| ActiveMac | Anonymous | Create a User | Reviews | News | Forums | Advertise |
|
|
|
User Controls |
|
New User |
|
Login |
|
Edit/View My Profile |
|
|
|
|
|
|
|
ActiveMac |
|
Articles |
|
Forums |
|
Links |
|
News |
|
News Search |
|
Reviews |
|
|
|
|
|
|
|
News Centers |
|
Windows/Microsoft |
|
DVD |
|
ActiveHardware |
|
Xbox |
|
MaINTosh |
|
News Search |
|
|
|
|
|
|
|
ANet Chats |
|
The Lobby |
|
Special Events Room |
|
Developer's Lounge |
|
XBox Chat |
|
|
|
|
|
|
|
FAQ's |
|
Windows 98/98 SE |
|
Windows 2000 |
|
Windows Me |
|
Windows "Whistler" XP |
|
Windows CE |
|
Internet Explorer 6 |
|
Internet Explorer 5 |
|
Xbox |
|
DirectX |
|
DVD's |
|
|
|
|
|
|
|
TopTechTips |
|
Registry Tips |
|
Windows 95/98 |
|
Windows 2000 |
|
Internet Explorer 4 |
|
Internet Explorer 5 |
|
Windows NT Tips |
|
Program Tips |
|
Easter Eggs |
|
Hardware |
|
DVD |
|
|
|
|
|
|
|
Latest Reviews |
|
Applications |
|
Microsoft Windows XP Professional |
|
Norton SystemWorks 2002 |
|
|
|
Hardware |
|
Intel Personal Audio Player 3000 |
|
Microsoft Wireless IntelliMouse Explorer |
|
|
|
|
|
|
|
Site News/Info |
|
About This Site |
|
Affiliates |
|
ANet Forums |
|
Contact Us |
|
Default Home Page |
|
Link To Us |
|
Links |
|
Member Pages |
|
Site Search |
|
Awards |
|
|
|
|
|
|
|
Credits |
 |
Unix reclaims server speed crown | |
|
||
| Time: 18:08 EST/23:08 GMT | News Source: News.com | Posted By: Jonathan Tigner | ||
|
IBM's p690 Turbo Unix server, newly overhauled with faster processors and other components, posted a score of 681,000 transactions per minute on the Transaction Processing Performance Council's TPC-C test. It edged ahead of the 658,000 score posted two weeks ago by Hewlett-Packard's Superdome running Microsoft's Windows Server 2003 operating system. "I saw (Microsoft CEO) Steve Ballmer stand in front of a German audience and crow that they had TPC-C bragging rights. Well, they don't," said RedMonk analyst James Governor. Microsoft isn't the only company taking a hit from the test result. Big Blue has long used Oracle's database software for its TPC-C testing, but in a significant strategic shift, the company this time used its own DB2 product. "It certainly underscores software's ascendance within IBM, not just from a marketing perspective but from a technology perspective," Governor said. "If the database wasn't good enough, IBM server folks would not be pushing it." The move also pulls the rug out from underneath Oracle's marketing effort, which will no longer be able to boast that IBM was forced to use Oracle's software when trying to wring the last scrap of performance out of a system during benchmark testing, he added. | ||
|
Write Comment
Return to News |
Displaying 1 through 25 of 309 Last | Next |
The time now is
5:29:57 PM
ET.
Any comment problems? E-mail us |
| #1 By 20 (67.9.179.51) at 5/9/2003 6:37:38 PM |
|
Unix vendors and Oracle: "Look at the TPC! We are awesome! M$ sux0rs! TPC is a high-quality, trusted benchmark and illustrates our performance superiority"
(MS destroys Unix numbers on the TPC and takes almost all the top 10 slots) Unix vendors and Oracle: "MS bought their way in, the TPC are bunch of a hacks and they are on the take. Besides, TPC doesn't prove anything, it's a useless, slanted, and unrepresentative process. We are still king!" (A Unix box retakes the #1 spot) Unix vendors: "See! TPC is awesome, we are number 1! M$ sux0rs! They cheated and the TPC is bogus for them, but it's the shining light of truth for us!" |
| #2 By 3339 (65.198.47.10) at 5/9/2003 6:37:39 PM |
|
These are as much a test of the processor as they are of the operating environment and database software. Why would IBM test its stuff on Intel processors, running SQL server?
HP did their most recent tests just two months ago--they don't have faster processors yet--the new Power4++ kicks ass... as it is outperforming the PA RISC chips at only slightly higher cost but with half the processors. (Which means much bigger savings in the long haul--maintenance- and license-wise.) HP is scheduled to retest using Unix. How will you feel if they beat their MS test results? |
| #3 By 16302 (207.195.39.195) at 5/9/2003 6:39:26 PM |
| It should be noted that the IBM server is only the fastest 'non-clustered' server; the HP ProLiant server running Windows 2000/SQl 2000 is still the king of the heap with 709,200 tpmC. |
| #4 By 20 (67.9.179.51) at 5/9/2003 6:46:29 PM |
|
#3: I don't think anyone said that MS has completely removed Unix from the datacenter, yet that's what you imply (in yet another strawman for your weak arguments).
