Sign in with
Sign up | Sign in
Your question

Conroe loses to X2 3800 in Multitasking Benchmarks?

Last response: in CPUs
Share
June 6, 2006 5:42:38 PM

Given that even today Intel still will not release samples to the community (Yes, Tom's recent article was on a machine provided by Intel where again, they were allowed only to push buttons), you know something has to be hiding inside all these Conroe benchmarks. It seems that a Chinese site got hold of a Conroe E6300 without Intel around and ran his own multitasking tests. And low and behold, Conroe loses to the X2 3800.

In SysMark 2004 test, the Con E6300 got 210 points. In comparison, an Athlon 64 X2 3800+ got 232 points. That's a 10.5% AMD win using Intel's own benchmark that Intel developed.

The chip.
http://detail.zol.com.cn/picture_index_51/index500822.s...
Conroes Result.
http://mb.zol.com.cn/29/295247.html
X2 3800's Result.
http://reviews-zdnet.com.com/Polywell_MiniBox2_939NP/45...

Now of course this is still only preliminary, but I suspect that when these chips are finally available, that Conroe will do awesome on the synthetic and single thread benchmarks that use less than 4MB of cache, but in real world multitasking and multithreaded benchmarks, both cores of the Conroe are fighting to use the same cache and performance then plummets. If this holds true, then it would make for very interesting discussions because most games are still single threaded, but more games are becoming multithreaded. Also, most everyone likes the idea of maximizing their multitasking performance and if Conroe's multitasking performance is weak it poses a big negative, even to some light gamers who enjoy multitasking at the same time.

Ask yourself this.
Why won't Intel just release the damn samples already!? Is it because they know there are holes in its performance and it won't be as good as the marketing hype that they've generated so they are hiding it for as long as possible until actual launch? I say they have it available so ship out the samples!!
June 6, 2006 6:05:22 PM

It could be competitive advantage. Maybe Intel does want their chip in AMDs hands. Then when they release they have just that much longer to be the king.

Heck, maybe they'll release the "tops" model that will be evem faster than the current rumors to even moreso overshoot AMD. Right now AMD has an idea of how fast Conroe is so they are planning ahead to outbest Intel. But if Intel launches with something even better it might outbest even what AMD has coming up next.

All conspiracy theory!
June 6, 2006 6:14:24 PM

The multitasking performance really is interesting...

Quote:
Why won't Intel just release the damn samples already!? Is it because they know there are holes in its performance and it won't be as good as the marketing hype that they've generated so they are hiding it for as long as possible until actual launch? I say they have it available so ship out the samples!!


However, I don't see why Intel has to officially release "samples" to everyone so soon.

Did AMD release AM2 samples 1 month before launch? Did they release K8 samples back before they were launched??

Does any manufacturer do that? Does nVidia? Or ATI?
Related resources
June 6, 2006 6:15:51 PM

Interesting data mike. Hopefully we will see some reputable 3rd party reviews soon. This wouldn't be totally surprising after the rosy picture they painted for Prescott after all.

Of course it also wouldn't be the first time that the best cpu for you may have more to do with the way you use your computer than anything else.

Don't know if your theory will carry over to multi-threaded stuff as conroe is handily winning the encoding benchmarks. May just be a multi-tasking issue as opposed to a multi-threading issue.

There will probably be new versions of the popular benchmarks released soon that are compile optimized for core 2 uarch anyway.
June 6, 2006 6:19:33 PM

I'm a little suprised that I don't see an onslaught of Intel fanboys raiding this post yet.
June 6, 2006 6:23:16 PM

Quote:
Interesting data mike.

Hey, thanks for your input.
But truly, I am not this Mike guy. :roll: :lol: 
June 6, 2006 6:23:19 PM

The fact that they share the 4MB cahce is an advantage not a dissadvantage, because they can both access it simultaneously, and there is no nead to hike across the hypertransport bus to peak at what the other processor has in memory. This comes in handy when you're running a multithreaded application because you halve the cache requirement. In straight up multitasking it's mot a dissadvantage, or advantage because the cache is accessible by both cores simultaneously.

