65 posts • joined 25 Aug 2009
Re: With same JDK
FYI. 2.8 GHz Intel Xeon E7-4890 v2 is brand new (what this article is talking about). And compared to previous generation Xeons, SPARC T5 shows a significant performance advantage whether per chip or per core based comparisons if you look at any real world benchmark, whether DB based or Java based. And considering that software licensing favors faster per core performance, the ~10-20% HW price difference between a SPARC T5-2 and a similarly configured 2.8 GHz Intel Xeon E7-4890 v2 2-socket system gets erased pretty fast considering the SPARC T5-2 has a 21% performance/core advantage (based on SPECjbb2013) than the very latest 2.8 GHz Intel Xeon E7-4890 v2 systems. If we ever see any DB benchmarks published on these Xeon E7-4890 v2, will be interesting to see how it compares to the leadership results already published with SPARC T5.
But with lower overall performance than these new Xeon chips? What are you smoking? Until theres actually any benchmarks published, SPARC T5 currently smokes recently announced Ivybridge-EP v2 and compared to previous fastest Westmere-EX, SPARC T5 shows up to 5x faster performance depending on the benchmark. A SPARC T5-4 4-socket recently demolished the 8-socket HP ProLiant DL980 G7 server on the 10TB TPC-H datawarehouse benchmark by 2.4x with 28% better price/performance running a real enterprise level database! That’s 4.8x better performance per chip and 3x better performance per core! And SPARC T5 runs all 8-threads per core with a full 16-cores utilizing critical thread technology-and oh, by the way, its been out for almost a year now. https://blogs.oracle.com/BestPerf/entry/20131125_t5_4_tpch_10000gb
Intel's marketing says butt is being truly kicked? Sounds more like that old Wendy’s commercial “Wheres the beef?” As in wheres the benchmarks to prove these marketing claims (Ivybridge-EX vs SPARC T5)? And wheres the systems to justify the price comparisons-Estimates again? If you read the fine print, and yes, Intels made sure its less than 6 point type, the benchmark claim and comparison to SPARC T5 is based on “Intel Estimate” using SPECint_rate_base2006. Nice to compare performance based on estimated RAW performance, but wheres the system level application benchmarks? Wheres SPECjEnterprise2010 or TPC-C or even TPC-H? From what I can see, SPARC T5 is kicking Ivybridge v2 ass. On SPECjEnterprise2010, SPARC T5-2 is 1.5x faster than just announced 2-socket X4-2 Ivy Bridge E5-2697 v2 system. On SPECjbb2013, SPARC T5-2 is 1.8x faster than 2.7 GHz Intel Xeon E5-2697 v2. And the SPARC T5-4 just blew the doors off the previous 8-socket Xeon E7-4870 based HP DL980 G7 on TPC-H at 10TB by 2.4x. Until Intel or its partners publish real world benchmarks, making false claims based on “Internal Estimates” just shows how desperate Intel really is these days. Heres the beef. https://blogs.oracle.com/BestPerf/
Re: Catching up on SPARC T5
Old Power7 has more than twice the performance per core of the new T5? Time to redo your math! On which benchmark? Which comparison? And why are there no Power7+ benchmarks to compare to on database benchmarks like TPC-C or TPC-H? And don't get caught in the trap of comparing a fully configured SPARC T5-8 versus a very low configured Power7 system and assume that per core/performance will scale linearly to higher end configs! If it did, Im sure IBM would have published the result! Unless of course you're buying an 8-core system.
On TPC-C OLTP benchmark, SPARC T5-8 achieved 8,552,523 tpmC w/128-cores. The closest throughput Power7 system achieved 10,366,254 tpmC w/192-cores (using 3x more CPUs!). SPARC T5 has a 24% per core advantage! SPARC T5 is also 2.4x faster per chip and 2.5x better $/tpmC than fastest (and closest comparable)Power7 result.
On TPC-H @ 3TB, SPARC T5-4 achieved 409,721 QphH with 64 cores while the fastest Power7 with 32-cores achieved 164,747 QphH. SPARC T5 is 24% faster/core and also delivers 2.1x more performance using half the # of CPUs than the Power780. Both TPC-C and TPC-H have price/performance metric which shows SPARC T5 being far superior to *any* Power6/Power7 results. And this price/perf metric includes SW & HW costs! And with no Power7+ results available, we can only assume its slower...
And as for AIX vs Solaris, Solaris is years ahead of AIX and now that IBM is so focused on delivering Linux on Power, AIX is certainly falling behind faster. Why after all would IBM release Power systems that don't support AIX?
And finally, do you really believe that Power8 will have superior performance/core than Power7 or Power7+? with 50% more cores @ same 4GHz, its unlikely that perf/core will improve, especially on larger systems. I guess we shall see sometime late next year!
Re: Catching up on SPARC T5
Well asdf, you clearly haven't done your homework. Locked-in to Larry's pricing? Have you compared systems pricing between SPARC and Power lately? Clearly not!
A SPARC T5-8 is 3-4x lower priced than an equivalent performing Power 770 or Power780 system. Even a high end configured SPARC M5-32 system is lower priced than the slower high-end Power 780 configuration! Maybe that’s why IBM's Power Revenue has been hemorrhaging these last 5 quarters?
And even software licensing is half the cost on SPARC vs Power whether running Oracle or IBM SW or even SAP for that matter! But as you say, Larry is a business man and a successful one at that being one of top ten richest people in the world. Would he invest in a dead end business? Doubtful.. Would he invest in a business where he can be successful, become #1 and make a lot of money? Absolutely! And whats making him (and Oracle) money (hint: Oracle has remained profitable and continues to show revenue gains since the Sun acquisition)? Selling engineered systems based on best of breed technologies like SPARC and Solaris and beating IBM and HP. Have you seen IBM's last 5 quarters of double-digit revenue declines on its Power Business? Even IBM's business as a whole has been down for last 5 quarters! How about HP's last 7-quarters of double-digit Itanium declines? Sure, when Oracle first took over Sun 3+ years ago, SPARC was in a decline. But not today. Oracle has invested $3+BN in R&D these last 3 years on SPARC with a solid roadmap (which is public by the way) to the future and it shows-Oracle is delivering. Fujitsu is also heavily investing in SPARC with its own SPARCX technologies. Why? Because there is huge opportunities for growth here and with Intel abandoning Itanium, and even showing signs of defocusing on high end Xeon due to ARM beating the crap out of them, who's left to address the commercial enterprise market? Oracle.