The fact that MS is even on that list, let alone #1 is a huge deal. It proves that MS is hear and it's a competitor in the Unix market place and, in many cases, outperforms Unix for a fraction of the cost. Note that the Unix solution only slightly edges out the Windows box, but for an extra $1.121 million. I bet if HP spent an extra $1 million, they could easily make up 22,336 tpmC. So, again, Unix barely eeks out it's seat at the top of the TPC, but at extrordinary cost and not much higher results. |
| #5 By 20 (67.9.179.51) at 5/9/2003 6:48:22 PM |
|
Wonder when .NET might start showing up on here?
I bet it could make a showing on the top 10, though probably at the bottom to start with. Though, on Windows 2003, .NET works significantly faster. |
| #6 By 3339 (65.198.47.10) at 5/9/2003 7:03:32 PM |
|
" I don't think anyone said that MS has completely removed Unix from the datacenter, yet that's what you imply (in yet another strawman for your weak arguments). "
What the hell you talking about, daz? How did I imply that MS is nonexistent in the datacenter world? I said: "These are as much a test of the processor as they are of the operating environment and database software. Why would IBM test its stuff on Intel processors, running SQL server?" Nope, don't see any implication there. How 'bout: "HP did their most recent tests just two months ago--they don't have faster processors yet--the new Power4++ kicks ass... as it is outperforming the PA RISC chips at only slightly higher cost but with half the processors. (Which means much bigger savings in the long haul--maintenance- and license-wise.)" Nope, that's all hardware talk mostly. I mention licenses--does that mean MS doesn't exist in the datacenter to you? Weird. To me: nope. Oh, I see. This part: "HP is scheduled to retest using Unix. How will you feel if they beat their MS test results?" Means MS doesn't exist in the datacenter? Whatever. If that's the type of interpretation you make, there's nothing I can do to help you. "Note that the Unix solution only slightly edges out the Windows box, but for an extra $1.121 million." I guess with all my wild implications you missed the part where I mentioned this. It also did so with half the number of processors. If you (and anyone else) can't tell, I'm largely concerned about the whole package: the OS, the hardware, and the DB. I didn't see where I said much at all about MS. |
| #7 By 415 (199.8.71.121) at 5/9/2003 7:20:01 PM |
|
The way you guys bicker just cracks me up! Thanks for the entertainment!
I'd rather buy a Windows server and keep the cool million in my pocket, 'ah thank you! |
| #8 By 61 (65.32.171.144) at 5/13/2003 11:03:46 AM |
|
Jerk:
The Power4 processors have 2 cores per die... essentially it is 2 processors on one chip. So it's not truely 1/2 the number of processors. |
| #9 By 3339 (66.219.95.6) at 5/13/2003 5:50:00 PM |
|
parker, that comment was in response to daz who was looking for some sort of apples to apples comparison when clearly in these tests IBM is trying to rate its hardware, its OS, and its DBase. Of course, IBM has many configs--but if IBM wants to say our complete package is the best, why would they use Intel procs and SQL. That was the point, slow one.
CPU, sure, it's true. You buy Power4's by the chip, not by the core. You cannot buy one core. The cost of the chip is comparable to similar high end chips. And there are tests with Intel procs using multithreading, pretending that they are more chips. Just because a chip has a particular advantage/architecture doesn't mean you get to count it as a disadvantage to benefit the losing processor. There are numerous multi-core chips, but I have never seen them counted as if each core is its own chip. |
| #10 By 61 (65.32.171.144) at 5/13/2003 7:27:27 PM |
|
jerk:
The fact that it has 2 cores per die means it is the equivelant of having 64 procs. instead of just the 32 it states. In other words, the IBM machine is not doing more work with half the processors. |
| #11 By 3339 (66.219.95.6) at 5/13/2003 7:55:47 PM |
|
But you are ignoring all of the benefits of reducing the number of processors--you are basically saying: those individual chips are as powerfl as two chips so I'm going to call them two chips." Multi-core means less chip interconnects, and fewer bus issues. This is cost savings--maybe if this escalated the price dramatically, there would be a need to acknowledge it, but the fact is there are substantial cost savings. The chips (with multi-cores) sell for a price comparable to chips with equivalent capabilities (whether or not they are single or multi-core).
There are plenty of multi-core chips out there--I have never seen a single benchmark refer to them, treat them, or even note them as more than one chip. Anyway, these benchmarks aren't just about performance. Cost and maintenance are huge factors as well--maintenance of fewer processors is a plus; the cost of licenses for fewer processors is a plus. (That's also worth noting: per processor licensing for multi-core chips doesn't multiply per core--it's still one chip.) If you'd like to tell yourself that a multi-core chip isn't one chip but many and that the performance of these two servers was the same, okay, enjoy yourself. |
| #12 By 61 (65.32.171.144) at 5/13/2003 10:37:15 PM |
| jerk: Reguardless of how you look at it, it's still essentially the same as running 64 separate chips that only have 1 core per die. |
|
Write Comment
Return to News |
Displaying 1 through 25 of 309 Last | Next |
The time now is
5:29:57 PM
ET.
Any comment problems? E-mail us |