One rumor going around as to why the chips arn't being released is because the chipsets are not quite ready yet. Then again, who knows? Conroe will be better than the A64... big whoop the A64 has been around since what 2003 and it's taken Intel 3, years to get off its but, and start competing. AMD will have a comperable part in less than a fith of that time.

Be happy A64 prices have been stagnant for to long. AMD needs a kick in the but, just like intel did. Both sides are going to be fighting kneck and kneck soon, and we'll all be better off for it.

We're going to see solutions from both sides that perform better, run cooler, suck less juice, and cost less. Make no mistake Conroe will kick ass but so will AMD's eventual response. A consumer level chip with 4MB of shared cahce is a good thing.

-peace
June 6, 2006 6:25:19 PM

Quote:
Did AMD release AM2 samples 1 month before launch?
Yes.
June 6, 2006 6:26:40 PM

...

Actually there are some Core 2 Duo samples outhere...
June 6, 2006 6:31:55 PM

Quote:
Did AMD release AM2 samples 1 month before launch?
Yes.

OK, it must have slipped my attention.

Can you please provide some links that review AM2 performance one month before its launch?
June 6, 2006 6:48:06 PM

Hey LMM, why don't you get a life?
June 6, 2006 6:52:44 PM

I have to ask.
Why is it everytime someone posts sumthing PRO-AMD u say hes lmm or mmm. But when some one post pro-intel you like "ok"?
June 6, 2006 6:54:03 PM

Quote:
Hey LMM, why don't you get a life?


Ahh cut him some slack. If it truly is MMM he will certainly self destruct sometime in the next week or so. If he manages to remain civil for that period, I will conclude that it is probably not Mike after all.

No offense to you itguy, its just that MMM was such a firm advocate of the conroe cache thrash coverup conspiracy
June 6, 2006 7:05:03 PM

Quote:
I have to ask.
Why is it everytime someone posts sumthing PRO-AMD u say hes lmm or mmm. But when some one post pro-intel you like "ok"?


Actually now that I think about it, MMM is probably the most hardcore devoted AMD fanboy ever. So for your average AMD fanboy to be identified with MMM, that could be considered quite a compliment. Although the real MMM might take offense.
June 6, 2006 7:06:19 PM

Quote:
Ask yourself this.
Why won't Intel just release the damn samples already!? Is it because they know there are holes in its performance and it won't be as good as the marketing hype that they've generated so they are hiding it for as long as possible until actual launch? I say they have it available so ship out the samples!!


I would bet that MB manufacturers and other industry partners have had samples for months. I think its the the NDA that your referring to. In fact Anandtech reports that some MB manufacturers have working Kentsfield samples already. The article in Tom's clearly says that the benchmarks were performed months ago, but they were just now allowed to release the results.

http://www.anandtech.com/tradeshows/showdoc.aspx?i=2770...
June 6, 2006 7:23:22 PM

Stupid AMD Fanboy. When you post two test results. Are you sure they are the same test.

Sysmark 2004 (partial scores) vs Sysmark 2004 SP2 (overall score)

Not the same. You can't compare results from different test.

My computer is faster at SPECviewperfect 7 than NASA computer in SPECviewperfect 9
a c 99 à CPUs
June 6, 2006 7:34:01 PM

No, if you really were MMM, you'd have written something like "AMD pwns Intel's CON-roe AGAIN!! Intel is teh crapper!!!!"

And this is not very surprising. The Core 2 generally ties the Pentium D 965EE in multitasking benchmarks, as AnandTech showed (I'd post a link, but I can't seem to find it as AnandTech's news stories from the previous days can be tough to find.) The X2s beat the Pentium Ds in most multitasking benchmarks, and if the Core 2 is roughly equal to the Pentium D, well you can do the math.

Hmm. I wonder if this was a reason that Intel didn't let the HW review sites do their own benchmarks. Intel's marketing arm is very, very strong and they have really piled the hype surrounding Conroe's launch. This has the effect of disrupting the CPU market before a new product launch, even when the product turns out to be a flop instead of a splash (i.e. the Prescott) like the hype led everybody to think. And there is nothing to kill hype like a couple of benchmarks that show your product below the competition, especially if the tests are in an area (multitasking) that your marketing people hype up as direly important.

