Monday, June 05, 2006

Conroe close to be really busted

Remember I told you that Conroe is only good for those single threaded applications with working sets comparable to 4MB, so they can run from cache? Now, there is more proof. A Chinese site got hold of one of this Con E6300 chip (1.86 GHZ) and benchmarked it with BAPCo SysMark 2004.

BAPCo SysMark was created by Intel. In July 2002, AMD was allowed to join in its development. Today, BAPCo's members include INTEL, AMD, Microsoft, HP, DELL, CNet, ZDNet, Nvidia, ATI and others. SysMark is about doing a series of real activities, mimicking real computer usage. These tests have large memory requirement, as they chew on large documents, etc. Also, multiple applications may be running at the same time during a SysMark test. Intel had avoided using this benchmark in all the Intel controlled benchmark runs.

In SysMark 2004 test, the Con E6300 got 210 points. In comparison , an Athlon 64 X2 3800+ got 232 points (2/1.86 = 1.07, 232/210=1.1). In Business Winstone 2004, the Con E6300 got a score of 25.2, an Athlon 64 X2 3800+ got 26.2. In Content Creation Winstone 2004, the E6300 got 32.2, the Athlon X2 3800+ got 33.2 (note that WinStone is not a multi-tasking load, it just runs through more memory). Recall that the X2 3800+ has only 512KB L2 cache per core. Suddenly, the Conroe doesn't look so rosy at all. Intel priced this chip lower than X2 3800+ for a good reason.

No wonder Intel so far doesn't allow reviewers do their own benchmarks. The most recent tomshardware.com comparison of Con E6700 and FX62 was again done with Intel configured hardware and benchmark suites. When a lot of amateurs and script kiddies cheered Intel's IDF guerilla benchmarketing, I smelled something fishy from the very beginning. Later I pin pointed the origin of Conroe's scores.

Why is this so important? Intel's Israeli team has bragged about the 4 issue core and various minor hacks. The amateurs cheered about a 33% IPC improvement. They say 4/3 = 133%. But anyone with a good understanding of CPUs know this is impossible. The Conroe has 14 stages, 4 issue, this means 56 instructions in flight at any moment, the amount of dependency is inherent in normal code. AMD has long concluded that 3 issue is efficient and cost effective.

So, again, just as we have seen from Woodcrest benchmarks, from SysMark results, we see CORE2's IPC is not any higher than K8. CORE2 marchitecture is no miracle. Clockspeed and bandwith will determine Conroe and K8's relative performance.

I expect that once Conroe is released in the wild, it will be soundly fragged by AMD's offerings in most multi-tasking loads. Remember, the whole purpose of going multi-core is for multitasking/multithreading.

Of course, the AMD 4x4 technology will permanently pin Intel at half of AMD's performance.

PS: I am glad that the Chinese have started running benchmarks. What da heck. It's just installing and running programs. Any kid can do it.

PPS: Let's enjoy some music.

PPS: FX62 benchmark results can be found here.

41 Comments:

Anonymous Anonymous said...

hey sharikou. I think the 1.8 GHZ conroe is just a 2MB cache. DOes it make any difference with 4MB?

10:15 PM, June 05, 2006  
Anonymous Anonymous said...

I agree with your views. I think it's very important for more people to understand such views.

However, what matters to the market sales is not how Conroe would really perform, but how it would appear to perform. IMO, the 4-issue and large L2 cache are Conroe's strength when it comes to convincing those benchmarking sites. (And who do you think people will listen to when they decide which computer to buy?)

No doubt Intel is doing benchmarketing, and Conroe was designed to play well on benchmarks and probably no others. All those Conroe superior webpages repeat only a limited set of programs, and that means little if any, and smells wrong.

But sadly, with Intel's marketing, they are able to give Conroe the gaming performance crown! You can't say Intel's lying when it says "Conroe's performance on xyz benchmarks is better," because it is, even though those benchmarks probably reflect none of real-life applications.

IMO, this "benchmark hype" is just p4's megahurtz hype resurrected. Intel finally becomes a marketing company.

10:25 PM, June 05, 2006  
Blogger Sharikou, Ph. D. said...

Intel finally becomes a marketing company.
I expect AMD 4x4 will make Intel look really silly.

10:28 PM, June 05, 2006  
Anonymous Anonymous said...