Re: The death knell for SPARC.
Hey Matt-Still working for HP? Hows that working out for you?Hows HP's server business doing today? With over a $1BN being invested in SPARC today by Oracle, and Fujitsu also investing heavily in SPARC, how could you justify such remarks without being seen as a troll?
My view on things is that Oracle is leaving x86 to Dell, just like TPM points out, so Oracle can focus on Engineered Systems and SPARC. Oracle clearly has Intels roadmaps on x86 futures, and its clear if you look at the entire x86 market, that almost every company that bet on it is suffering today including Dell, IBM and HP. Even IBM is trying to abandon its x86 business to Lenovo and Intel, with its new CEO, is clearly going after ARM, which will continue to defocus them from the enterprise space which has significantly lower volumes which makes it hard for Intel to get any ROI. They’ve already abandoned the science project they called Itanium and its been over two years since they’ve updated its "EX" class of processors, not expecting an update till late this year, maybe early 2014 - 3 years from Westmere-EX. If you look at this weeks Haswell announcement, its all about mobile computing. What about the Enterprise Space?
Re: Here are some benchmarks
SPARC M5 is not a match to the Power 795? You’ve got to be kidding me! SPARC M5 is PCIe Gen3, Power 795 is still PCIe Gen1! The I/O on Power795 is the same as the previous Power 595 introduced back in 2008! SPARC M5 is 8-socket glue less and supports up to 4 hardware domains for full electrical isolation, not possible in 795. The list price of a 32-socket SPARC M5 with 4TB of RAM has list price of just over $1M, while a *base* config P795 with just 32 cores active and *1TB* of RAM is roughly same list price. Sure discounting will be high here, but IBM will need to discount a minimum of 70% to match Oracles SPARC M5 *list prices* And then theres the performance comparisons.
Re: M5-32 is not the only one with 32TB RAM
And SGI's systems run enterprise software like Oracle DB?
Re: Phil 4 SPARC lol
Although most of the press and analysts talk in terms of Revenue marketshare, this can be quite misleading, especially if you sell very expensive systems but NOT necessarily sell many of them. Selling 10 x $10M systems will show higher marketshare than selling 10,000 $9.5K systems.
~50% of IBM's UNIX/RISC revenue (~$2.5BN/Year) comes from just 3 very expensive systems that has had no real competition in last 3+ years. Power 770/780/795. The Sun/Oracle M-Series was Sun/Oracle's revenue generator systems trying to compete against high end Power7, but it came out in 2006/2007, over 6 years ago. Maybe that’s why Oracle HW has been in big decline these last 3+ years because theres been no real replacement to M-Series in 3+ years? However if you look at the volumes of these systems, its less than 10% of the total volume. For Oracle SPARC/UNIX, SPARC T4 represents ~65% of the volume but less than 40% of the revenue. Not enough to show revenue growth. And Oracle Engineered systems are still ramping up with longer sales cycles, and in many cases doesn’t replace high-end SMP systems.
Whats important though for ISV's, and customers, is *UNIT* marketshare. Why? Because more volume translates to more license sales, lower costs for smaller more powerful systems with lower pricing on components and ultimately a more healthy eco-system. Its only when you have enough volume, do you become really relevant, as has been the case for Xeon. SPARC/Solaris still has the #1 UNIX volume marketshare and with SPARC T5 and SPARC M5 as well as SPARC M10, Oracle now has the opportunity to gain not only more volume marketshare, but revenue marketshare. And realize it took IBM over a year to release the entire Power7 line, and 4+ months to release Power7+ line, and no Power7+ on Power795. Lets see if Power8 is released *next* year!
Re: SPARC lol
Look at IBM's last 4 Qtrs of Financial statements and question how is it that IBM's Power systems business is down so much? Maybe SPARC is BACK?
IBM reports 2012 first-quarter results
“Revenue: $24.7 billion, flat”
“Revenues from Power Systems were flat compared with the 2011 period”
IBM reports 2012 second-quarter results
“Revenue: $25.8 billion, down 3 percent”
“Revenues from Power Systems were down 7 percent compared with the 2011 period”
IBM reports 2012 third-quarter results
IBM total revenues for 3rd Qtr of 2012 of $24.7 billion were down 5%
“Revenues from Power Systems were down 2 percent compared with the 2011 period”
IBM REPORTS 2012 FOURTH-QUARTER AND FULL-YEAR RESULTS
“Revenue of $29.3 billion, down 1 percent”
“Revenues from Power Systems decreased 19 percent compared with the 2011 period”
Well Mr. Nelson, what really matters is how much performance the SPARC T5-8 delivers *and how much does it cost to buy and run* vs equivalents from IBM or others. So while in certain benchmarks there may be advantages or disadvantages in per core performance, the end result is TCO or $/performance. And by the way, SAP licenses per user not by core so the per core calculations are irrelevant!
If you were to price out the SPARC T5-8 and Power 760+ SAP configurations (with HW, OS, Support), SPARC T5-8 beats Power7+ by 3x in Price/Performance!
Chips Cores $US List $/SAPs
IBM Power 760 8 48 $718,801 $5.16
SPARC T5-8 8 128 $365,489 $1.65
TPC-C - Top Ten Performance Results has full disclosures
SPARC T5-8 $0.55
IBM Power 780 (MHB) $0.69
IBM Power 780 (MHB) Clustered $1.38
Re: Fastest ^H^H^H^H^H^H^H processor..?