The Intel Conroe may or may not be a great chip (I am betting it will be at least good based on the architectural details we know about) but I am sure that it won't be radically better in all applications than the K8 line as the marketing weenies have led us all to believe.
a c 99 à CPUs
June 6, 2006 7:37:16 PM

And if you remember, the early AM2s performed like crap as the DDR2 IMC had problems and there were doomsday predictions about how AM2 would perform far worse than Socket 939. That and AMD tested it with high-latency DDR2-667 instead of the lower-latency DDR2-800 that most people will be feeding to the AM2s.
June 6, 2006 7:50:23 PM

If this is the best complaint there is against Conroe, then I'm ashamed of the AMD fanboys on this one. There should be a more creative response than this!

There are so many errors in this methodology, it's not even funny. :oops: 

1) You are comparing benchmarks from two seperate sources. This is never a good idea.
2) The E6300 that was previewed was an A1 revision. The retail versions will be a B0 or higher - and B0 has a noticable gain over the A1 revisions according to XtremeSystem benchers.
3) The E6300 costs $183 in 1000 quantity. I can't find the 1000 quantity for X2 3800's, but considering most outlets won't sell it for less than $300, I'd expect it to be $260ish at the very least.
4) I thought synthetic benchmarks were useless? Or is that only the case when your favorite processor maker is losing in them? :roll:
June 6, 2006 7:53:45 PM

Good point. It's like comparing apples to oranges(k,oranges with an apple like taste). You need :
2 systems that are similar appart from mobo and CPU
1 test used for both
the same environment
Even in those conditions the results probably won't even reach 95% accuracy.

Also those are sintetics. They don't count anyway cause if they did Intel wouldn't have had to release Conroe since the Pentium D is a good sintetics performer.

And no need to prove yourself Itguy. You're obviously MMM waiting for every opportunity to somehow bash Intel.

" lower-latency DDR2-800 that most people will be feeding to the AM2s."
No bud. Every day people will probably not spend that amount of cash on memory. I can assure you that most people worldwide that will buy AM2 will go for 533 or 667 with high latencies. Memory isn't considered important. Most people would rather go for a better CPU or gfx card.
June 6, 2006 7:55:00 PM

Quote:
If this is the best complaint there is against Conroe, then I'm ashamed of the AMD fanboys on this one. There should be a more creative response than this!

There are so many errors in this methodology, it's not even funny. :oops: 

1) You are comparing benchmarks from two seperate sources. This is never a good idea.
2) The E6300 that was previewed was an A1 revision. The retail versions will be a B0 or higher - and B0 has a noticable gain over the A1 revisions according to XtremeSystem benchers.
3) The E6300 costs $183 in 1000 quantity. I can't find the 1000 quantity for X2 3800's, but considering most outlets won't sell it for less than $300, I'd expect it to be $260ish at the very least.
4) I thought synthetic benchmarks were useless? Or is that only the case when your favorite processor maker is losing in them? :roll:

Yup, not only are they from two separate sources, but they aren't even the same benchmark! Itguy, you have two options, did you:

a) not pay attention and quickly flail about and post this as soon as you could fabricate a story to further your agenda?

b) know that you were purposefully deceiving folks by posting two different benchmarks from two different sites, and thought that no one would call you on it?

Lame. Grasping at straws.
June 6, 2006 8:01:04 PM

Quote:
If this is the best complaint there is against Conroe, then I'm ashamed of the AMD fanboys on this one. There should be a more creative response than this!

