are less than 3x faster than HDDs when benchmarked on a real application workload. So much for "replacing dozens of spinning disks" with one SSD. So glad that nonsense has finally been dispelled.

2) Price/performance does not improve by switching from HDD to SSD. In fact, without the Oracle software "benchmark special" price reduction, the price/performance of this system would be worse than if it had been built with HDD.

0
0
Big Brother

RE:Kebabbert

*SIGH*

You accuse people who don't agree with you of being liars and spreading FUD.

But remember when you point at someone there are 3 fingers pointing back at yourself.

Lets take one of your claims:

POWER6 uses 500-600 watt per CPU, and with CPU I presume you mean CHIP, hence a

Now doing the math this would mean that just the CPU power usage would make a power 550

use 2400 watt. So why don't we check out your 'facts' with one of the tools that are

actually used where I work to calculate power usage.

http://www-912.ibm.com/see/EnergyEstimator

A fully loaded power 550 with

8x 5GHz cores

256 GB of RAM,

DVD-RAM,

Tape,

Dual 10Gbit,

6x15K SAS disks

5 PCI Cards

uses around 1429 Watts.

And if we strip away all the RAM and other stuff then it uses 878 Watt.

If we use the 3.5GHz Power6 then we hit 738 watt, hence going from 3.5 GHz

to 5GHz uses 35 Watt extra per Chip, as the machine has 4 chips.

So can you answer me why the p550 isn't using 3000-3500 watt as it should,

if your numbers were right ?

And another one.

"And for you FUDers who say that Niagara doesnt scale beyond 4 sockets and therefore produces inferior solutions: well you are WRONG as evidenced by this article."

You don't get it, do you ?

This benchmark is done using Oracle RAC, it is a cluster. You could have done the same with a processor that didn't support SMP. Hence your argument

that cause you cluster a bunch of 4 socket T5440's using RAC then it proves that the T2+ processor scales beyond 4 sockets is.. well.. not valid,

I would even go so far as to call it ignorant.

And last "Who can afford $17 million for the P595 machine?"

You didn't bother clicking on the links I posted for you. This benchmark is all about selling software licenses.

Server price:

Now the price of the power 595 is 5.833.031 USD with 3 years of support.

Now the price of the 12xT5440 are 2.113.991 USD with 3 years of support.

So the T5440 are a good deal cheaper as expected. But we need software to run on the benchmark.

Now the price of the Ora software is 7.954.800 USD with 3 years of support

Now the price of the IBM software is 1.277.477 USD with 3 years of support.

So for the physical servers and the software the cost is.

Ora price Servers + Server Software is 10.068.791 USD

IBM price Server + Server Software is 7.110.508 USD

The reason why oracle's price end up being lower for the whole solution is that they use flash drives, something that wasn't really available when the power 595 benchmark was done.

What you can see is that the Oracle software costs more than the Power 595 + IBM software. And this is the whole reason why Oracle

is pushing Niagara, based clusters. To sell more software.

And if you look closer at the Oracle prices then you will notice that you only get to lease the software for 3 years. After that you

have to buy yet another lease. And the software support is Metalink (web only).

If you were to buy the software the price would be around 20MUSD - Discounts.

Furthermore the IBM software stack also has 3 Tivoli magement products priced into the solution. ( Management, TADDM and accounting)

Again you can't back up your claims, this is actually getting a little boring.

// Jesper

0
0

@Jesper - liar/ignorant

Do you have prolems understanding text, or do you lie deliberately? Where am I "not backing my statements up"? If you are going to spread FUD or lie about me, then you should be able to prove your claims. Otherwise you are lying, or FUDing or are ignorant. Dont try to spread FUD about me. OTOH, there are FUDers here. They claim that Niagara suffers from a too small a cache and therefore is not fast. Now, if that is FUD I dont know. Who is fastest right now? How is that possible if the FUDers are correct?

Regarding your lies about me, you have to prove that I lie, otherwise it is YOU that is the liar and FUDer. Because I dont lie, and you can NOT prove that I lie, so therefore you can join the ranks where Mattie Pattie Laddie is. Because you are lying about me. And I shall now prove that you lie about me. And, you can not prove that I lie. Try it, I dare you. Who is the liar, then?

.

Proposition: Jesper is lying about me.

Proof. Jesper claims that I can not back my statements up, and he claims I have written similar false things several times earlier. He writes:

"Again you can't back up your claims"

He then argues that the Power6 does not consume 500-600 Watt. And presents some numbers showing that the Power6 does not consume 500-600 Watt. And draws the conclusion I am wrong. Ergo I lie.

I have not accepted your calculations. Maybe they are wrong maybe not. Maybe the Power6 really uses 500W, but youre calculations are wrong?

Well the point is Jesper Fresper, that I never claim that Power6 uses 500-600 Watt. I am ASKING. I write "Is 500-600 watt per Power6+ CPU realistic?". I dont claim it is true, I ask and I speculate. Earlier Ive asked why IBM refuses to publish Power6 power usage, and I asked if the Power6 uses 500 W? Check every post from me about this and you will see that I never claim that Power6 uses 500W. I am merely asking and speculating - which I make clear. I use question marks, etc. But you missed that, obviously.

Now, I dont claim that Power6 uses 500Watt. So dont come here and say that I - again - doesnt back up my claims, because I claim nothing. I dont have to back up anything. And dont try to imply that I earlier times have claimed false things. If you want to claim something about me, then you prove it. I strongly dislike lies.

Ergo, Jesper Fresper is lying about me. Stop it, ok? Or, prove that I have lied. Good luck. And, then you must prove that I lied repeatedly. Good luck again. If you can not prove youre claims, then stop lie.

.

.

Regarding the pricing of the TPC-C setups. Ive read that the single IBM P595 machine costed $17 million USD, on IBM's TPC benchmark page containing information about the configuration IBM used. Why do you claim the P595 for the TPC cost $8 million? Are you lying again or is this pure ignorance? Or confusion from your side?