Yeah and even funnier that IBM hasn’t published a high end non-clustered Power7 config ever since the Power 595 result was released eons ago. Why? Maybe because Power7 and Power7+ has worse per core performance than Power6 and IBM doesn’t want everyone to realize this? It was nice in the old days to be dual core extremely high GHZ, great for single thread, poor on throughput but todays world has changed with more throughput challenges being addressed by multi-core and multi-thread CPUs. Problem is, Power7/Power7+ cant do both. Either high GHZ with few cores/CPU but high per core perf or low GHz but with 8- cores but with poor per core perf but high throughput. You have to pick one.
Re: who cares about specint for mainframes
And why hasn't IBM benchmarked a fully loaded non-clustered Power 770/780/795 or even Power7+ system? Because their per core performance would all be worse that the lowly 2-socket Turbocore Power780 config they ran a few years ago. IBM whats everyone to size from the smallest configs using rPerf which as we all know, oversizes on workloads that have heavy I/O
Re: And just how much is the Oracle RDBMS going to cost?
Actually, now that SPARC T5 has proven performance leadership from integer/floating point to DB, SAP, OLTP and Java plus *all* Oracle software (Siebel, Peoplesoft, JD Edwards, etc), the SPARC T5 is now the reigning price/performance leader *including* Oracle Software that’s licensed on per core basis. SPARC T5 has the same 0.5 multiplier as Xeon and is half that of the Power7/Power7+ 1.0 multiplier so you'll require atleast 2x more licenses running Power7+ for the same performance, and compared to Xeon, you'll need roughly 25-50% more licenses on Xeon based on latest results depending on workload. You can see the details here: https://blogs.oracle.com/BestPerf/
SPARC T5 not only runs at 3.6GHz, but SPARC T5 has 16 x cores/CPU and 8 x threads/core-more than any CPU in the world. So theres more to performance than just GHZ. Both SPARC T5 and M5 run circles around the 5.5 GHz zEC12 chip. SPARC T5 leap frogged IBM's latest Power7+ and latest Xeon SandBridge from integer and floating point performance to SAP, OLTP and Java.
Just look at the 17 world record benchmarks just published!https://blogs.oracle.com/BestPerf/
Re: who cares about specint for mainframes
Its clear from your comments that you don't follow SPARC, as this year, SPARC celebrated 25 years, so clearly SPARC has survived longer than any other processor out there. Even Solaris has been around for 20+ years. If there is any processor (and OS) that has "staying power", in the enterprise, its SPARC/Solaris. http://www.youtube.com/watch?v=IKB9zV8TXuQ
Re: IBM Power is extremely expensive so they must drive perception of "low cost"
Oracle is in the enterprise software business not HPC. If you look at the SPEC CPU definition here: http://www.spec.org/cpu2006/Docs/readme1st.html , you will see that "SPEC CPU2006 focuses on compute intensive performance, which means these benchmarks emphasize the performance of:
• the computer processor (CPU),
• the memory architecture, and
• the compilers.
"SPEC CPU2006 is not intended to stress other computer components such as networking, the operating system, graphics, or the I/O system". When you can tell me which enterprise software environment, like database, middleware, ERP, OLTP, datawarehouse, BI, etc does not require networking, I/O or an operating system, I'll tell you why I think Oracle didn’t publish SPARC T4 SPECint2006rate scores. Furthermore, SPECint_rate2006 benchmark is actually based on applications including chess playing, video compression, PERL programming, event simulation, physics: quantum computing and other bizarre tests. Great for CPUs like Power and even Xeon which run in gaming consoles and PC's for playing games and HPC, but not enterprise software. None of these tests should be of concern for Oracle customers.
So if you want to hide your systems poor I/O throughput, poor networking or OS scalability performance, run SPECint2006rate because its not tested. Otherwise run benchmarks that simulate customers real workloads.
Oracle has published over 20 benchmarks on SPARC T4, many of them world records, and run across all of Oracles SW, so if you're an Oracle customer, you're covered. Should be asking IBM why there are no OLTP, datawarehouse, or *any* Oracle based benchmarks on Power7+? Does Power7+ run Oracle SW? And dont tell me Oracle isnt allowing IBM to publish Oracle benchmarks. Don't you think IBM would talk to the press if this was the case?
Re: IBM Power is extremely expensive so they must drive perception of "low cost"
Your comparisons are a little flawed Jesper...
Regarding "POWER7+ does run circles around the T4, a POWER 740 seems like a good match. Again if you take the SPECjEnterprise2010 benchmark you'll see that a POWER7+ core is aprox x2 of a T4 core."
The Power780+ @ 4.42GHz SPECjEnterprise2010 score of 10,902 EJOPs does not compare to a SPARC T4 *clustered* score of 40,105 EJOPs, and certainly doesn’t beat it. Sure, if you calculate per core performance, the Power7+ has higher per core but thats with a *single* Power 780+ system that’s half configured versus 6 x SPARC T4-4 servers clustered with Oracle RAC which clearly has overhead and more complexity to manage. IBM is playing tricks versus previous Power7 result. How else could IBM get a ~2x per core increase over previous Power 780 configuration when rPerf shows roughly~19% increase/core? Also, a Power780+ doesn’t compare to a Power740+ as you suggest. Why didn’t IBM try to beat Oracles SPARC T4 result? Because they can't. Why didn’t they run a Power 740+ to show its per core superiority to SPARC T4? Because they cant. IBM wants you to believe that you can extrapolate from a specially configured high end, million$ setup that avoids all bottlenecks of the system. And I would bet if IBM used a fully configured system (16 x sockets, 128 cores) instead of a base config of just 32 cores, the performance/core would drop significantly due to the latency overhead of the CEC fabric. Plus, a Power 740+ as you suggest, is running a slower Power7+ @ 4.2GHz and is limited to half the memory of a SPARC T4-4 for example. Its not an equal comparison.
Show me a benchmark with same SPARC T4 4-socket and Power 4-socket system configurations and then we can talk about who's CPUs and cores are faster. Where are the TPC-H, TPC-C, SPECjEnterprise2010 benchmarks on Power7+ with mid range?