There are so many errors in this methodology, it's not even funny. :oops: 

1) You are comparing benchmarks from two seperate sources. This is never a good idea.
2) The E6300 that was previewed was an A1 revision. The retail versions will be a B0 or higher - and B0 has a noticable gain over the A1 revisions according to XtremeSystem benchers.
3) The E6300 costs $183 in 1000 quantity. I can't find the 1000 quantity for X2 3800's, but considering most outlets won't sell it for less than $300, I'd expect it to be $260ish at the very least.
4) I thought synthetic benchmarks were useless? Or is that only the case when your favorite processor maker is losing in them? :roll:


Ahh once again the devil is in the details. Props to you ethernalite for educating me on the error of this methodology. You would make a much better detective than I.
a c 99 à CPUs
June 6, 2006 8:06:54 PM

DDR2 has gotten a *lot* cheaper in the last few months, especially DDR2-800 and better. So I will be willing to bet that most people will put DDR2-800 at 5-5-5 or less in AM2 units. The enthusiasts are using 4-4-4 or better today, and it will be a while before enough AM2 units get out in the OEM and general-consumer market. By then, 4-4-4 or better DDR2-800 should be at $75-85/GB or so that DDR-400 at CL2.5 or 3 is. AMD demo'd the AM2 engineering samples with 5-5-5 DDR2-667.
a c 473 à CPUs
a c 115 å Intel
June 6, 2006 8:09:39 PM

I though this would be a thread that show Conroe going head-to-head against a X2 3800+ in the same test environment.

*** YAWN ***

Wake me up when there's a real comparison between the two from a reputable site.
June 6, 2006 8:12:20 PM

Just another attempt by the fanboys to debunk the Conroe.

it's getting sad and pathetic if you ask me. They get all excited when they see stuff like this. Sad, very sad.
June 6, 2006 8:26:16 PM

Nice post ITGuy.

I suspected this a long time ago but now I'm totally convinced that conroe will suck at multitasking thanks to cache trashing.

In conclusion, Conroe lead is really shrinking... even before it's released.
June 6, 2006 8:29:09 PM

I bet that if AMD had introduced the shared cache first, it would be the greatest innvoation in the history of CPU's to 9-inch. But since it was introduced by Intel, it suffers from terrible cache thrashing! :roll:
June 6, 2006 8:30:34 PM

I'll need to see that to believe it. It has got cheaper but not to that extent.
And the enthusiast market is about 1% of the market. Probably 10% out of all the average people who own computers are enthusiasts. So even if all enthusiast do actually go for DDR2 800 4-4-4 goodness, the number is still small. Anyway, obviously, for the best performance you would need tight timings rather than bandwidth. So i expect the best performing memory would be a 667 on steroids(for dual core at least). Yes 667 5-5-5 is pure exaggeration but let's not now expect everyone to go crazy for fast memory because they won't. DDR2 800 5-5-5 is a bit expensive and pointless for AM2. DDR2 800 4-4-4 still left much to be desired for at that price.
DDR2 prices should come down though. But slowly... There are still way too many people using DDR and upgrading their memory,and many are building AMD 939 systems right now using DDR. The transition will take time.
June 6, 2006 8:33:12 PM

Quote:
If this is the best complaint there is against Conroe, then I'm ashamed of the AMD fanboys on this one. There should be a more creative response than this!

There are so many errors in this methodology, it's not even funny. :oops: 

1) You are comparing benchmarks from two seperate sources. This is never a good idea.
2) The E6300 that was previewed was an A1 revision. The retail versions will be a B0 or higher - and B0 has a noticable gain over the A1 revisions according to XtremeSystem benchers.
3) The E6300 costs $183 in 1000 quantity. I can't find the 1000 quantity for X2 3800's, but considering most outlets won't sell it for less than $300, I'd expect it to be $260ish at the very least.
4) I thought synthetic benchmarks were useless? Or is that only the case when your favorite processor maker is losing in them? :roll:


I didn't expected this to come from you. I was rather expecting someone mentally challenged like TopotaMadre to come up with such a lame excuse.

