Last message on previous page: I have a question for you guys, I was talking to a friend last night in regards to the intel xeon 3.2ghz 130W dual core processor 65nm.
On paper this looks like it is faster than the core 2 duo extreme however I am sure that this probably isnt true based on the prices.
Can someone explain to me what is a faster chip for a workstation pc between say the xeon I mentioned and a core 2 chip (not core 2 extreme perhaps the E6400 or something like that which is in a similar price range to the xeon that I mentioned above.
3.2ghz with 4mb cache sounds good to me but im guessing there is something in the cpu architecture of the xeon which is either older or doesnt run as fast or something when compared to the cor duo's
I have a question for you guys, I was talking to a friend last night in regards to the intel xeon 3.2ghz 130W dual core processor 65nm.On paper this looks like it is faster than the core 2 duo extreme however I am sure that this probably isnt true based on the prices.
No, it is based on the older Netburst architecture.
Quote :
Can someone explain to me what is a faster chip for a workstation pc between say the xeon I mentioned and a core 2 chip (not core 2 extreme perhaps the E6400 or something like that which is in a similar price range to the xeon that I mentioned above.
On pure CPU performance, the E6400 is probably 10-20% faster than the Xeon 3.2GHz(Dempsey), from what I can see in a quick lookup in THG's CPU guide.
But the Xeon platform (MB and RAM) is much more expensive than the Core 2 platform. So, for the same money, you will get much more performance from a Core 2 platform. If you need a two processor platform or masive amounts of RAM (Over 4GB), then consider the slower clocked 5100 series Xeons which are Core 2 based. They will still be faster than a Netburst based Xeon.
I just wonder if someone has took a look on the 4-way (dual core) servers. It seems Intel has not yet made available its micro-architecture to its Xeon MP series.
As a result, for 4-way dual-core servers, AMD still beats Intel so hard.
Please look at earlier discussion in this thread before posting. Yes, XeonMP is based on Netburst, and therefore being deficient. Tigerton is Intels codename for the Core2 based XeonMP 4 socket, 16 CPU :!: platform.
Thanks. And, I know what you said.
I just wanted people know the current situation. Tigerton has not yet available to beat Operton in 4-way Dual core servers.
Hello THG readers. I must agree with most of you: This review/benchmark is a waste of time. First of all, the benchmark wasn't fair. Probably they had no idea that the new AMD Opteron 1000/2000/8000 series are already out.
Second of all, real benchmarks include Unreal Tournament 2006 (is that the latest version? I'm not a gamer, sorry ) and rar/unrar some huge ass sourcecode trunk, like firefox or similar.
Another thing that I cannot understand is why people consider Opterons and Xeons "server" processors. most of us know that they're just desktop processors, SMP-capable and sometimes with a higher FSB speed and some improvement brought to the chipset (buffered/unbuffered ECC/NON-ECC memory). You can always throw a some 7800GTX videocard in a dual Xeon 5160 and game your brains out.
off-topic: It's true that the average user prefers 'desktop' processors wrapped in a nice box along with motherboards from DFI or ASUS which always have hentai/<insert animation pictures here> on them. Oh, lets not forget about the video cards and especially the pictures on the GPU coolers. the more colors you have in those pics, the cooler your GPU is going to be.
Personally I think DP systems are actually suited for desktop usage/workstation and even small servers.
:?: Are your tests valid? Did you misconfigure the SuperServer 6015B-T?
As show in your picture, the DIMMs were installed in slots A1, A2, B1, & B2. According to SuperMicro's manual, the DIMMs should have been in slots A1, B1, C1, & D1. Otherwise you would have disabled interleave. I don't know how much performance was lost.
By the way, this is an issue with all 5000? MCH chips.
the article was fine but i havent seen any articles relateing to more then 2 cpu systems like
Processor: Dual-Core AMD Opteron™
Socket: 940
Model Number: 885
Frequency (GHz): 2.6
L2 Cache Size: 2 MB
Stepping: E6
Manufacturing Tech (CMOS): .09 micron SOI
Wattage (W): 95.0
Integrated Memory
Controller (GHz): 2.6
Number of Cores: 2
the last one i saw here was a year ago i think puting AMD in the lead and they were useing vertual ? along with 4 cpu's
intel has brought out some excelent new arcetecture but i think its just catch up
what wood you use more then 4 cpu computer for anyway ?
Always asking "why is it so"
They use the Opteron 880, which is a little bit slower than the 885. You can extrapolate based on the clock speed difference (2.4 / 2.6GHz)
Quote :
the last one i saw here was a year ago i think puting AMD in the lead and they were useing vertual ? along with 4 cpu's
intel has brought out some excelent new arcetecture but i think its just catch up
In 4+ CPU configurations AMD wins as the Intel Xeon MP chips is still Netburst based. The latest generation Xeon MP is doing quite well for a Netburst, even winning a few benchmarks. But there is no doubt that Opteron is still the best if you need a server with more than 2 CPU's.
Though, with the just introduced Clovertown Xeon DP's it gets less relevant to go the four CPU socket route. With two clovertowns you will have a 8 core server that will be better performing than a four socket Opteron system. If you need more than that, then wait for the Tigerton platform (Core2 based Xeon MP's) in Q2 2007. I'm sure that AMD in the same timeframe will introduce quad core Opterons, so they also can deliver a 4 socket, 16 CPU system.