If I were to use your comparison however, a *large config* SPARC T4-4 32 core 1TB RAM lists for approx $117K, where as a P740+ 4.2GHz /1TB system with half the cores (16), lists for ~$150K and that’s without PowerVM and AIX Std Ed. A ~$30K difference. You probably didn’t configure with same RAM, memory riser cards, DVD drive, RAID card and all the other "options" needed to compare to whats built-in on SPARC T4. Also, a closer comparison would be a Power 750+ with 32 cores, 1TB RAM which lists for ~$270K when configured identically to a SPARC T4-4 and again, performance wise, should be close considering the SPARC T4-4 beat previous Power 780 8 CPU/32core @ 4.14GHz on TPC-H @ 1TB and 3TB for example. And note that SPARC T4 is a year old with SPARC T5 coming out soon with well over 2x more performance and scalability, not the minimal 20-30% or so that Power7+ gets over Power7 almost 3 years ago. So regardless of your comparison, IBM's Power systems are expensive and price/perf sucks compared to SPARC.
Re: Power750 finally has PCI-Express 2.0 slots?
Wow-they’ve caught up to 2007 PCIe Gen 2 technology while Intel and soon Oracle will have PCIe Gen3 across their product lines. Power 795 is still PCIe Gen 1!
IBM Power is extremely expensive so they must drive perception of "low cost"
The IBM Power 710 for just $5,947 is a joke and is IBM's bait and switch system. IBM will lose money if they just sell the box as is with nothing attached or added. The system is an extremely basic 4-core, 8GB server with no OS, no virtualization, no SW maintenance, no HW management, and basically everything is optional-DVD drive optional, memory card risers optional, RAID optional, etc etc etc. By the time you add all the options, that are really needed to run the system properly, you're looking at over $20,000! And if you compare a like for like, core to core configuration against Oracle SPARC T4 systems for example, the IBM Power7+ systems are 30% to several times more expensive. Have you priced out the new Power 760 system? We're talking $400K plus for a 48-core system! Even a Power750+ 4-socket 32-core box is well over $240K. Watch out for IBM's "Option Gotchas". The devil is always in the details
Re: investment protection
SPARC roadmap has been frustrating for years? Have you been living in a box these last 3 years? The SPARC roadmap hasn’t changed since 2010 when Oracle took over and has been consistent ever since-delivering to the roadmap like clockwork. I think IBM's bewildering choice of 22 different Power7/Power7+ CPU GHZ and core counts that aren't available in same system is frustrating! How do you choose? Which is best? So much complexity.
Continuity? Upgrade Path? The same M-Series has been shipping the *longest* since 2006 and has received 4 in-chassis CPU upgrades (that can be done without powering down system) over its lifecycle of more than 6 years. In the same timeframe, IBM's p5 595 required pretty much an entire overhaul with downtime to "upgrade" to Power6 and then to Power 795, again, another major "overhaul" upgrade to Power7. Do you really believe that Power8 is going to be an easy-in chassis upgrade? It certainly won't be without incurring costly downtime. And we wont talk about Itanium...Poulson is still MIA and Kittson is DOA. So Fujitsu's (and Oracles) SPARC M10 systems are the next generation M-Series (APL) systems. You did notice the Fujitsu/Oracle logos on the chassis? And whats nice, unlike Superdome2 or Power795, you aren't locked-in to this massive chassis. SPARC M10-4s is totally modular 4RU building blocks plugging into standard 40RU rack just like the Power 770/780 CEC's but with 4 times the scalability to 64-sockets and 8x the memory capacity and with Gen3 PCIe. SPARC M10-4S surpasses the Power795 in every respect including 2x more performance. And with SPARC M10 sharing same sun4v architecture as T-Series and soon SPARC M4, you'll be able to do live migration (and do it securely using built in encryption accelerators) between all 3 platforms. Plus they all run the same Solaris 10 or Solaris 11. Binary compatibility assured.
Re: investment protection
SPARC64 is not going to be around? Did you not read the article? Fujitsu=SPARC64=SPARC64VII+=SPARC64X=SPARC M10. SPARC64X is the latest Fujitsu developed SPARC, where the previous two versions were developed for supercomputers -noting that the K computer, based on SPARC64 VIIIfx was ranked #1 world's fastest supercomputer in 2011. The M is the classification of the systems, as in Mission Critical or Mainframe class. These new SPARC M10 servers *are* the true successor to the previous M-Series which were based on Fujitsus SPARC64VII+ in its latest version. Oracle is developing the SPARC M4 which is also Mission critical but as you mention, is based on the successful S3 core that’s also in latest SPARC T4 and SPARC T5 coming out. But SPARC M4 is not a T-Series CPU. Its designed for highest end mission critical workloads and will have different RAS, performance and scalability characteristics.
Re: investment protection
You're blaming Oracle for not having a non Oracle product on the Oracle roadmap? As far as I know, Oracles SPARC roadmap is for Oracle developed products.
Fujitsu has their own development plans and roadmap for SPARC and so maybe you should have the discussion with Fujitsu? Sure, Oracle and Fujitsu cross-sell each others SPARC products, and that SPARC M10 is another SPARC based architecture, but that’s whats great about all of this. Its called choice. Choice of vendors, choice of SPARC technology, yet they remain 100% binary compatible as long as you're running Solaris 10/11. You cant get that with Power nor Xeon for that matter. Unless you consider AMD as reasonable Xeon alternative.
Whats clear is that since the Sun acquisition, Oracle has been *very* transparent in its SPARC developments (making the roadmap public with dates and expected performance increases) and has been very consistent to the plans they laid out back in 2010, delivering almost like clock-work-already 3 new CPU generations delivered and two on the way *very* soon. Not something that IBM or even Intel can claim with their latest slips (i.e.: Power7+ and Sandy-Bridge).
And finally, if you look at the replays of Oracle OpenWorld 2010, 2011, 2012 and especially the latest keynote presentation with Larry Ellison and president of Fujitsu, it was made quite clear that the partnership is as strong as ever and that the two companies are working on many new technologies together like software on silicon. Stay tuned. This is just the beginning and I expect things will get clearer very soon.