It doesn't matters if the tests were run on different siteds since the results will be the same d@mn thing. Even an error margin of 5% in favor of conrunt it would still lose against the X2 (make the math your self). :wink:
June 6, 2006 8:34:50 PM

Very weird that Conroe would lose to X2 3800. I wonder what we'll all see in the months to come. Is Conroe just hype after all and on a new board?
June 6, 2006 8:35:18 PM

Quote:
I bet that if AMD had introduced the shared cache first, it would be the greatest innvoation in the history of CPU's to 9-inch. But since it was introduced by Intel, it suffers from terrible cache thrashing! :roll:

I'm not totally convinced about shared cache, specially the way how intel has it implemented on their new line of processors. 8)
June 6, 2006 8:42:08 PM

Quote:

The chip.
http://detail.zol.com.cn/picture_index_51/index500822.s...


I suspect that when these chips are finally available, that Conroe will do awesome on the synthetic and single thread benchmarks that use less than 4MB of cache, but in real world multitasking and multithreaded benchmarks, both cores of the Conroe are fighting to use the same cache and performance then plummets.
This is the 2MB L2 Conroe, not the 4MB.
June 6, 2006 8:49:22 PM

Quote:
If this is the best complaint there is against Conroe, then I'm ashamed of the AMD fanboys on this one. There should be a more creative response than this!
There are so many errors in this methodology, it's not even funny. :oops: 

This is no worse than the methodologies that Intel has crammed down the public's throat comparing 2 systems built and configured by Intel where everything is different about the systems; you can't even open them up; and only select benchmarks that Intel chooses are tested instead of all benchmarks. Another claimed Intel favorable review tested systems on other sides of the world, again with Intel employees setting up the systems. Heck, Intel even flat out lied on their www.intelstartyourengines.com page where they had a pretty chart which said a Woodcrest could beat an Opteron running with both systems running sql server 2005 64bit edition. The problem was the hardware configs were completely and substantially different and the Opteron system was actually running sql server 2000 32bit edition (Intel claimed it was running 2005 64bit edition too) while the Woodcrest ran sql server 2005 64 edition. They changed it a few weeks later once people discovered that they were flat out lying. Yes, Intel's marketing arm is big and it has spikes on it.

What people need to understand is that benchmarks are like Statistics Class. You can take the exact same hardware and make a positive case for each one and even back it up with benchmarks to say that one beats the other. All you need is unscrupulous morals or the incentive of BILLIONS of dollars. Intel stands to gain hundreds of millions of dollars in pre-sales alone as long as the public perception of Conroe's performance stays the same, so they have a tremendous incentive to control the tests and hide potential weak areas of their new processor line.

This is why no benchmarks should be admissible unless they are fairly and properly designed by an independent 3rd party with zero control by the chip manufacturers. Even then, you need multiple full independent benchmarks to get a clear indication of where the 2 products stand. That's also why I am advocating that Intel release the samples of their new chips so benchmarks can be run and we can see where Conroe is great and where it may not be. Right now any supposed benchmarks that Intel has pushed out are no more believable than a political ad days before the election. FWIW, I would be just as skeptical if AMD did the same shenanigans for it's benchmarks. But they send out their chips for independent review and don't "pre-configure" the hardware and restrict the benchmarks. Again, there is a reason Intel is not releasing samples and that is to make more money in pre-orders before the "true" benchmarks are released that show Conroe's true performance is not what Intel has led everyone to believe.
I am not knocking Conroe because I think it is a great chip and will be a great value, but I am dissapointed in Intel's chosen method of advertising their products which promotes deception by controlling reviews and sometimes borderlines on outright lies.

1Tanker wrote:
" This is the 2MB L2 Conroe, not the 4MB."
Thanks for that addition. I concur.
June 6, 2006 8:56:07 PM

Oh this is tragic.
June 6, 2006 9:04:29 PM

I agree that Intel's numbers must be taken with a grain of salt. And I am not at all saying that these numbers are not the least bit interesting. I'm sure Intel has displayed their processors in the best light possible. And I'm sure that the FX-62 will beat, or tie, the Conroe in a couple of benches that Intel has not allowed previewers to run. There is, however, plenty of third party data from people that got their hands on the ES samples to prove that Intel's numbers are in the proper ball park. Attempting to deny it simply proves that you are mindlessly supporting a multi-billion dollar corporation for no particular reason.