Quote :
what would you use more then 4 cpu computer for anyway ?
When faced with heavy server workloads, the first option should always be to scale out - dividing the work among multiple servers. If that is no longer possible, you will have to scale up using 4 socket systems. The most prevailant application on 4 socket systems is database servers.
Another application for the workstation crowd is HD video rendering, which can be done on big workstations or rendering servers.
Also 4 socket systems are typically build with more redundancies (dual powersupplies, RAID RAM and so on) for better reliability.
what wood you use more then 4 cpu computer for anyway ?
Always asking "why is it so"
Any application server with a reasonable amount of load will use not only 4 but 444 CPU's without an issue.
Database servers (like mentioned above), Internet application servers (ASP.NET, etc.), other application servers based on Citrix are typical examples where multi-cores really make a difference.
Multiple cores improve timely responsiveness of a server. Server loads are not uniform. A single-core powerful CPU versus a less powerful multi-core CPU during a peak load is like the difference between having a Porche or a Bus where you need to rush 20 people to the stadium before the game starts.
A single-core powerful CPU versus a less powerful multi-core CPU during a peak load is like the difference between having a Porche or a Bus where you need to rush 20 people to the stadium before the game starts.
I'm sorry, I'm talking to myself but following the same analogy, the difference between an Opteron and any Intel CPU is like Intel's bus' having a much narrower door for people to get on or off.
Sorry guys, I know a lot has already been said on this article, but I can't stand this kind of tests. It (so conveniently) backs up Intel's marketing arguments whithout any clear objectivity, and most of all, ends with a biased purchase advice !!
Facts :
1- you test last generation of processors for one brand and not the other (why not Opteron 2000 family ?)
2- consequence of point 1, one system runs last generation of RAM chips and not the other
3- you don't test both systems with an OS that we're sure not to be Intel optimised
4- you test so-called server machines with apps I've never seen running on any server in any company that I've worked for in 10 years. Did you intend to bench geek machines ?
5- you test with Intel's own compilers ................ no comment !!!
6- you make a test with a software you recognise as BEING Intel optimised : again, useless for comparison between Opteron and Xeon (which is the goal of the article !!)
7- The power consumption test is totally unclear. What do you measure exactly : procs only, procs+chipset, procs+chipset+RAM, system's total ?
8- Why do you try to apologize for Intel's FB-DIMM high power consumption ? It's the technology Intel have chosen, period !!
Considering those points, my conclusion is that this benchmark is incomplete and partial !!
"Intel has retaken the throne from AMD in the server/workstation sector and sent AMD back to the drawing board." .... man, what a conclusion !!!!
I know it's hard for all Intel fans to admit it, but their favorite company had never considered AMD as a real threat in the server segment and sat on her laurels for many years. Actually, I think Intel has great engineers but they waste their talent !!
Once again, they've managed to gain back the first position in the desktop processors' race. It's just the race that's talked about most, but it's not the most important to be won !!
As far as I know, for the Core design, Intel has enhanced the processor architecture, the efficiency of internal caches and pipelines, of predictive algorithms, ...
But how long do you think one can keep an old design at the top by "patching it" or spending $ billions to get the latest micro-electronics state-of-the-art technologies ?
What happens when the competitors get them too ?
Woodcrest 80% more powerful than Dempsey ?
OK, so here are some simple maths :
Woodcrest_Proc_data_processing x N processors --> Woodcrest_NB-FSB_data_processing = Dempsey_NB-FSB_data_processing x 1.8 x N x A
"A" being a multiplying factor taking into account the increasing switching time lost by the NB, managing multiple data clients (procs), as their number and individual data processing capabilities increase.
I really don't think one can keep raising indefinitely the factors of such an equation !!
My point is that they'd do better focusing on building something really new, like the Itanium platform was meant to be !!
What a lot of non-sense. Talk about shooting the messenger instead of accepting the message. There is NO competition between Woodcrest and Opteron. Woodcrest is better on every benchmark. There is no attacking the benchmarks which can explain that fact away.
Well .... sure, you can make Woodcrest beat any Cray supercomputer in a bench .... depending on how you present the facts or how you do the tests !!
I'm not saying Woodcrest superior performance is a lie ; all I'm saying is :
1- it's temporary (as usual in the processors war),
2- you can't shout Intel's sweeping away all competitors on all figures, by making a biased comparative test !!!
You should know by now that there's no definitive solution, no miracle processor that's superior in every way : all is about what you intend to do with the machine.
I criticised that bench for lacking this objectivity and being all dedicated to the glory of "Almighty Intel" !!
I also criticised the way in which they bench a server board like a PC board. This gives all geeks that don't have the "professional point of view" (the real world of enterprises) the impression that Intel is going to put AMD out of the server business : it's absolutely not true. I call this "Intel evangelism" and it's not acceptable for a site like THG.
I tried then to explain why in my opinion, Intel's design will have to be reviewed deeply to be able to kepp up with the competition in high-end multiprocs servers (at least four procs I mean).
But obviously, I'm facing blind Intel fans out here, so you guys believe whatever you like ..... I won't argue any further !!