And regarding all your pricing talk, so what? Why do you calculate the pricing? What are you trying to prove? Are you just confused? For your information, I have never claimed that SUNs solutions is cheaper or more expensive than IBM's. I have only talked about the performance. I wanted to show that the people here saying that Niagara is slow, are wrong. And this TPC score proves I am correct. I dont care about the pricing. I only care about performance, just to prove that Niagara is fast. In fact, maybe you missed it, but I am very clear about the pricings. That no one can afford these computers, and therefore they are not interesting setups. So why do you talk about pricing with me? I am not interested in discussing pricing. Confused? Are you often confused?

0
1
Anonymous Coward

More analysis about this benchmark

http://www.c0t0d0s0.org/archives/6015-Some-takeways-from-the-SunOracle-TPC-C-benchmark.html

http://www.c0t0d0s0.org/archives/6012-IBMs-reaction-to-SunOracle-TPC-C-announcement.html

0
0
Big Brother

@Kebabbert

Dear kebabbert

Listen I stopped flaming people on the internet 15 years ago cause life is way to short to use it on bringing unhappiness to people. And I'm not going to start now. Although I must admit you last post here did tempt me a tiny bit :)=

Making fun of my name is cute btw.

I would never spread FUD about you, I simply said that you have a nasty habit of calling people FUDders and liars,when they don't agree with you. Now I have never called you a liar, I have simply stated that you were wrong. There is a difference, lying is stating something that you know is wrong, and I don't know if you are lying.

You seem to ask the "If POWER6 uses 500-600 watt ?" question again and again.

You seem to have a habit of asking that same question anywhere you can sneak it in. Here are 3 links beside this debate:

http://www.theregister.co.uk/2009/07/21/ibm_power7_details/comments/#c_543822

http://www.theregister.co.uk/2009/07/22/sun_sparc_t_crank/comments/#c_543769

http://techworld.idg.se/2.2524/1.242411/osaker-framtid-for-suns-16-karniga-processor?articleRenderMode=listpostings

And such questions might get you cheers, at a SUN xmas lunch, but not here.

But at least now I have directed you to where you can find an answer to your question so now you don't have to keep asking it. Right ?

And for your whole little 17 Million for a machine, let me correct you yet again, it is the price for the whole power 595 benchm arked configuration. With software, disks, servers and all. So it's not the power 595 server. So now you know that.

And I cannot understand your claim that nobody can afford these machines. Where I work we have a lot of M9000,M8000,Superdomes,p595 and power 595. Well it's not ours all of them, many of them are owned by our customers.

And I still don't think that the T5440 is so stunning. 7717511 tpcc for 12 machines gives 643126 tpmcc per machine, now the old 4.2 GHz version of the power 550 does 629159 tpcc. So I gues it's a pretty good bet that the 5.0 Ghz version would do 700000+ tpcc.

// Jesper Fnesper Lesper Desper Klesper whatever.

0
0
Happy

@Jesper Frimann RE:Kebabbert #

Jesper posted at 09:36 GMT, "The reason why oracle's price end up being lower for the whole solution is that they use flash drives, something that wasn't really available when the power 595 benchmark was done."

IBM used racks of drives in an unorthodox way, that would never typically be used in a production environment, to gain a benchmark advantage.

I think your defense is quite good, in this case, indicating that Sun & Oracle may be using new technology in an unorthodox way, to gain an advantage.

I think your argument weakens in one space: other vendors could have used racks of hard drives in a similar way to IBM to compete, but they did not -- Sun & Oracle have invested in innovate software (Oracle 11g, OpenSolaris, and ZFS) technology to bring a new technology (Flash) to existing markets (Databases through Oracle 11g), as well as general application performance enhancements to operating systems (Solaris and OpenSolaris), as well as performance enhancements to other non-native applications and competing vendor operating systems (through OpenStorage.)

In short, the technological way that IBM used to achieve it's benchmarks could have been emulated by others, but it was not because it was cheezy... and the technological way Sun/Oracle used to achieve their benchmarks will be emulated by others application vendors, can be leveraged by applications which run under Solaris, and can even be leveraged by applications which can not run under Solaris!

Sun pioneered in this arena, Oracle saw the investment possibility (by trying to purchase Sun), and now the rest of the market will need to play catch-up.

That's OK - Sun pioneered in heavily multi-core'ed and muti-threaded CPU architecture... and the rest of the market continues to emulate with Intel being the fastest to catch up.

Sun's innovation in benefited and will benefit the entire global computing community for years to come while IBM's innovation benefited only themselves and a few customers for a short period of time.

There is good reason to argue that this behavior is a primary reason why Sun was not consistently profitable and IBM typically is.

Sun = Innovation ; IBM = Business

0
0

This was a fun read.

Jerry Springer would love this thread! :-)

Kebby - a 500W processor - you should know no such processor exists - it sure as heck couldn't be air cooled! Geesh...must I post a Paris picture again! You and Mr. Halko must have gotten hammered a few times courtesy of a Sun credit card...

Neither of these systems is going to land on a floor - though I know I'll see a lot more 595s than 12-node T2+ systems around - IBM sells a lot of 595s (according to them and IDC etc.) because its a general purpose machine. The fact that a little over a year ago IBM needed as many storage units and spindles to produce their result is an impressive statement with respect to the 595s I/O capabilities. Heck, even HDS used a 595 to produce a SPC disk benchmark! Side note - HP had to show they used a 595 too when they used the HDS benchmark result to show off their rebadge storage box - what a hoot! :-)

Halko - calling Sun an innovator with respect to Niagara - do some fact checking - Sun bought Afara to get Niagara technology - its no Sun innovation!

Were flash storage in the IBM kit bag when the 595 result was published, I suspect it would have been used for the obvious price/performance benefits Sun/Oracle showed.

This benchmark, like many others shows me that core for core, which is how I pay for all my database software (perpetual licenses only please!) T2+ can't keep up with POWER6. Small caches, low memory bandwidth per core - things might get better with DDR3 and extra memory controllers etc etc...down the road.

Just look at POWER7 for a timely 8-core processor. Everything falls nicely into place - faster memory, faster busses, bigger caches, on chip L3 via EDRAM - no question Afara was a little ahead of its time and Sun made a good purchase, but Sun an innovator - not in server hardware...

0
0
Terminator

Larry and company are going to get fined again and probation

Ok Larry...you got away with a $10K fine for your $100K ad in the WSJ.