Re: Where are the people complaining that 2x-4x performance increase is impossible?
Yeah typical. And no one is questioning why theres not a single *public* benchmark released on Poulson systems and many of the HP Integrity sites are still not showing the Poulson update. Was this rushed to market to make it before end of 2012 so HP/Intel cant be called liars?
Re: It's plenty clear.
Jesper, the roadmap is a processor roadmap and not a system roadmap. The roadmap shows next M-Series CPU @ +6x throughput and +1.5x Thread strength compared to previous generation, CPU to CPU. Oracle is talking about per socket performance. Where do you get comparisons to M9000-64? The linked slides show 5-6x better performance *per socket* than M9000-32.
Re: Take no bets
Atleast Oracle has the balls to put out a public roadmap with anything on it! You don't get that from either Intel or IBM. Isn't the bottom line of CPU advancements all about performance improvements? Well, the way I understand the roadmap, based on what Oracle already announced with SPARC T4, the next T-Series, expected to be the SPARC T5, shows +2.5x throughput and +1.2x single-threaded performance, which I believe would be compared to previous generation (SPARC T4). So maybe its not very detailed on what these really mean, nor what technologies are under the hood, but clearly, a little more than a year later than SPARC T4, you'll see significant improvements in performance-which is quite a feat. if you look at SPARC T3 versus SPARC T4 on real world app performance like Siebel or Java for example, SPARC T4 delivers almost 50% faster Siebel performance and 41% faster Java jvm performance. And seeing that SPARC T3/T4 are still holding several world records on several benchmarks like TPC-C, TPC-H, SPECjEnterprise2010, Specjvm, Oracle E-Business Suite, Peoplesoft, Siebel, JD Edwards, etc, SPARC T5 performance gains is good news.
Re: Take no bets
Bogus FUD. Do you really believe Oracle, with its massive R&D investments in SPARC and software on silicon will stop now? They wouldn't release a brand new public roadmap http://www.oracle.com/us/products/servers-storage/servers/sparc-enterprise/public-sparc-roadmap-421264.pdf
with future CPUs showing up in 2015-2016 if they planned to stop developments after SPARC T5? Larry wants to control the whole stack and make certain its #1 in performance and value to customers. Remember his stated goal. He wants to beat IBM on the high end. And he wont stop till he gets there. And now with Fujitsus SPARC64X getting prepared, IBM's in for a double-wammy.
Re: Smoke and Mirrors anyone?
No SPEC CPU2006 benchmarks released on SPARC T4 because Oracle is focused on developing SPARC processors for systems to accelerate commercial software, not HPC, not games. SPEC is a RAW CPU benchmark that doesn’t test anything but the CPU. SPEC CPU2006 does not test I/O nor networking nor *ANY* SW including the Operating System, database, application server, Java, etc. It doesn’t test encryption/decryption if its on-chip. It doesn’t test flash cache algorithms. It doesn’t test system I/O-critical for database performance. It is good for evaluating the compute performance of the CPU, but not necessarily the system as a whole and Oracle is focused on selling optimized systems, demonstrating the values of the red stack, and not CPUs.
There is little correlation between SPEC CPU2006 results and database performance. You can have leading SPEC CPU2006 results and poor database performance.
If you look at the definition of the SPECint_rate2006 benchmark, its actually based on applications including chess playing, Video Compression, PERL Programming, Event Simulation, Physics: Quantum Computing and other bizarre tests. Certainly not markets that Oracle plays in nor cares about.
Oracle has published over 14 #1 world record benchmarks on the SPARC T4 so far and I don't believe any of them have yet to be beaten by Power7+. If IBM really wants to prove superiority, let IBM showcase this running the same benchmarks and configurations as Oracle. IBM is after all, one of Oracles leading HW/SW partners. http://www.oracle.com/openworld/partners/ibm/index.html
Since 2010, when Oracle acquired Sun, Oracle has released 3 newly developed SPARC CPUs, and in each generation, has proven significant performance improvements well above 50%, in some cases, above 2x. That’s not bad considering that it took IBM over 2 years to update Power7 to Power7+ and show performance improvements of 20-30% and only on two almost identical systems. When you consider that Power7+ has 75% more transistors, 2.5x more L3 cache, 25% GHZ increase and took 2+ years to develop and only get 20-30% performance improvement, you have to wonder? And where are the database benchmarks on Power7+?
And finally, OpenWorld may be coming to an end but Oracle never stated that SPARC T5 would be announced at OOW. 2012 isn't over yet. SPARC T5 is going to knock IBM's socks off.
Smoke and Mirrors anyone?
Well, maybe we need to give IBM the benefit of the doubt since the webcast hasn’t even started yet and most of the launch materials have been released early. What I find amazing are the following!
1) VERY CONFUSING GHZ/cores/socket configurations between Power 770+ and Power 780+ - Time for a new decoder ring! How the hell do you figure out which is best for your workload? More work to feed the IGS folks!
2) The 8-core Power7+ chip runs at a slower clock speed than Power7 (3.7GHz vs 3.86GHz) but the 4-core Power7+ chip runs at faster clock speed than Turbo-core Power7 chip it replaces? (4.42GHz vs 4.14GHz TC). So much for 8-core superiority?
3) Only Power770 and Power780 systems have been updated-Wheres the rest of the product line or will we have to wait a whole year for Power7+ to be released across the product line?
- Theres no Power7+ DCM based systems, No improvements to blades, Power entry and mid range systems untouched-maybe next year?
4) Looks like Turbo-core mode appears to be DEAD!
5) Many of the benchmark results on Power 780 Power7+ were run with 4 cores/CPU or 6 cores/CPU and not 8 cores/CPU, hiding many of the scalability challenges of doubling cores per socket.
6) Theres no database benchmarks (unless you call SAP SD-2 Tier a DB benchmark), no TPC-C, TPC-H, no SPECjEnterprise2010, no Oracle SW benchmarks, no Sybase, No Peoplesoft, No Siebel, no Websphere, etc.