I think that Conroe is a great chip. Because of cache. Because of performance per watt. They did a chip which matches AMD platform. But nothing more than that. Because K8 architecture is even greater chip.
Now it is urgent to match marketing efforts and manage this "Conroe issue".

10:36 PM, June 05, 2006  
Anonymous Anonymous said...

You are comparing a $300 AMD to a $183 Intel chip which runs at 2Ghz v 1.86.

Really you either need to compare similar speeds or similar price points depending on whether you want to do a comparison for your dollar or for the comparative chip.

Need to see more back to back tests with games and file conversions etc.

10:38 PM, June 05, 2006  
Blogger Sharikou, Ph. D. said...

I think that Conroe is a great chip.
Getting close to K8 IPC level is no small feat. I should commend the Israelis too. But the problem is they hyped about 20% IPC advantage from the beginning, and now they must try every kind of cheats possible to make that promise seem fulfilled.

10:39 PM, June 05, 2006  
Blogger Sharikou, Ph. D. said...

You are comparing a $300 AMD to a $183 Intel chip which runs at 2Ghz v 1.86.


2GHZ is 7% faster than 1.86 GHZ. E6300 is priced at $209. I expect X2 3800+ price to drop from current levels in two months, as AMD introduce higher speed grades.

10:43 PM, June 05, 2006  
Anonymous Anonymous said...

These 2MB Conroes are of course priced lower than the X2 3800+ because they are considered low end. Get a grip.. Often prices plays some role in customer's purchasing decision. From what I heard all Conroes are gonna priced lower than their competitive equivalents.. I mean MUCH LOWERRR. Brace for price wars ahead, methinks...

10:44 PM, June 05, 2006  
Anonymous Anonymous said...

Everything depends on Conroe availability till end of the year.
Everything. If yelds are low, Intel is doomed because people will be trying to get a chip which is unavailable.
So you wait or buy AMD chip...

11:26 PM, June 05, 2006  
Blogger Sharikou, Ph. D. said...

Everything depends on Conroe availability till end of the year.

Yes. Conroe may be the Intel killer. If Intel can't deliver Conroe at 50% volume by the end of the year, it has substantial probability of going under.

11:30 PM, June 05, 2006  
Anonymous Anonymous said...

Maybe Intel CAN sell Core 2 Duo's at lower prices because it's doing well on the 65nm fabrication.

After all, a 65nm chip is about half the size (and thus price) of a 90nm one with the same number of transistors.

12:43 AM, June 06, 2006  
Anonymous Anonymous said...

If Intel can't deliver Conroe at 50% volume by the end of the year, it has substantial probability of going under.

That isn't going to happen anytime soon. AMD can't supply the whole x86 market by themselves; if Intel went under in the timescale you've suggested before, the effects on the PC market would be utterly catastrophic.

If Intel is in danger of going under in the timescale you're suggesting, odds are that IBM will wait until it can buy Intel for cheap, then do so. As a result, the market situation would be the same as ever - in fact, it might be worse for AMD, as IBM's additional manufacturing facilities and design expertise combined with what Intel already has could make them a far more deadly foe than they're facing now.

2:03 AM, June 06, 2006  
Anonymous Anonymous said...

Now Shariku, the music link of Li Yundi is one of your beter posted links - thanks!

5:34 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

To those who quote Aand: we have proven Anand a paid Intel cheerleader, his opinion has zero credibility.

8:18 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

That isn't going to happen anytime soon. AMD can't supply the whole x86 market by themselves

Right now, AMD is desperately in need of capacity. Intel has been doing massive advertisement for AMD for the past 3 months. Intel made a lot of people know that P4/Xeon is crap and Core Duo is 32 bit and the future lies with CONROE. The crash of INTEL stock made these facts even more well known. The coming months during which Conroe ramps is the time of Gold for AMD. You have to remember AMD's ASP is below $100. If AMD succeed in denying Intel the oppurtunity of dumping its legacy chips, Intel will be stuck in the red for years to come. By Dec 2006, AMD will have 65nm parts...

8:29 AM, June 06, 2006  
Anonymous Anonymous said...

Sharikou I don't expect the 4x4 platform to be as cheap as you say it will be.

Initially I believe it will be a premium product for AMD and eventually scale down prices. Remeber that 'common joe' won't need this thing.

But my question is this:

Can I buy a dual-core processor now and buy a quad-core later on for the second socket. Or does the configuration have to be uniform (two 2-core or two 4-core)?