Your continued usage of the advertisement at Oracle openworld and mockery of the council will force them to prevent future publishing. That TPC-H you tried to pretend was ExadataV2 was actually HP and you will have to wait for a new TPC-H. You cannot be a member if you break the rules. You also will lose your Sun seat as soon as the EU approves this MySQL killing acquisition.

Larry at Oracle OpenWorld: "I think we're better. In fact, I'm certain we're better and John

said, probably. So, I said, great I'm running the ad. And up went the ad and IBM actually

went to the Transaction Processing Council and said because Oracle ran this ad, they

should not be allowed to publish any benchmarks for six months -- was the IBM solution.

Everyone else said, well, let's charge them 10 grand and maybe they won't do this again and

we ended up paying 10 grand and okay. And we promised not to run anymore ads saying

we're going to be faster and prove it later on"

http://ondemandpreview.vportal.net/main.cfm?m=browse.au&auid=64

Larry...you cannot use the Ad! Hello? Just because you are worth billions does not mean you are above everyone else.

There is a revolt starting. Oracle's $47K per core for EE and $22.5K per core for RAC which makes Exadata $7 MILLION per rack is causing customers to move to DB2.

You will not have me over a barrel, you can take your Exadata and your outrageous maintenance fees and get out of my datacenter. I now have two coffee cups on my desk.

IBM Power and IBM DB2, don't waste my time talking about adding more software to my ELA.

http://www.businessweek.com/magazine/content/09_38/b4147052120632.htm

Go ahead Larry...Make my day!

0
0
Big Brother

RE:David Halko

"That's OK - Sun pioneered in heavily multi-core'ed and muti-threaded CPU architecture... and the rest of the market continues to emulate with Intel being the fastest to catch up."

Well it wasn't really SUN, SUN bought the company that made the TX000 servers, it was AFAIR one of the 'original' SUN guys who made that company. Just like the highend servers which they got when they bought Cray's Unix business back in .. 1996 I think it was, wasn't their own creation. And yes the TX00 were innovative and the Niagara family is *GREAT* for what it was designed for. But there isn't not a general purpose box.

As for pioneered, multicores and multi threaded cores, Weeeel.. then the Power RS64 IV had up to 4 threads back in 2000, and you forget that the cell microprocessor came out at the same time as Niagara. And power4 was the first CMT processor back in 2001, where as the UltraSPARC IV (Jaguar) is from 2004.

"That's OK - Sun pioneered in heavily multi-core'ed and muti-threaded CPU architecture... and the rest of the market continues to emulate with Intel being the fastest to catch up."

I don't think that either HP nor IBM could have convinced their customers to give up their single threaded performance. And as for catch up, U forget that SUN is being absolutely trashed in the UNIX markedplace, sad but true. So there are no business reasons to emulate SUN's CMT. If IBM were really worried about the T5XXX line they could have put out a 'Q' version of their POWER6 servers as they did with POWER5+. Actually the 1.8GHz p560 has a higher socket throughput with 0,56 rperf than the power6 power 570. And such an MCM based Q server would have rocked, in benchmarks. But in the real world the server wouldn't have had enough memory to be useful, as when you run in a virtualized environment you use a hell of a lot of memory.

And early here next year we will have POWER7. 8 cores 32 threads on one chip, but still with the same or better single threaded throughput as POWER6.

So right now a T5XX0 might be 1.5 times faster per socket than a power 5x0 box. But with four times the cores and a bit more umph on each core, then it will be something like a factor of 2-3 the other way around. Sure rainbow falls with 16 cores will make a difference but IMHO it's just more of the same 128 statically scheduled threads versus 32 SMT threads where one thread actually can use the whole processor.

And as for the whole storage setup on the T5440 RAC benchmark, I tip my hat at that one.

But for the fact that you can quickly burn your way through a flash card if you write alot. Now I know that some SSD vendors simply just reserve half of the drive so as to extend the lifespan of the drive. And that would work for me.

Now the whole 'lots of disks on the benchmark' is a hack, In my world you have Big Tin Disk boxes from Hitachi,HP,STOR^H^H^HUN,EMC,IBM and Netapp that can do a lot of IO/sek by using cache. But as the TPCC benchmark is also about price, they are to expensive to use on such a benchmark. Hence you use a lot of disks. I find the disk price of the benchmark irrelevant, as in my world the disk systems are a constant, no mather what server or vendor we use.

"Sun's innovation in benefited and will benefit the entire global computing community for years to come while IBM's innovation benefited only themselves and a few customers for a short period of time.

There is good reason to argue that this behavior is a primary reason why Sun was not consistently profitable and IBM typically is.

Sun = Innovation ; IBM = Business"

I think you are very wrong. I grew up on BSD and Solaris Unix wise. I learned like a good UNIX boy at university to hate IBM, cause they were the big bad mainframe monopoly. And they were in the late 80'ies. But in the early 90'ies in my first IT job I had to use AIX and power, I lobbied for a HP or SUN box, but actually I got to like AIX with it's logical volume manager, SMIT menu interface, and easy to use cli where everything was like chxx mkxx lsxx rmxx. And you were the underdog as a AIX admin, that had to listen to how good Solaris, HPUX and Whatever name digitals Unix used at the moment. But us in the AIX admin department could "outadmin" the other admins, cause while they were editing files, and sending signals to deamons we just did a chxx and then a refresh -s subsystem. Or when they were formatting drives, and tying in sectors and sh*t, you simply did an extendvg and then that was done.

This is where my preference for AIX on power was founded.

But you seem to forget that the whole basis of SUN+Oracle's business depends on IBM inventions. IBM and others (not SUN) invented the RISC processor. IBM invented the relational database and SQL. And if they were such a monopoly, they why are SUN and Oracle in business at all ? Not to mention EMC, cause IBM invented the hard drive.

And if it hadn't been for the IBM PC, then neither Microsoft nor INTEL would be the giants they are today. What ground breaking technology have SUN invented besides Java ?

If you search for "SUN inventions" on google, the whole first page is about a novel.

How many nobel price winners have SUN fostered ? IBM is 5 or something, and not in economics, but in physics with discoveries like the tunneling microscope, and high-temperature superconductivity.