For you to criticize Intel's marketing/preview methods and then go and cite two seperate benches of an old revision on a cheaper chip on a single benchmark (that was probably cited as "useless" since it was synthetic) and then drawing sweeping conclusions ("Conroe sucks at multithreading") is so wrong that it makes Intel look like an honest little boy. That's all I'm saying.
June 6, 2006 9:12:02 PM

Quote:
...

Actually there are some Core 2 Duo samples outhere...


Actually theres not, not in unbiased hands; Intel claimed to officaly release samples but the just had reviewers come to them and work on their set-up boxes.
June 6, 2006 9:32:40 PM

Now that Intel is finally putting the pressure on AMD in the performance arena after 3 years, we will all win. Let the performance war and price war begin.
:lol:  :p 
Now that Conroe is looking good the Intel flame boys can finally beat their little chests. Let them have their day. They have been beaten up for years by over priced proc's and poor performance.
AMD proc's have been out performing Intel proc's and running 500mhz to 1 ghz slower at the same time for years now. The tables look to be finally turning in that area.

Think back a couple of years to when AMD had cheaper proc's, better performance, and the proc's ran slower and still beat Intel, up until now(
that is IF Conroe is all that it says it is).
Remember the AMD camps arguements? We are cheaper, perform better and give up 500 hgz to 1 ghz and still beat Intel.
IF Conroe is all that, then the Intel flame boys can throw that back at the AMD camp.
In the end, we all win.

:D 
June 6, 2006 9:32:41 PM

Quote:
Given that even today Intel still will not release samples to the community (Yes, Tom's recent article was on a machine provided by Intel where again, they were allowed only to push buttons), you know something has to be hiding inside all these Conroe benchmarks. It seems that a Chinese site got hold of a Conroe E6300 without Intel around and ran his own multitasking tests. And low and behold, Conroe loses to the X2 3800.

In SysMark 2004 test, the Con E6300 got 210 points. In comparison, an Athlon 64 X2 3800+ got 232 points. That's a 10.5% AMD win using Intel's own benchmark that Intel developed.

The chip.
http://detail.zol.com.cn/picture_index_51/index500822.s...
Conroes Result.
http://mb.zol.com.cn/29/295247.html
X2 3800's Result.
http://reviews-zdnet.com.com/Polywell_MiniBox2_939NP/45...

Now of course this is still only preliminary, but I suspect that when these chips are finally available, that Conroe will do awesome on the synthetic and single thread benchmarks that use less than 4MB of cache, but in real world multitasking and multithreaded benchmarks, both cores of the Conroe are fighting to use the same cache and performance then plummets. If this holds true, then it would make for very interesting discussions because most games are still single threaded, but more games are becoming multithreaded. Also, most everyone likes the idea of maximizing their multitasking performance and if Conroe's multitasking performance is weak it poses a big negative, even to some light gamers who enjoy multitasking at the same time.

Ask yourself this.
Why won't Intel just release the damn samples already!? Is it because they know there are holes in its performance and it won't be as good as the marketing hype that they've generated so they are hiding it for as long as possible until actual launch? I say they have it available so ship out the samples!!


Why should they? They have everyone guessing and talking about their Conroe chips as we speak. It far outshines any announcements right now from the competition. In my opinion there is no better way to sell your product than by getting everyone’s curiosity and anticipation up. I have seen many tests performed and many performance numbers released by many sources. Some by sites I trust and some that I’m not too sure about. However, I personally believe that once released the Conroe processors will be something special at least this go around. It seems that there are so many people that want this new chip to fail in some way. I for one want this processor to take the lead and dominate. This in turn will force change in the competition. Only when this happens will we as consumers see leaps in performance and possible lower prices. So until someone has solid numbers one way or another I consider posts like this simple spam. We all need to wait to see the true numbers once released.
June 6, 2006 10:05:15 PM

Biased opinion but in my home one HP w8400 with 2 Dual-Core Intel® Xeon® Processor 5080 leave one HP xw9300 with 2 Dual-Core AMD Opteron™ Processor 275 far behind it.