9:08 AM, June 06, 2006  
Blogger "Mad Mod" Mike said...

@Shari: If they tested against the Opteron 165 or X2 4000+, than we could see more cache to cache comparison

@Jeach!: AMD said themselves that typical Opteron 2P systems were too expensive for enthusiasts and that's why they made 4x4, so logically, it should be cheaper.

You can put a Dual-Core and a Single-Core in a Dual Socket 940 board, but a few applications error out and it is mostly unstable. I would surmise the same for any platform.

9:14 AM, June 06, 2006  
Anonymous Anonymous said...

E6300 = 1.86GHz / 1066MHz 2MB L2 Cache (2006-07-23 @ USD $183)

The BAPCO SYSmark 2004 SE Score you found for an AMD Athlon X2 3800+ is the highest I've seen for that processor. Anand Lal Shimpi recent SYSmark 2004 result for this processor is 212.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2736&p=5

(scroll down to see the overall score of 212)

Others have the Athlon X2 3800+ scoring 220 in BAPCO SYSmark 2004 (overall)

http://www.theregister.co.uk/2005/08/12/review_amd_x2_3800plus/page2.html

Anand also has an older 2.0 GHz Core Duo (Yonah) scoring 211, against an Athlon XP 4200+ 2.2GHz. scoring 222 here:

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2648&p=7

(again, scroll down for the relevant chart)

You wrote:

"BAPCo SysMark was created by Intel. SysMark is about doing a bunch of activities at the same time, mimicking real computer usage. In this test, the each core of Conroe will get less cache. Intel had avoided using this benchmark in all the Intel controlled benchmark runs."

Is it a stretch to think that a benchmark with "2004" in the name, a benchmark, in your words, "...created by Intel." might be "optimized" to flatter the "NetBurst" architecture Intel was marketing at the time? If, as you say, "Intel had avoided using this benchmark in all the Intel controlled benchmark runs.", mightn't Intel be aware of this particular benchmark's "optimization" to flatter "NetBurst" might not yield similarly flattering results for Conroe?

I'm no Intel fanboy, and I agree that the bus bottleneck is a serious limitation of Intel's current and forthcoming architectures, but if you wish to be taken seriously as a truthful observer of the industry, wouldn't it be prudent to avoid comparing the results of a chap in China with a different set of results at CNet/ZDnet?

Moreover, if you would ordinarily have discounted the BAPCO SYSmark 2004 benchmarks as suspect or flawed during the "NetBurst" era, why trumpet them as dispositive now?

9:23 AM, June 06, 2006  
Anonymous Anonymous said...

I went to the Taipei Computex today, and took a look at the Conroe with my own eyes. There are some weird things I found.

1. Conroe runs at a pretty high temperature. Since I didn't have a camera, I couldn't really take a picture. Conroe actually idles around 40C, while 60C under load. The temperature was read by a program, set up by Intel representatives.

2. Intel again put a supposedly "FX-62" aside with a Conroe system. However, when I asked if I could run any benchmarks, they refused. They also refused to run benchmarks themselves. The only thing they allowed me to do was to run CPU-Z. *So intel set up these two machines... for CPU-Z, and brag about their "75%" advantage over AMD?*<--- yep... 75%, claimed by a Intel fellow.

3. Conroe actually can run without a CPU fan!!. Sorry I don't have the photo with me, but Conroe could actually run *idle of course* without CPU fan, but with heatsink, for a significant amount of time.

Now, some words from Intel representatives.

1. "we currently don't see the need for IMC, because 4MB cache actually reduces the need to read from memory"

2. "we believe we can achieve higher scalability using FSB. As proven on Conroe, we beat AMD by a huge, undisputed margin"

3. when asked about dual-die approach for Clovertown, "i understand your question, but i believe you are actually comparing a platform with a processor technology"

4. "we are building warehouses, to store Conroes, and other NGMA chips. We have high yield from these 300mm wafers."

5. "IMC has its cons. one being that you can't have large amount of cache if IMC is present. Since Conroe doesn't have IMC, we can put a significant amount of cache on the processor."

I had some interesting discussion with the representatives there. I'll probably go tomorrow, and hopefully bring a camera. XD

9:33 AM, June 06, 2006  
Anonymous Anonymous said...

You wrote:

"To those who quote Aand [sic]: we have proven Anand a paid Intel cheerleader, his opinion has zero credibility.