And you say SUN is innovating for the global computing community. Ok, that is why SUN was the very first to finance SCO's attack on Linux. Simply great, buy the rights to make Solaris code public, while trying to put Linux in the grave. No it wasn't Microsoft who were the first to go after linux using SCO as a proxy. It was SUN.

http://news.cnet.com/2100-1016_3-1024633.html

And we are many pasionate linux users who hasn't forgotten that move.

So yes IBM is all about business, but damn so is SUN. And when it comes to innovation, SUN is, or rather was, no match for IBM.

// Jesper

0
0
Silver badge
Happy

RE: RE:David Halko

Isn't the Sunshiners' insistance that Sun were somehow the greatest powerhouse of innovation ever just so funny? And the especially sad bit is they really believe it, despite all historic evidence. The really deluded ones end up like Kebabfart, just accusing anyone that disagrees with them of lying.

In the meantime, I'll speculate as to what Larry is up to. Well, he's not stupid, he's realised one of Sun's big problems was that customer confidence fell off the cliff. He also knows there's no such thing as bad publicity. The TPC-C benchmark session probably cost him a fair whack to run, much more than the timid tap on the wrist he got from the TPC. But it's done what he wanted - got people talking about Oracle and SPARC. Larry's aim is to get Snoreacle hardware salesgrunts back into those executive boardrooms that closed their doors to the dying Sun, and headlines on TPC records - even if they're just computer drag-racing - will probably help. Looking at Oracle's previous tactic of heavily discounting the database to push other Oracle products, I'm expecting offers of heavy discounts on Snoreacle software if it comes as a bundle on their hardware. What Larry loses in the short-term he'll plan to make back in a long-term relationship, especially from support charges.

The good news for us customers is that this could spring a nice pricewar in software and hardware, with genearous discounts all round. Thanks, Larry!

0
0

Apologize to Jesper

I see that you are not like Mattie Pattie. Therefore I want apologize for my answer. We may disagree but clearly you are no aggressive FUDer and liar. I accept that you disagree with me, that is totally ok. I dont like when people spread FUD though. But I dont think you are that kind of person.

As for your reply to me, I dont agree with you about the details. But never mind them.

Regarding my "nasty habit" of calling people FUDer and liars, as I see it, I have no nasty habit of that. As I see it, there are lot of people having nasty habits of lying about Niagara, about how slow it is, how it suffers from a small cache, etc. But how can that be true when facts show that Niagara crushes other CPUs on some work loads? I try to tell those morons that they are wrong and explain why they are wrong, but they just dont get it. I post benchmarks, I post articles, etc - but still they claim the Niagara is slow. As I see it, they LIE. I have proven them wrong, but still they persist.

Now, what do you call their behaviour? They are proven wrong, but still they persist. Are they lying then? It is like "this car with the new revolutionary jet engine can never be fast. Sure, it reaches 800mph which benches show, but I promise you, it is slow." I call that lying, dont you?

So I mean that some people have a nasty habit of lying because I prove them wrong, and still they continue - which is a nasty habit from them. Then I start to call them FUDer and liar. Is this wrong? Are they not FUDers and liars, you mean?

Your heart may be with IBM, mine will be at open source SUN. As long as people dont start to spread FUD, I have no problem with people disagreeing with me. Therefore, I wish you lots of beer and 6 in a near future! :)

0
1
Silver badge
Happy

RE: Apologize to Jesper

".....Regarding my "nasty habit" of calling people FUDer and liars, as I see it, I have no nasty habit of that....." LOL!!! I've lost count of the number of times you've called other posters liars or morons. Of course, that's not a problem when you suffer from an obsession verging on the religeous. Kebabfart, it's old advice but ever so pertinent - you really need to go get laid!

"....As I see it, there are lot of people having nasty habits of lying about Niagara...." Kebabfart, if someone is relating their experiences, and they don't match with your religeous beliefs, that does not make them a liar, it just means they have real practical experience. Some day, if you survive in computing long enough, you too will be able to speak from a position of experience, rather than just recanting whatever has been spoonfed to you by your Sun rep.

"....They are proven wrong...." All you have proven is your value as an unintentioned comic. Technically, you have proven nothing.

"...It is like "this car with the new revolutionary jet engine can never be fast. Sure, it reaches 800mph which benches show, but I promise you, it is slow." I call that lying, dont you?..." If you had an 800mph car you might be asked how it drove around town as a real indicator of its worth, rather than just relying on a top-speed figure from a test-track. You see, I don't doubt that Sun have dozens of carefully crafted "benchmarks", such as the TPC-C record, which bear no reality to what really happens in the enterprise. My experience is of benchmarking Niagara against Itanium, Power and Xeon with a real enterprise application stack we use, in a real enterprise environment, with real storage layouts. In those tests, Niagara sucked like the proverbial.

/SP&L

0
0
Happy

@VirtualGreg,@Jesper Frimann: Afara & Innovation

DavidHalko posts, "That's OK - Sun pioneered in heavily multi-core'ed and muti-threaded CPU architecture... and the rest of the market continues to emulate with Intel being the fastest to catch up."

VirtualGreg posts, "calling Sun an innovator with respect to Niagara - do some fact checking - Sun bought Afara to get Niagara technology - its no Sun innovation!"

Jesper Frimann posts, "Well it wasn't really SUN, SUN bought the company that made the TX000 servers, it was AFAIR one of the 'original' SUN guys who made that company."

VirtualGreg and Jesper Frimann, you may not be aware, but I wrote the original article with most of the content and references on Wikipedia for Afara!!! ;-)

http://en.wikipedia.org/wiki/Afara_Websystems

Afara did not have a piece of SPARC silicon when Sun invested in them. Sun brought SPARC CoolThreads into fruition. Almost a half-decade later and Sun is still the most significant octal-core CPU vendor on the market.

DavidHalko posts, "Sun = Innovation ; IBM = Business"

Jesper posts, "I think you are very wrong."

Your opinion is very fair - I can appreciate your point of view, I think we just disagree.

While Sun was doing the heavy-lifting by engineering heavily multi-core and multi-threaded processors into a single piece of silicon, others were cobbling together multi-core using multi-chip modules.

Some may argue that the MCM's were innovative engineering, disagree and suggest MCM's are a pragmatic business-man's short term solution to a technical problem.

0
0
Terminator

Oracle RAC sucks... But don't despair, Larrry: drill, baby, drill!!l!!