7) Theres no improvements to I/O / PCI-Express - no PCI-E Gen 3! IBM continues to use outdated PCIe Gen2, I/O backplane that was introduced with the model MMC.
8) And heres the kicker. While Oracle continues to get kicked on making benchmark claims, IBM isn't? when stating "IBM Lab testing"
Where are the public proof points on real world (new) benchmarks showcasing the 20-30% performance improvement claimed? Theres no OLTP ore Java enterprise results published so IBM cant prove these statements publicly. Publishing results on 6+ year old benchmarks is lame.
From IBM press release:
"IBM lab testing showed clients have a denser compute node with 20 to 30 percent per core performance improvement and the new POWER7+ processors deliver improved application performance. Expectations are as much as 40 percent improvement with Java based workloads and as much as 30 percent with traditional OLTP and ERP environments compared to POWER7"
@ Jesper Re: The secret sauce is Oracle's storage to compute Infiniband I/O
Yeah, many vendors may be supporting Infiniband connected storage for years but no one besides Oracle has produced an integrated/engineered solution using Infiniband to connect storage to compute nodes (host connectivity) that’s even leveraged by the database ie: flashcache. Even IBM's latest Puresystems don't support infiniband connected storage to compute and neither does HP's cloud matrix or VBLock.
And FYI, IBM's XIV only supported infiniband with its Gen 3 models which was announced last year. Not years ago. And have you checked what a base IBM XIV with just 72TB Raw price lists for? Almost $500K! And for compute? How much for Power7 system with same # of cores as a half rack SPARC SuperCluster (64 x cores)? 2 x Power 750's ( 32 x cores each) list at ~$400K. That’s close to a $1M combined. To put things into perspective, a complete half rack SPARC SuperCluster that includes 2 x SPARC T4-4s plus 108TB of raw data disk capacity plus over 1TB smart flash cache lists for $685K fully integrated.
Re: The secret sauce is Oracle's storage to compute Infiniband I/O
HUH? That’s the whole point! Regardless of whether you compare Oracles Engineered Systems to the current latest competition from IBM or HP or with older systems for consolidation, its clear that Oracles Engineered systems including SPARC SuperCluster have a significant competitive advantage in performance as touted by this and many other reports.
The secret sauce is Oracle's storage to compute Infiniband I/O
Whats interesting in most of these articles questioning Oracles Engineered Systems performance -and like here on SPARC SuperCluster, is that they never mention the one technology that separates Oracle from every other vendor. While all of Oracles competitors continue to fumble with (potentially very fast) storage attached to compute via fibrechannel, limiting the performance potential of the storage, Oracle fully leverages InfiniBand attached storage. Infiniband interconnects move 16x more data than GigE and 5x more than 4Gb Fibrechannel. So, Oracle is in a unique situation having developed InfiniBand attached storage connecting a multi-tier storage architecture. This is where much of the performance differences come from.
Re: IBM do it too...
Did you actually read (and understand) the fine print or just want to FUD Oracle regardless?
Heres the actual text from http://www.oracle.com/us/corporate/press/1851992
"Up to 56 percent lower acquisition cost versus new IBM POWER7 and HP Superdome2-based configurations(1)."
(1) Comparison based on current published list pricing of a half rack SPARC SuperCluster solution: one SPARC SuperCluster half-rack with eight 3.0 GHz SPARC T4 processors, eight channels, eight cores each with built in storage (36 × 600GB) and one Exadata Storage module versus IBM configuration example of the IBM POWER 780: one IBM Power 780 with eight 3.8 GHz POWER7 processors, 1 channel, 8 cores, and 512GB memory and IBM storage DS8700 (384 × 146GB FC, one base cabinet and one expansion cabinet) and HP Superdome2: two Superdome2 each with sixteen 1.73 GHz Itanium 9350 processors, two channels, eight cores, and 1TB memory with HP EVA8000 storage (182 × 300 GB FC, one base cabinet and one expansion cabinet). Actual deployments and configurations will vary.
And the second claim which is about *consolidation*, which in my dictionary, means older systems, usually 3-5 years old. So clearly, that means Power6 systems.
"Up to 40 percent lower five-year TCO when consolidating existing IBM and HP infrastructure, while reducing rackspace footprint up to 73 percent(2)."
Re: Of course there is spare CPU cycles...
Maybe there are just too many ignorant readers? If you read the press release and understood what it says, you'd see the following: http://www.oracle.com/us/corporate/press/1851992
"The new solution helps customers meet service level agreements (SLAs). A half-rack SPARC SuperCluster easily supports hourly throughput of approximately 200,000 Oracle Financials transactions with 2,000 concurrent users in the application tier domain with 80 percent additional capacity available for higher loads or further consolidation"
So * 20 per cent utilization in the application tiers of the cluster* just means theres a ton of headroom left to support even more transactions (during peak demand for example), x more users or run/consolidate other workloads to consume the remaining CPU cycles.
Re: It's hilarious
Jester, Its clear that you’ve been drinking too much of that blue coolaid like TPM, but what I find most amusing that you don't seem to have any clue on what SPARC SuperCluster is all about, but yet you have so much to say against it. Probably a good idea to do some research first before making such uneducated responses.
First of all, SPARC SuperCluster is an engineered system with servers, storage, networking, flash, infiniband, etc that’s all integrated into one rack, all engineered to run any Oracle SW and the 1,000's of apps that support Solaris faster. It cant compare directly to high end SMP systems like SD2, M-Series or Power systems without adding all the other components into the mix. The closest comparison would be IBM's Puresystems which includes a mixture of 2-CPU or 4-CPU nodes, networking and storage. However, Puresystems does nothing about faster performance, just faster delivery times.
Theres nothing more proprietary than running IBM Power which locks you into running AIX or IBM mainframe or running HP-UX which locks you into running Itanium, a dead architecture.
A single SPARC T4-4 (there are up to 4 of them in a SPARC SuperCluster) outperforms Power 750's, Power 770's and even the turbo-core Power 780 on several public benchmarks running DB's, Java and Weblogic environments, and a SPARC T4-4 outperformed a SD2-16 by quite a high margin, so comparing a SPARC SuperCluster which includes 4 x SPARC T4's to a SD2-16 or single Power 780 is really whats hilarious!.