Anand Lal Shimp wrote:

"As far as the processor at hand is concerned, Intel has done a reasonable job with the Pentium EE 955, but with Conroe not too far away, we just can't justify recommending it. If you absolutely must upgrade today, the Athlon 64 X2 is still probably going to be a better bang for your buck. However, as we have seen in the benchmarks, there are advantages to being able to execute four threads simultaneously."

"It is pretty much a toss-up at this point, but we'd recommend sticking with AMD for now and re-evaluating Intel's offerings when Conroe arrives. If all goes well, we will have a cooler running, faster processor with Conroe that may provide some even tougher competition for AMD's Athlon 64 X2."

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2658&p=14

If, as you say "we have proven Anand a paid Intel cheerleader,..." his paymaster doesn't seem to have been getting value for money. Perhaps Anand's nuanced verdict is insufficiently prejudicial to Intel for your taste.

9:37 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

if you wish to be taken seriously as a truthful observer of the industry, wouldn't it be prudent to avoid comparing the results of a chap in China with a different set of results at CNet/ZDnet?

Come on, you can check the configurations on the Chinese site. Any kid can do a benchmark. I would normally use an English source due to people's prejudice, but Intel doesn't give the western people the liberty to run SysMark and Winstone on Conroe.

Moreover, if you would ordinarily have discounted the BAPCO SYSmark 2004 benchmarks as suspect or flawed during the "NetBurst" era, why trumpet them as dispositive now?

SysMark was originally created by Intel to favour Netburst, this is a fact everyone knows. Van Smith had a paper on that. However, later AMD and others joined the SysMark development and it's becoming fairer. In any case, Intel created this benchmark and is still the most influential member of this benchmark. Why is it not using it for Conroe?

9:40 AM, June 06, 2006  
Anonymous Anonymous said...

sharikou... are you a chinese? :P

9:42 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

If, as you say "we have proven Anand a paid Intel cheerleader,..." his paymaster doesn't seem to have been getting value for money.

While Anand is a paid pumper, he must also maintain an impression of fairness from time to time, or he loses readership. What he said about X2 and P4 965 is obvious, everyone is saying the same. He is not so stupid as to recommend a worse chip.

9:58 AM, June 06, 2006  
Blogger Ajay S. said...

Hmm, I dont understand this.

How could a chinese guy get to do a independent conroe review while jurnos in US and Europe are still being fed only Intel configured setups?

I know revenue from chinese market is growing fast but since exactly when has it overtaken the west? Will pissed off bigshot hw review sites be calling intel up now?

Maybe Intel is just channeling more energy to the west to keeping its biggest markets brain washed!

Whatever, I do hope AMD 65nm parts roll out sonner than later

10:02 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

How could a chinese guy get to do a independent conroe review while jurnos in US and Europe are still being fed only Intel configured setups?
Well, the truth is probably this. In the west, Intel can sue anyone violating its NDAs and "trade secrets" into the stone age. I assume some wetsern jurnos managed to get a chip in their hands by now. But western jurnos are afraid. In China, if you bring suits like this, Courts would probably laugh at you.

10:07 AM, June 06, 2006  
Anonymous Anonymous said...

You know a good analogy for Conroe's 4MB cache dependency would be 0-60MPH times for cars.

If you have 5MT vs. 6MT, usually the 5MTs will win b/c there is one less shift required. However, the 6MT car may have significantly better times for 0-70, 0-80, and heck quarter-mile times as well.

But, if you restricted your view to only 0-60, the 5MT car will always look better.

10:13 AM, June 06, 2006  
Blogger "Mad Mod" Mike said...

If Intel reps said they don't need IMC because of high cache and high cache reduces need for it, how come their Quad-Core Itanium 2 has IMC and CSI and also 24MB of L2 Cache (6MB per core)? I'd say it's simply because the Itanium 2 is the size of a graham cracker and they don't care what they do with those PoS WoM's.

That scalability is hillarious. Look at benchmarks of the Bensley Dempsey, even with 2 1066MHz FSB's and QUAD CHANNEL RAM, it can't even get 6GB/s Memory Bandwidth...so I have to wonder, does memory truly get aggregated across the 2 FSB's? Or is Intel's NorthBrige just THAT crappy?

10:14 AM, June 06, 2006  
Anonymous Anonymous said...

More recently (April 07, 2006) Anand wrote:

"Gamers looking for a temporary upgrade should honestly look to the Athlon 64 3000+ instead, as very few games have boarded the dual train as of now.