I love this thread! I'm learning a lot! Thanks to all sunshiners and ibmers for their passionate bellicism... Needless to say that I'm in one of these two camps [ehem.. is it clear enough??]

So, after a good laugh around the tricks with the pricing of this pooooor benchmark, I asked a friend to have a professional look at this Oracle RAC. This is his professional diagnosis:

- **Oracle Turned Off the Oracle RAC Cache Fusion** - Despite using Oracle RAC technology, Oracle actually turned off Cache Fusion (a key component of Oracle RAC) by setting gc_files_to_locks=“\1217-1276=42EACH”. Quite simply, Oracle isn’t passing data pages over the network between nodes, instead they are operating in Oracle Parallel Server (OPS) mode - a precursor to RAC – and pinging the pages off disk between the nodes.

wOW! wOULD YOU DO THAT to YOUR ORACLE RAC?? (this is me, not him... you just can't imagine the laughs when he shared this pearl ...)

- Excessive Logging Hurts Performance: DB2 Logs 1/8th of Oracle RAC per Transaction. TPC-C full disclosures include a measurement of the amount of log space consumed during an 8 hour run of the OLTP workload. This allows us to determine the number of KB of log space that is consumed per transaction. Why is this important? Logging is the most critical I/O factor in high volume OLTP applications. Assuming that systems are expertly tuned (and they are for TPC-C), then the logger can be the choke point for higher and higher performance levels. Quite simply, reduced logging means increased performance. DB2 has the most efficient logger in the business. The best DB2 TPC-C result logs about 2.4 KB per transaction. Oracle ran a TPC-C workload in a non-RAC Oracle 11g environment and logged about 4.9 KB per transaction. The recently announced Oracle RAC 11g TPC-C result logged a whopping 19.5 KB of log space per transaction.

- Oracle Partitioned the Data and Routed Transactions to Individual Nodes in the RAC Cluster. Despite the marketing message that you don’t need to change your applications to efficiently run on an Oracle RAC cluster, in their latest TPC-C result, Oracle directed transactions to specific nodes using a mastering technique. It wasn’t enough that Oracle used partitioning to create data zones to get around the inefficiencies of Oracle RAC, they changed the application! If you want true application transparency, DB2 pureScale delivers near-linear transactional scalability without application change.

- Oracle Turned Off Page Integrity Checking. No one ever said that TPC-C is “real-world” – but there are things that Oracle does in this benchmark that makes this down right dangerous in the quest for performance. For example, Oracle set db_block_checking=false and db_block_checksum=false. By turning these parameters off, Oracle does not check the integrity of a database page. Per the Oracle documentation, “block checking can often prevent memory and data corruption.” It’s interesting to note that you can’t turn page integrity checking off in DB2 – even in benchmarks. Very few [not insane] customers would run a production environment without page integrity checks; however, we can see why Oracle would want to in a benchmark: Oracle's documentation claims performance degradation for block checking between 1% and 10%. I especially like the fact that they set they set _gc_integrity_checks=0 which turns off integrity checking for global cache services.

Not bad... I hope you enjoyed your show, Larry, but this is much worse than what we expected.

We're very disappointed!!!

0
0
Welcome

@Jesper Frimann--- cutting teeth, SMIT, FACE, FMLI, XFMLI, Volume Mgmt, ZFS, and Linux

Jesper posts, "I lobbied for a HP or SUN box, but actually I got to like AIX with it's logical volume manager, SMIT menu interface, and easy to use cli... us in the AIX admin department could "outadmin" the other admins, cause while they were editing files, and sending signals to deamons we just did a chxx and then a refresh -s subsystem. Or when they were formatting drives, and tying in sectors and sh*t, you simply did an extendvg and then that was done."

I think the SMIT interface is great! It was just like the FACE interface bundled in AT&T SVR3!

I cut my professional teeth under AT&T UNIX SVR3 with FMLI based system configuration and it was awesome. I thought NCR UNIX SVR4 with XFMLI was a great addition. Sun's X based "admintool" was pretty neat for awhile, but it was never extended sufficiently (I think this is a failing of Sun - slowing the uptake of new user in the community.)

I really wish the Open market and Linux would have standardized on character based menu interface for administration (i.e. like FMLI), extended it to include X Windows (i.e. like XFMLI), and extended it again to "HTML" (no such FMLI extension that I am aware of) - to provide a standard way for all platforms across all protocols.

With FMLI and Admintool leaving the Solaris source base, it would be awesome if IBM open-sourced SMIT! woo hoo!

I agree that the un-unified configuration file bit is pretty ugly. The movement of Solaris 10 to Services (from inetd, inittab, /etc/rc, and varios cron jobs) makes a very nice unified cli. The XML basis makes it very open.

I am honestly glad I don't have to deal with Volume Managers any longer. I really like ZFS and the benefits it continues brings - it may not be perfect, but it is really helpful. Storage Management operations is becoming unified under Solaris ZFS. It will be nice to see ZFS appear on other operating systems, much like NFS had, due to Sun's friendly licensing model.

Jesper posts, "Simply great, buy the rights to make Solaris code public"

Sun has been continually going back to MANY Solaris contributors to buy or ask them to release their rights to licensed code to open source Solaris... or re-writing the code that could not be released, from scratch - it was not just SCO. The C-Net article you cited also indicated, "Sun needed the software for its version of Solaris that runs on Intel servers". When Linux code could not be used for Intel Solaris, there were very few commercial vendors left to go to, to get more code.

The Linux community looking at Sun in a negative light for "buying the rights" to open-source Solaris seems rather myopic. Sun has always been a significant Linux contributor, contributed A LOT of their own code & engineers into Linux, and ships a significant number of Linux platforms. From the C-Net article you cited, Sun paid to protect their Solaris & Linux's customers, "Sun's complete line of Solaris and Linux products...are covered by Sun's portfolio of Unix licensing agreements. Solaris and Sun Linux represent safe choices for those companies that develop and deploy services based on Unix systems."

The conclusion that Sun was the "very first to finance SCO's attack on Linux" from purchasing their own rights appears sensationalist to me. Great headlines for journals. Great propaganda for Sun competitors. The people who gain from that propaganda refused to go through the pain of open sourcing their OS's or move their proprietary OS's to Intel/AMD.