The two workstaions cost me 6200$ and I test the two in order to buy another 9. If a Presler based workstation can do it the better Woodcrest can make the Opteron slow.

Same hardware vendor, same money, a lot difference. Or AMD cut prizes or I forgot it when I get workstations at the end of the mounth.
June 6, 2006 10:20:04 PM

Those site seemed a little bit dodgy to me. Its was all in Chinese and it’s the first time I've ever come across this mb.zol.com

I think this is a case of AMD fanboyism response to Intel Conroe
June 6, 2006 10:22:46 PM

Quote:
Biased opinion but in my home one HP w8400 with 2 Dual-Core Intel® Xeon® Processor 5080 leave one HP xw9300 with 2 Dual-Core AMD Opteron™ Processor 275 far behind it.

The two workstaions cost me 6200$ and I test the two in order to buy another 9. If a Presler based workstation can do it the better Woodcrest can make the Opteron slow.

Same hardware vendor, same money, a lot difference. Or AMD cut prizes or I forgot it when I get workstations at the end of the mounth.
8O :?
June 6, 2006 10:26:00 PM

Quote:
Biased opinion but in my home one HP w8400 with 2 Dual-Core Intel® Xeon® Processor 5080 leave one HP xw9300 with 2 Dual-Core AMD Opteron™ Processor 275 far behind it.

The two workstaions cost me 6200$ and I test the two in order to buy another 9. If a Presler based workstation can do it the better Woodcrest can make the Opteron slow.

Same hardware vendor, same money, a lot difference. Or AMD cut prizes or I forgot it when I get workstations at the end of the mounth.


Really? Which applications does one beat the other in, or is it just the famous "snappiness"? I am wondering because even though the Xeon is almost double the clock speed, I have not seen many reviews showing the Xeon come that close in most server apps. Plus, I was not able to find an xw9300 with 2x 275's in it.
June 6, 2006 10:32:15 PM

I've seen so many PI benchmarks around, why don't they do some Multithreading tests, i want to see real performance not one cpu with 4mb of cache available for it.
I play games 15 percent of my time on a PC, the rest i work, and i use 5-8 big software packages running on my desktop, with outlook open, my system is crawling like a snail in heat.
June 6, 2006 10:34:15 PM

Quote:
Biased opinion but in my home one HP w8400 with 2 Dual-Core Intel® Xeon® Processor 5080 leave one HP xw9300 with 2 Dual-Core AMD Opteron™ Processor 275 far behind it.

The two workstaions cost me 6200$ and I test the two in order to buy another 9. If a Presler based workstation can do it the better Woodcrest can make the Opteron slow.

Same hardware vendor, same money, a lot difference. Or AMD cut prizes or I forgot it when I get workstations at the end of the mounth.


Really? Which applications does one beat the other in, or is it just the famous "snappiness"? I am wondering because even though the Xeon is almost double the clock speed, I have not seen many reviews showing the Xeon come that close in most server apps. Plus, I was not able to find an xw9300 with 2x 275's in it. I think he is just spreading personal experience FUD, much like 9-Inch does with everything else.
June 6, 2006 10:36:35 PM

Thats one of the reason's I called him out. I want to see what apps he is talking about, and if they exist/qualify as useful.
June 6, 2006 10:41:53 PM

He is full of crap. An Opteron server would cost much more than a Xeon server. And if they don't, they should because Opteron's kick Xeon's in the balls.
June 6, 2006 10:44:22 PM

Well, I looked it up, an equal server with dual 270's costs about half as much (that would probably sell me right there, I'd buy two, and set them up in a grid), and with 280s is a couple hundred more.
June 6, 2006 10:53:34 PM

Quote:
Well, I looked it up, an equal server with dual 270's costs about half as much (that would probably sell me right there, I'd buy two, and set them up in a grid), and with 280s is a couple hundred more.
Your kidding me, right? I thought Opterons were always more expensive. Though to tell the truth, I never have actually looked at or for a Xeon. 'Bit ironic considering I have an Opteron in my high-end system. Just out of curiosity, link please? :?:
!