At the higher end of these value offerings, the Athlon 64 X2 3800+ continues to be a tremendous performer with a relatively low price tag. If you can afford the approximately $300 entry fee, the X2 3800+ will truly impress you across the board. Interestingly enough, we found that for the most part the Opteron 165 just isn't worth it compared to the Athlon 64 X2 3800+. Thanks to AMD's on-die memory controller, the higher clock speed of the 3800+ is more useful than the larger L2 cache of the Opteron 165. (Overclocking makes things a bit more interesting, naturally.)

So there you have it, if you are a multitasker or run multithreaded applications, and you want a great low cost solution, then the Pentium D 805 makes a wonderful stepping stone to a future AM2 or Conroe platform. If you're a gamer that doesn't care about multitasking while gaming, the Athlon 64 3000+ is still a strong value. And finally, if you've got the budget for it, the Athlon 64 X2 3800+ still can't be beat.
"

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2736&p=11

Vanishingly subtle bias for someone you describe as "a paid pumper". Heck you don't seem to lose readers and your attempt to "...maintain an impression of fairness from time to time..." is confined to grudging backhanded compliments like "Getting close to K8 IPC level is no small feat. I should commend the Israelis too. But the problem is they hyped about 20% IPC advantage from the beginning, and now they must try every kind of cheats possible to make that promise seem fulfilled"

Why is the geographical location and the nationality of some of the "Core" architecture team so peculiarly noteworthy? You don't find it necessary to label AMD's personnel as Texans, or Americans, or the Dresden personnel as Germans? It's a sincere question, I don't presume to know or imply its answer.

I expect many readers return to your blog, as I do, despite you wearing your heart on your sleeve, because you serve as a useful aggregator of links to AMD News. Some, no doubt, come for your hyperbolic analyses and colourful language ("frag" et cetera). The danger here is in becoming a cartoonish buffoon, like a Rush Limbaugh, or a Howard Stern: someone whose audience giggles when you write something outrageous that they privately agree with but don't have an audience to say it to. You'd be more effective if you didn't editorialize and just reported the facts without the outrageous commentary. The facts against Intel, Dell, and some of the Wall Street firms and analysts you've mentioned are damning enough. Your readers (the ones worth having at any rate) are smart enough to connect the dots. When you go for the laugh lines you undercut your own hard-earned authority, Dr. Sharikou, and occasionally come across as an embittered clown. You're smarter than that.

10:52 AM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

Some, no doubt, come for your hyperbolic analyses and colourful language
Damn, I tried to respond twice, twice this blog software crapped out, and I lost the msg.
1) Anand got himself burnt for sure. Crowning conroe before even touching an eng sample of it is self-destruction of credibility.

2) This blog is about sharing and exchanging info and hammering out the truth. I am not trying to establish any authority here. Though humans have the tendency to follow authority, we are not going to exploit that weakness.

3) I believe a blog like this can provide value. Technology will level the playing field against mass media.
4) Let's don't take those "Israeli" talk as anything political.

12:28 PM, June 06, 2006  
Anonymous Anonymous said...

If we assume the scores are directly proportional to clock speed then the scores of all the benchmarks can be obtained by the formula:
Score at 1.86 GHz= score at 2GHz*0.93.
N.B. => (2GHz*0.93=1.86GHz)
The scores would be:
SysMark 2004-x2 3800=215.76, e6300=210
Business Winstone 2004-x2 3800=24.37, e6300= 25.2
Content Creation Winstone 2004-x2 3800= 30.88, e6300= 32.2
Assuming the aforementioned criteria are correct the x2 3800 falls behind the e6300 in all the results shown by sharikou except SysMark 2004.
Numbers aside, the benches are old and the results are within the margin of error.

1:22 PM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

Business Winstone 2004-x2 3800=24.37, e6300= 25.2
Content Creation Winstone 2004-x2 3800= 30.88, e6300= 32.2

Winstone is not even a multi-tasking benchmark. It just needs more memory, and already Conroe can't hold lead. Try find a Opteron 165 or 170 benchmark. I bet the 1MB cache can make up the slight difference.

1:33 PM, June 06, 2006  
Anonymous Anonymous said...

Ummm... just one thing: if Conroe is so busted, why suddenly the need for TWO FX-62s on one board? Quad core is just around the corner, isn't it?