Had Sun been trying to assault Linux, code contributions to Linux from Sun would have ceased contributions (instead of continuing to increase Linux contributions), and Sun would not have paid to protect their Linux customers.

Have a good day, Jesper!

0
0
Joke

@matt bryant -- no good experience nececssary

--- matt bryant - if someone is relating their experiences, and they don't match with your religeous beliefs, that does not make them a liar, it just means they have real practical experience... I don't doubt that Sun have dozens of carefully crafted "benchmarks"

if mattie's experience don't match real world benchmarks, perhaps mattie is not a liar, but it begs to suggest mattie is incompetent. niagra is winning a lot of published real world benchmarks lately. http://www.sun.com/benchmarks/

poor mattie can't spell, references other threads where he admitted to having no experience, referenced old threads where he admitted to not being employed, talks to people who don't exist, can't find references to things he religiously believes is true, admitted to migrating off of sun long ago, and now suggests he benchmarked recent niagra chips.

can't believe anything from mattie who seldom provides a verifiable reference - where's the beef, mattie pattie? http://www.youtube.com/watch?v=Ug75diEyiA0

0
0
Flame

Matt Bryant = hypocrite

matt bryant ~ I've lost count of the number of times you've called other posters liars or morons... Kebabfart, if someone is...

matt bryant misspelled his name, hypocrite

0
0
rch

@Halko

Halko wrote:

"Iam honestly glad I don't have to deal with Volume Managers any longer. I really like ZFS and the benefits it continues brings"

Funny then that a volume manager and not ZFS is used in this tpcc benchmark. Sun is using the dreadful Solaris Volume Manager which even Sun employees admit is a POS.

0
0
Big Brother

RE:David Halko

"Afara did not have a piece of SPARC silicon when Sun invested in them. Sun brought SPARC CoolThreads into fruition. Almost a half-decade later and Sun is still the most significant octal-core CPU vendor on the market." (octo-core) I guess SPARC still operates in binary :)

We were talking about inovation. So you would agree with me that the actual inovation behind the TX000 and came from Afara. SUN did what you should do as a big company that buys a small inovative company, they turned the inovation into a product. And as I have said again and again, there is nothing wrong with the Niagara based servers if you use them for what they were intended for. But they were not designed for running a moderate number of heavy independent threads that do a lot of cross communication between the threads.

So, the reason why I am p*ssed, is that I am tired of having fanatics trying to get me to use these servers in my designs where it isn't appropriate. And the worst part is that I've lost some battles, and most of those solutions now have to be redone.

"While Sun was doing the heavy-lifting by engineering heavily multi-core and multi-threaded processors into a single piece of silicon, others were cobbling together multi-core using multi-chip modules. Some may argue that the MCM's were innovative engineering, disagree and suggest MCM's are a pragmatic business-man's short term solution to a technical problem."

I think you have missed the point all together. It's not multicore | MCM. It is multicore & MCM, if you have read some papers on MCM technology, you would see that there are quite a few benefits.

The MCM module that have been in the press the most, besides those from the mainframes, are the POWER4 and POWER5, which basically allowed you to plock 8 cores into a socket. Now nobody has claimed that those were in a single socket, or that IBM somehow beat SUN to the octo core by using these MCM modules. On the other hand the Quad core modules found in the power5+ based p5X0Q servers, did plug into a standard pSeries socket. And you might call that for a business man solution, but damn it worked, both Oracle and IBM counted it as being one socket. So a box like the p560 could run Oracle standard edition, just as the T5440 can today. And you might make fun of MCM's but for example the 505Q servers were just as fast as the T2000.

You make the creation of the Niagara chip sound like it a revolution. Hey, the original T1 was 8 SPARC II cores, that each rans 4 threads in a statically scheudled round robin multithreading mode. Clever and a good piece of ingeneering. But not a revolution.

Now as for the whole Linux/SUN/SCO/IBM deal...

"Had Sun been trying to assault Linux, code contributions to Linux from Sun would have ceased contributions (instead of continuing to increase Linux contributions), and Sun would not have paid to protect their Linux customers."

Ehh what code contributions to Linux ? I think you are kind of in the dark here. Try to find SUN's contributions to Linux 2.6 kernel here:

http://www.linuxfoundation.org/publications/linuxkerneldevelopment.php

They are not even in the top 31. Lets see how many sun developers that are active in maintaining parts of the Linux kernel...

[frimann@roedmette ~/linux-2.6.31.4]$ grep -i sun.com MAINTAINERS | sort -u | wc -l

1

One guy...

Or lets try to scan the whole source code for Sun contributions...

-> 11 hits on 6 different persons.

So lets just put the SUN was/is a major contributor to Linux in the grave.

And for fun I did the same with IBM, and there are thousands of contributions by hundreds of people.

And you state:

"The Linux community looking at Sun in a negative light for "buying the rights" to open-source Solaris seems rather myopic"

In law inforcement you normally operate with a concept called 'Cui bono', 'Cui bono' means who benefits in latin. Lets look at the timeline.

January 2003 IBM Software Boss starts talking about merging Linux and AIX.

In early 2003 SUN buys the Unix licens from SCO that gives SCO the financials to sue IBM.

In march 2003 SCO sues IBM and generally threatens everyone involved with Linux.

June 2003 the SUN marketing machine starts to crunch

http://www.sun.com/datacenter/migration/aix/AIX_WSJ.pdf

January 2005 SUN launches Opensolaris

Now what have SUN benefitted from this whole deal ?

// Jesper

0
0

JuanMaBlue

Let me ask you, have you asked your friend how IBM did their TPC-C benchmarks? Do you really think that IBM did their benchmark differently from Oracle? Think, McFly, think! Jesus.

Someone withdrew their TPC-C bench recently, who was that? Was it IBM?

0
1
BQ

prices and stuff...

IBM Power 595 64way 5ghz with 1TB of memory costs bit over 7M USD list price with 3y maintenance (24x7 is standard with these babies)... included in price are PowerVM enterprise & Management edition for AIX...

also regarding innovations please visit...

http://www-03.ibm.com/technology/index.html

and also... who has 100% market share on Mars and Moon? guess who... IBM :-)

regarding green and things... latest published list

http://www.green500.org/lists/2009/06/list.php

(pay attention on 4th system on list... look it's no.1 from Top500... )

and YES they are all based on IBM Power technology...

btw I work for IBM... moved there from HP where I really enjoyed working with Alphas (which where very good systems and they are withdrawn in favor to Itanic crap)... and do u know why I moved to IBM... it's because of TECHNOLOGY and INNOVATIONS... and I haven't been wrong...