A SPARC SuperCluster with 4 x SPARC T4's, which has a total of 4TB of RAM, is a fraction of the cost of a fully loaded Power770 or Power780 which doesn’t include any additional storage or networking for that matter!
I dare you or anyone else reading this post to do a side by side price comparison of a SPARC SuperCluster versus anything equivalent (from a like for like spec and performance perspective) and prove me wrong - no one offers a better price/performance solution than SPARC SuperCluster for running Oracle SW and even SAP for that matter!
Power7+ DCM looks to have poor scalability
TPM-You missed one big issue regarding IBM's Power7+ DCM performance
Looking at IBM's Power7+ performance (estimation) gains slide, if you compare Power7+ SCM gains versus Power7+ DCM gains, theres less than 40% improvement in performance (Power7+ DCM vs Power7+ SCM) - for what is essentially a doubling of cores/cache/thread per socket. Clearly, the DCM is bandwidth starved and so CPU scalability is impacted.
Estimating the bar chart, I see ~15% faster OLTP, ~27% faster ERP, ~26% faster Integer & ~39% faster Java . 15% to 39% scaling for doubling-up of cores/CPUs doesn’t look like a wise investment unless its priced aggressively. And certainly not for running software based on per core licensing.
Crippling the code? Hah! Itanium users have been claiming this for years!
Or atleast HP can now blame Oracle when Itanium performance sucks running Oracle SW. Either way, Itanium's performance sucks and not even Poulson will be able to save it from despair.
Re: RISC Chips
First of all, I doubt that AIX will go away even if IBM is truly focused on driving Linux on Power. IBM is all about driving services revenue so the more options, OS'es, virtualization technologies, etc, the more complexities, the more chances for them to sell their services. After all, more than 50% of IBMs revenue comes from selling services and this % has been growing every year for quite a long time.
So just as you claim that Oracle wont do anything unless it can make a profit, same is true for IBM and almost any company wanting to be profitable and stay in business. The question is what are these companies doing with their revenues and profit?
If you look at the balance sheets, Oracle appears to be investing a higher % of their revenue on R&D than IBM is (more than 2x from the looks of it), which highlights whos really investing in technologies for its customers.
What I dont agree is your statement on SPARC. Oracle is clearly investing in SPARC as seen by the several SPARC releases since the Sun acquisition and with the upcoming SPARC T5 and SPARC M4, its going to pose new and bigger threats to Intel and IBM. Oracle would have put a public roadmap out there on SPARC unless it was serious about delivering-after all, its reputation would be on the line.
With Oracles leading SW marketshare, offering a leading CPU architecture to run it is a win win for Oracle and its customers.
So while IBM may be making higher revenue on Unix than Oracle currently is, that just shows you who's making more money off of its customers. What I believe is the more important indicator is the volumes, and SPARC is still leading Power and Itanium in volumes.
For Oracle to succeed, it needs SPARC (and x86) to win. SPARC for a truly engineered end to end environment, and x86 for everything else.
Re: Red dwarf
Yes its true. Sun died a long death after the dot com era because they couldn’t figure out how to make money off of giving away software for free (Solaris, Java, Openoffice and many others) and the hardware business that they were driving to, was becoming more and more commoditized -it came to a point of trying to sell hardware without much differentiation (especially x86) and making razor thin margins which made it difficult to remain profitable and invest in R&D. IBM has reinvented itself by changing from a HW focused company to a services focused company selling hardware and software to drive services. Redhat, HP and many others are focused on the same business model. And as long as IT complexity remains king, services revenue will continue to grow. Oracle is headed in a different direction of trying to eliminate the need for many of those services by engineering the entire stack and selling you a Red box that’s easy to manage and maintain over the lifecycle, reducing management and administration costs a long the way.
Figures don't lie, but liars figure
I agree. This whole analysis is fluff. The bottom line is that Oracle continues to be profitable and continues to drive profit and revenues on both HW and SW. Selling other peoples products & technologies doesn’t fit into this equation and trying to figure out how much revenue (and if there any loses compared to pre-Sun days) has been made on HW is irrelevant. Why hasn’t there been any analysis on IBM's massive mainframe revenue declines or HP's Itanium declines vs PA-RISC days? Its clear it’s a new world with new trends, including the cloud and convergence and clearly Oracle has turned around the Sun HW part of the business and is remaining successful with continued growth after the acquisition.
Re: What *has* Oracle done with Sun?
The question should be-what has Oracle done with Sun (since the acquisition)? As far as I am concerned, its converted a failing business model of selling HW at razor thin margins and giving away software for free(Java, Solaris, OpenOffice, etc), which might have drove high revenue (for HW atleast) but drove very little profit (hard to invest in R&D when theres little profit) to a very profitable business, selling only select HW that is viable (SPARC & Eng systems), that’s competitive and that help drives Oracles business model of making profit (hardware and software engineered together) so it can invest in R&D to innovate and continue to compete. Oracle has disinvested in selling x86 servers which have no margin(but still investing in Exa* systems based on x86), selling other companies products like HDS, Fujitsu etc. This clearly has an effect on HW only revenue, but if you look at the profit of the company, profit from HW, its margins and increases in revenue of Oracle, Oracle continues to grow at strong rate as seen by its latest financial reports. The trend is clearly towards converged systems and Oracle is clearly a leader in this trend.
Re: Oracle could stop ignoring customers...
Sounds like you're not calling the right folks at Oracle-selling maintenance is quite high on Oracles priority list just like it is for IBM, HP, Redhat and others where profits are high.
As for the new world, see my post at end of this comment section
Re: Seems like IBM has just redefined the word complexity
I guess you haven't looked through IBM's full announcement. IBM's Pure Application Systems is *exactly* what Oracle has been doing with its *Exa* systems. http://ibm.co/HCmK34
Its essentially a PureFlex (Blade) system with IBM bundled middleware including DB2, Websphere, Tivoli, etc. So, IBM is validating Oracles approach to selling software through hardware. And as far as I can tell, the Pure Application Systems don't support Oracle SW ;-) Oracles SPARC SuperCluster Engineered System runs full suite of IBM SW as well as SAP and other 3rd party ISV's.