How come 4x4 is coming now, AFTER NetBurst and a particular E6600/E6700 vs FX-62 comparison on Hexus?

1:34 PM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

a particular E6600/E6700 vs FX-62 comparison on Hexus?

The kids at Hexus can't be trusted. Tomshardware's results on similar benchmarks didn't yield the 40% performance lead Hexus got.

I think all the evidence is quite conclusive. For Conroe to show lead, the following conditions must be met

1) application load must be single threaded so full amount of cache can be used by one core

2) application load must be cache sensitive


4x4 is a solution to put Intel permanently at about 50% of AMD's performance on desktop.

1:56 PM, June 06, 2006  
Anonymous Anonymous said...

yep must agree to the last statement of sharikou.4x4 will own the real games crown.

I have a dual 280 and it outperforms a oc'ed conroe @3800 by 10-15% in benches like 3dmark on cpu power.

and maybe the benches aren't 100% accurate it realy shows what is stated a long time. conroe is nice for benches, in real world stay away from it and keep your k8, it's not worth "upgrading".

same as curent core duo, bringing a nice core to the market that doesn't have 64bit support, intel what the hell are you thinking.. that you can screw the market for another 3 years because you have the name tag???

3:15 PM, June 06, 2006  
Anonymous Anonymous said...

I am by no means an Intel fanboy but I, like most consumers is more concerned about bang for the buck than absolute performance. That said, when Conroe launches and prices on the Pentium D falls to absurd lows I plan on buyig a Pentium D 940 for around $190 and oc'ing it to around Pentium D 960 levels or higher (virtually all Presler CPU's can reach 3.6 G's at close to or at stock voltages). Now that's bang for the buck!
Notice from reviews around the web that the Pentium D 940> is more than capable at performing most tasks at reasonable levels.

3:42 PM, June 06, 2006  
Anonymous Anonymous said...

hey All Sharikou,looks like your favorite Intel Paid pumper has got (finally) a Real Conroe....and the results are somewhat inline with the IDF preliminary benchmark....check the link...
http://anandtech.com/cpuchipsets/showdoc.aspx?i=2771

6:20 PM, June 06, 2006  
Blogger Sharikou, Ph. D. said...

Coverage on Anand.

7:45 PM, June 06, 2006  
Anonymous Anonymous said...

Dear anonymous ( of the last reply to our PhD guy)

I think you need to read more CLEARLY what they state there
again they were mere BUTTON PUSHERS.
check again the SECOND PARAGRAPH of the webpage you give
at :
http://anandtech.com/cpuchipsets/showdoc.aspx?i=2771

also check the UGLY MEMORY they used for their "personal test"
I wouldnt say "ohh shit its pure lie" if someone acussed anand of going "you guys got told by intel to use these horrible memory modules right?"

10:15 AM, June 07, 2006  
Anonymous Anonymous said...

Well. The reason why Chinese sites got ES CPUs faster is because almost all the motherboard manufactures are in Taiwan or China. Intel and AMD both have to send them newest samples to test compatibility on the motherboards. After testing, the CPUs got leaked. Why would Intel or AMD send ES to reviewers who mostly don’t know shit about electronic engineering? They need people with proper equipments and knowledge. A lot of Chinese sites are actually operated by daytime engineers actually working in the industry. Intel or AMD won’t send you one. Why, because you are not qualified.

12:45 AM, June 15, 2006  
Anonymous Anonymous said...

Here is a good site in Taiwan. They also have an English forum if you like to ask questions.

AM2 FX-62 VS Conroe E6700

http://www.coolaler.com/modules/news/article.php?storyid=470

12:51 AM, June 15, 2006  
Anonymous Anonymous said...

Threads like this are great to revisit months later. Now that Conroe is shipping in quantity and AMDs best parts are being killed in every real world benchmark by Conroe based solutions that are significantly cheaper, where is all the discussion of Conroe being "thread count bound", "cache bound", etc?

What happened to all of the genius expert predictions? I'd like to see some explanations of how the predictions that have bounced around here could have been so far off. Or maybe some excuses instead?

All of my machines are currently AMD based, but Conroe performance is undeniable. It's silly to pretend it isnt what it appears or make excuses (or even worse, ridiculous "architecture elegance" comparisons) when real world performance is demonstrating this cores superiority despite AMD having played all its cards (integrated MC, LDT, DDR2, etc)

7:49 PM, September 11, 2006  

Post a Comment

<< Home