0
0
Silver badge
Happy

RE: @matt bryant -- no good experience nececssary

Aw, has ickle Kebabfart got so embararssed by his own posts he's now posting anonymously?

".....if mattie's experience don't match real world benchmarks, perhaps mattie is not a liar, but it begs to suggest mattie is incompetent. niagra is winning a lot of published real world benchmarks lately. http://www.sun.com/benchmarks/...." <Yawn> Back to the cherry-picked benchmarks AGAIN? You really just don't know when to quit, do you? Let's look at the one real benchmark for success - profits. Hp and IBM make them, Sun hasn't for years. Why? Because hp and IBM sold their boxes because customers decided they were better than the Sun ones, which didn't sell. Now, just for fun, please show me the last time Sun made four consecutive quarters with a profit.

".....references other threads where he admitted to having no experience....." Please post a link to the thread where I stated any such thing. Or we can look at what you posted in the thread http://www.theregister.co.uk/2009/07/22/sun_sparc_t_crank/comments/, where you admitted your own company uses IBM in preference to Sun kit, so you don't even have experience of the Niagara kit!

"....referenced old threads where he admitted to not being employed...." The only thing I admitted to was not being employed by hp or IBM. I'm quite confident in saying I not only have more experience than you, but also a better-paid job.

"....talks to people who don't exist...." Are you upset at me calling you Kebabfart? Thin-skinned or what! You really need to grow up, there's a lot of nastier things that people must think about you.

"....admitted to migrating off of sun long ago, and now suggests he benchmarked recent niagra chips...." Yes, we did migrate off E10ks and E25Ks. When we came to replace them we had to prove to management that what we put in place was better than what Sun was offering. We did so by shoot-outs, and Sun didn't win any of them. We have used shoot-outs since as a quick and easy way (for us, not the vendors!) to effectively do proof-of-concept on their offerings so we can see which one really lives up to the hype. That means I have put Mseries and Niagara servers up against other vendors' iron. You can guess the way those tests worked out by the fact that we haven't bought any Sun for years, but we have bought IBM and hp (and please note we bought IBM despite my pref for hp). Now, I'm not saying our tests are reflections of what everyone runs in all cases, but we do try and use as many generic apps as possible as it is simpler to support than mountains of inhouse code (believe me!), so they could be taken as fairly representative of general enterprise use. And, seeing as I have at least seen the boxes running real apps with real data, whereas all you do is repeat selected benchmarks and Sun marketing, I think I can comfortably say you have no comparable reference.

"...where's the beef?...." That's exactly what the Sun shareholders were asking all year. After all, they were slightly upset they bought shares in Sun when it was worth $200bn, and yet, within a decade, they had to settle for Oracle grudgingly buying them for chump change. All McNeedy's and Ponytail's talk turned out to be just so much empty bun, with "Solaris and SPARC and nothing else" being exposed as the recipe for financial disaster. And then Oracle didn't even want the wole of Sun, they just wanted the tiny burger of the software bizz, but instead they had to take the stale bun as well. And now Larry is trying to sell that same stale bun and mini-burger with a bit of Oracle relish and some flashy marketing. Larry may be hoping there's a sucker born every minute, but the problem seems to be that, like Kebabfart, most of them are not in a position to be making purchases.

In the meantime, shall we start a litlle pool on when hp or IBM post their own flash-boosted TPC-C record?

/Enjoy!

0
0

Mattie Pattie,

I dont post anonymously, other than when I post "i work at a large bank, and I love the Power6, but now we migrate away to Niagara".

.

Does it surprise you that there are more people than me, claiming that you lie? Regarding your benchmark: profit. So what? Ive want to show you that your are FUD by claiming that Niagara is slow, and by writing Slowlaris. If you were true, I wouldnt be able to post benches showing that Solaris on Niagara easily kills AIX and Power6. But now I am able to post such benches, which means you are lying. The Solaris and niagara is in fact many times faster than IBM, on certain work loads.

So let me ask you again: which is the fastest? Will you stop spread FUD and lies? Clearly, benches show which is many times faster: Niagara.

0
1
Silver badge
FAIL

RE: Mattie Pattie,

"......Does it surprise you that there are more people than me, claiming that you lie?...." Not really, it's quite a common experience when dealing with those blinkered by religeous belief, they tend to treat any attempt to explain reality to them as a personal attack and hence get shirty and insulting.

"....Regarding your benchmark: profit. So what?...." Try saying that to the 3000 Sun employees that just go the boot Then think about the gaping holes in capability that must mean have just been added to the already well-holed Sun support, development and sales structures, and then try saying "so what" to Sun's current and possible future customers. Then again, thinking doesn't seem to be too much of a strong point for you, you're probably waiting for the Sun rep to call and tell you what to think about the layoffs.

"....Ive want to show you that your are FUD by claiming that Niagara is slow, and by writing Slowlaris. If you were true, I wouldnt be able to post benches showing that Solaris on Niagara easily kills AIX and Power6. But now I am able to post such benches, which means you are lying...." Yeah, and I came up with that Slowaris moniker all by myself - not! And Niagara has very low penetration anywhere outside the webserving niche, with Sun even having to bring in the M3000 to satisfy customers that complained the T-series were pants compared to their old Netras. Care to dispute those facts?

"....The Solaris and niagara is in fact many times faster than IBM, on certain work loads...." LOL! Look at the last phrase - "on certain work loads" - do I detect a little bit of wriggling going on? Just the sentences before you assured us "If you were true, I wouldnt be able to post benches showing that Solaris on Niagara easily kills AIX and Power6", so which is it, does Niagara "easily kill" AIX and Power or just on a few workloads? Come on, at least make this hard for me, there's no sport in it when you keep shooting yourself in the foot with such amusing regularity! Just for fun, I want to tell us exactly how many comparisons on whatever "work loads" you like you personally have actually run between Power and Niagara (hint - if you claim any then we already know you're lying).