And as for IBM services attached to these new "Pure" systems, I believe the marketing term IBM is using "Built-in expertise" translates to IBM Services. Looking at the Pureflex details, you'll see some special services included “Lab Services” and “Acct Advocate” and “3 yrs w/1 microcode/yr". Doubt this is done by VARS.
Re: Seems like IBM has just redefined the word complexity
No flaws in my arguments as far as I can tell but I do agree with you that software also dries massive profits for IBM. I was stating revenues first of all and the statement I countered was "... profits from systems".
Looking at IBM's financial report for 2011, states "More than 90 percent of our segment profit in 2011 was from software, services and financing". Don't see systems anywhere in that sentence.
And I think your statement "IGS is often sold on a cost basis with little or no profit at all just to drive hardware and software sales" is backwards. IBM sells its hardware and software to "DRIVE" its services and hence why I referred to the Trojan horse approach,. If you look into the details of what these new "IBM PureFlex Systems", they are ingeniously tied/locked to IBM SW and services. Its taken me all day just to figure out what is actually new and offered here. Maybe I should call IBM services for help. Oops, its gonna cost me.
Seems like IBM has just redefined the word complexity
First, its interesting to make such a misleading statement as "IBM still derives the bulk of its prestige and profits from systems" but then again, perception drives reality .
Has anyone looked at IBM's latest Financial reports? http://www.ibm.com/annualreport/2011/bin/assets/2011_ibm_annual.pdf
>60% of IBM's annual revenue is derived from selling services and is growing rapidly every quarter while Systems and Technology is just ~18% of IBM's revenue stream. IBM's business goal is to sell services where the software and "technology" will drive more services.
So clearly this announcement isn't about better technologies to make it easier, less complex for data centers, otherwise, these customers wouldn’t need any of those expensive (IBM) services. Too much choice, too much flexibility, too many options only increases complexity. It’s the IBM Trojan Horse. After all, the code name of this project "Project Troy" is quite meaningful. http://en.wikipedia.org/wiki/Trojan_horse
So although, these new PureSystems sound "great on paper", the reality is, you're gonna pay for IBM services to get these systems installed, working and running. Otherwise, IBMs IGS would be impacted.
Re: shiny but $$$$
You get what you pay for. And I'm sure you're driving a Yugo too http://en.wikipedia.org/wiki/File:Yugo-US-poster.jpg
Re: Oracle didn't buy Sun for the hardware business
Maybe its you that doesn’t have a clue about where the financial industry is moving to?
Gartner has published their market share numbers for 2011 based on total software revenues. According to Gartner, Oracle
• Is #1 in worldwide RDBMS software revenue share
• Holds more revenue share than its seven closest competitors combined
• Grew at 18.0 percent, exceeding both the industry average (16.3 percent) and the growth rates of its closest competitors.
And sure looks like Oracles hardware is selling quite well in the financial sector.
Re: "Ramp-up" or selling
A few corrections here.
If you look at all the earnings call transcripts, you'll see the following statement that if you read it, would realize that further #'s are never guaranteed and clearly things can change.
"As a reminder, today's discussion will include forward-looking statements, including predictions, expectations, estimates or other information that might be considered forward-looking. Throughout today's discussion, we will present some important factors relating to our business, which may potentially affect these forward-looking statements. While these forward-looking statements represent our current judgment, these statements are also subject to risks and uncertainties that may cause actual results to differ materially from statements made today. As a result, we caution you against placing undue reliance on these forward-looking statements"
After all, if we believed IDC, Itanium would be a $100BN+ business by now, considering IDC forecasted $30BN+ in Itanium sales by 2002 with a massive ramp!
And you'll find quite a few mis-leading statements from the Intel leader at the time who's now leading EMC.. Interesting.
I guess its all marketing after all. And many will say "Figures don't lie, but liars figure"
So in the end, the proof is in the pudding http://blogs.barrons.com/techtraderdaily/2012/03/20/oracle-fyq3-beats-2/?mod=BOL_qtoverview_barlatest
Wonderbra-whats up with your rant on SPARC?
Its clear you are an IBM gnome. Why would you say that Oracle will pull the plug on SPARC anytime soon?
Since Oracle published the *public* SPARC roadmap, now 2 years ago, they have delivered on every milestone and even Larry, yesterday, stated that the future of Oracle is with SPARC. Future SPARC CPUs will integrate Oracle SW components, which is where Larry believes he can innovate and differentiate.
With over a 1,000 engineers, http://www.oracle.com/us/corporate/innovation/innovator-hetherington-191304.html, Oracle has even pulled in several developments like the SPARC T5, now expected end of this year and the performance gains on SPARC T4 60% greater than what they promised, beating Power7 and Westmere-EX on many relevant benchmarks.
Its clear that Oracle is having difficulty increasing server revenue, especially when a new SPARC T4 system is a fraction of the cost of the older M-Series systems, competing against Xeon on price, while they are less than half to a third the cost of an equivalent IBM Power system but I think with the SPARC T4 systems ramping in volume, http://www.marketwatch.com/story/oracles-sparc-t4-servers-in-high-demand-in-the-communications-industry-2012-01-31, and SPARC SuperCluster showing massive ramp up as well (Oracle announced a 139% revenue increase in engineered systems yesterday), the revenues will be returning very soon. As Mark Hurd and Larry Ellison said on yesterdays great Q3FY12 performance, Q4 and FY13 will be the year of Oracles Systems.
- The land of Milk and Sammy: Free music app touted by Samsung
- The long war on 'DRAM price fixing' is over: Claim YOUR spoils now (It's worth a few beers)
- Privacy warriors lob sueball at Facebook buyout of WhatsApp
- Dell thuds down low-cost lap workstation for
cheapfrugal creatives or engineers
- 20 Freescale staff on vanished Malaysia Airlines flight MH370