/SP&L

0
0
Happy

Afara, Multi-Chip Modules, Linux

Jesper posts, "We were talking about inovation. So you would agree with me that the actual inovation behind the TX000 and came from Afara. SUN did what you should do as a big company that buys a small inovative company, they turned the inovation into a product."

Afara had an idea, they did not have silicon, nor did they have the expertise to gain the funding - because people with the money were not certain if they could make it past technical hurdles. They did bring some old Sun SPARC expertise in, before they were purchased, to work past those hurdles.

Anyone can have a good idea, but when there are technical hurdles, it takes innovation to overcome them, to bring an idea to fruition. It also takes an innovative spirit in order to select a good idea to bring it to fruition.

Jesper posts, "I am tired of having fanatics trying to get me to use these servers in my designs where it isn't appropriate."

Join the club. I feel the same way about various other servers.

Jesper posts, "if you have read some papers on MCM technology, you would see that there are quite a few benefits."

Doubtless, there are many benefits to multi-chip modules, as well as drawbacks.

Jesper posts, "And you might make fun of MCM's but for example the 505Q servers were just as fast as the T2000."

I don't think I made fun of multi-chip modules, at least I didn't try to. I merely said it was a business-man's approach. MCM's require less engineering, get faster to production, uses older technology effectively, reduces risk, but does not perform as fast if engineered it onto a single piece of silicon.

David posts, "Had Sun been trying to assault Linux, code contributions to Linux from Sun would have ceased contributions"

Jesper posts, "what code contributions to Linux?"

Sun is in the top 30 corporate contributors, according to the August 2009 update from the Linux Kernel Development. They are listed on the printed hard-copy, pages 11 & 19.

Have a great day Jesper! Always great chatting with you!

0
0
Big Brother

RE:Afara, Multi-Chip Modules, Linux

"Jesper posts, "I am tired of having fanatics trying to get me to use these servers in my designs where it isn't appropriate."

Join the club. I feel the same way about various other servers."

My whole point was that I don't think that Niagara based servers are general purpose servers.

So you are saying that you don't feel that one of more of Itanium, POWER, Xeon or AMD based servers are not general purpose servers ?

"Doubtless, there are many benefits to multi-chip modules, as well as drawbacks"

And the drawbacks are ?

And originally you said:

"Some may argue that the MCM's were innovative engineering, disagree and suggest MCM's are a pragmatic business-man's short term solution to a technical problem."

Short term solution ?

MCM modules have been used in mainframes since the 70ties.

Intel used MCM's on the PentiumPro, and later on Presler,Dempsey,Clovertown, Kentsfield and Yorkfield. And who knows they might do it on Itanium also.

IBM used POWER2,POWER4[+],POWER5 [+],POWER6[+] and POWER7.

And from what have been written here on the register then we are gonna see a 4 chip MCM power7 module. Now that is going to be a 32way MCM module which you can have in your hand

like this: http://upload.wikimedia.org/wikipedia/commons/1/1f/Power5.jpg

And that MCM module will most likely have the several times the Umpf, than a current M8000 does. Hence such a module will have all the busses, all the infrastructure and part of the backplane that goes into the CPU part of seven or eight CMU boards of a M9000.

Oh no that is not innovative. I suspect that the only reason you say it's not innovative is cause SUN doesn't do MCM's.

"Anyone can have a good idea, but when there are technical hurdles, it takes innovation to overcome them, to bring an idea to fruition. It also takes an innovative spirit in order to select a good idea to bring it to fruition."

IMHO there is nothing wrong in that, large corporations like HP, IBM and SUN do it all the time.

But it still doesn't change that the original innovation was done by another company.

My problem is that when people say things like this, it treated as a sacrilegious act, by the followers of the SUN. Sometimes I feel like Solaris and SPARC are religous icons, when speaking to the followers of the SUN.

And I can say SlOwLARIS just as well as I can say HPsUX and AIntuniX, TooNIX which is what the OS'es are known at where I work. It kind of softens things up between the different fractions.

"Sun is in the top 30 corporate contributors, according to the August 2009 update from the Linux Kernel Development"

Ok, found it, so this year they made it into the top 30. Wheeee!

Hardly impressive when compared to their main competitors INTEL/HP and IBM.

But still it doesn't change the fact that the number of people at SUN that are maintaining a part of the Linux kernel is one. So it's not like SUN has invested a great deal of effort in Linux.

So your original statement:

"The Linux community looking at Sun in a negative light for "buying the rights" to open-source Solaris seems rather myopic. Sun has always been a significant Linux contributor, contributed A LOT of their own code & engineers into Linux, and ships a significant number of Linux platforms."

Still doesn't stand. It is simply not true, period.

// Jesper

PS: And I think it's very sad that SUN has to let more people go.

PPS: Bigbrother cause something is rotten in the state of Denmark.

0
0

Mattie Pattie Laddie

I dont post anonymously.

.

If we talk about performance and best tech, does a big sales record prove that the performance is best? No. Windows outsells everything else, does it mean that Windows is the best OS? No. Regarding performance and the best tech, I would appreciate if you posted hard links to hard benches, instead of posting "your opinions". You can not conclude that a certain CPU is slowest based on an opinion - especially when the benches show it to be fastest in the world.

0
1

Jesper Frimann

"...My whole point was that I don't think that Niagara based servers are general purpose servers..."

But I dont think anyone here believes or claims that? SUN has made it clear that on certain work loads the Niagara is really good. It turns out that Niagara is many times faster a several times higher clocked CPU, on some work loads. And there are workloads where Niagara sucks. Hence, it is not a general purpose server. But OTOH, which server CPU is that? For instance, if you have a large enough work load so the Power6 cache will thrash, swap things in and out all the time, then Power6 will not perform good. It depends on fitting all work load into it's cache to achieve good performance. Otherwise it will suck. Hence, Power6 is "not a general purpose server CPU" - just as Niagara.

What I am trying to say is that every solution has it's weaknesses and strengths. There is no CPU which is good on every task. SUN has made clear of Niagara weaknesses. IBM has not made clear of Power6 weaknesses. "My whole point is that Power6 is not a general purpose CPU, as it has weak sides just like every CPU has. Including Niagara".

0
1

Page:

This topic is closed for new posts.

Forums