Conroe is looking to be impressive, but the Pentium 805D's price looks even more impressive! :)

LOL! I have never heard the word fanboy in one topic as much ever! :)

Anand still is a paid intel bumber.
fx has lower scores than usual :D
PcStats almost only one which we can trust. Until they are paid some $$$$ too.
Conroe is impressive, but not that impressive folks!
Beware of marketing bull**** :D

Intel themselves said that they are only gonna be producing 35% conroe by the end of the year

#111 - This has no effect on current apps, except a few theoretical benchmarks at this point. Later, this may prove to deliver extra performance, when apps can take advantage of it. How many games actually use SSE? All the gaming performance advantage is because of the cache.

32-bit netburst xeons -> 64-bit netburst -> 32-bit p3 tech -> 64-bit p3 tech -> 64-bit netburst tech (tulsa). It it looks like it was designed by a bunch of dancing monkeys.

470W dual core paxville heaters, 32-bit yonah processors for blade servers, woodcrest server with one model number higher than netburst servers 50XX vs 51XX. They gotta use netburst with 16 megs of shared cache for 4p because woodcrest can't do 4p at this point. Looks like a complete mess to me. Intel's track record in the server space is crap, and people are starting to realize it. Many companies are tied into DELL, and wouldn't know a good processor if it bit them in the ass, so you may have a point

Message to AMD fans....you wont see K8L till middle of 2007 which is about when Intel will be releasing Conroe II. (thats if AMD actually has no probs with the new 65nm processes and the K8L architecture.
So if you are "waiting on K8"....I wouildnt hold your breathe :)

Lets clear a few things up.

1. The only reason Conroe is so fast is because of the cache. This is prov en: when two instances of a single threaded benchmark are run, the scores are significantly worse because the cache is shared between cores. This will make conroe really good for current games, which are still mostly single threaded. For multithreaded apps, conroe will not be much faster than AM2.

2. Even Intel stated that even by the end of the year conroe will only be 35% of the desktop chips intel produces. How many of these do you guys think will be 4 meg cache models? The yields have got to be bad at this point. The 2 meg models are not going to be overly impressive, but will probably represent the majority of conroes sold. My point: AMD will not go under because it can't match the top models of conroe at release. 4 meg cache desktop models of conroe will represent a very small fraction of CPU's sold by intel for a long time, which will keep prices of these models a lot higher (Think 7800 GTX 512) as the demand will be high. On most other chips intel doesn't make a lot of margins (think $90 on 805D). Between the netburst crap intel intel will be trying to sell everybody until well into 2007 and the few available conroe chips, amd will have plenty of sales on the desktop.

3. On the server side, people that use opteron now are not gonna switch to woodcrest right away. By the time people realized Woodcrest is there, the next big thing will be out. Also, intel's Server roadmap is a complete mess. 32-bit netburst xeons -> 64-bit netburst -> 32-bit p3 tech -> 64-bit p3 tech -> 64-bit netburst tech (tulsa). It it looks like it was designed by a bunch of dancing monkeys. People who buy servers look for upgradability in a platform.

Therefore it does not take Fanboys for AMD to live on

1. The only reason Conroe is so fast is because of the cache. This is prov en: when two instances of a single threaded benchmark are run, the scores are significantly worse because the cache is shared between cores. This will make conroe really good for current games, which are still mostly single threaded. For multithreaded apps, conroe will not be much faster than AM2.
Where has this been proven? Woodcrest has no problem beating Opteron in highly threaded server apps and there's no significant difference in Conroe's superiority in single-threaded and multi-threaded tests in PCMark 2004.

http://www.xtremesystems.org/forums/showthread.php?t=97609

As for the 2MB cache, a 1.86GHz Conroe with 2MB cache is pretty competitive:

http://www.pconline.com.cn/diy/evalue/evalue/cpu/0605/791941_8.html

Probably would be a good match for a 4600+.

If we didn't have fanboys and joe-buy-from-xxx then one company would go under when their performance was behind for a year or two.

We need those people.

Fanboys are tiresome. Intel, after being behind for years, pull ahead, and the only explanation is a conspiracy. Geez, get a life.

My X86 chronology has been Intel, AMD, AMD, Intel, Intel, AMD, AMD.

Almost certainly Intel this time. I simply buy the best bang for the buck at the time I get a new PC.

Anyone who blindly adheres to one brand is limiting themselves, to publically go into denial mode, is pathetic.

Intel gave him the hardware, or how in the hell he got those stuff?
And why memory with 5-5-5 latencies?
Maybe fx would look better with 4-4-4, dunno :p

Btw. Intel quoted Anandtechs and hexus' comments about conroe
and showed them in wall street. hmmm......
Intel cannot blackmail OEM's anymore (lawsuit)
So Intel uses it's enormous cash piles to bribe reviewes.

Intel is just desperately trying to get money from conroe before AMD's ownage 65nm's arrive.

4 issue wide probably does the least good, something like 5% increase in performance.

x86 code is known to be difficult to extract ILP from, anything past 2 pipes and diminishing returns kick in hard and fast.

Single cycle SSE execution will only improve performance in situations where SSE code is actually used, and outside of a few things like gaming or video encoding it really isn't used that much...

Cache helps, but its not a silver bullet either. Sort of like how high clockspeed or IPC is no gaurantee of high performance.

Most of Conroe's performance comes from its memory disambiguation hardware to keep those ALU's fed, Macro Op Fusion, and various tweaks to the x86 decoders.

AFAIK K8L will include AMD's take on alot of these things but its not coming out for a year at best so until then Intel will be the one to buy.

You know, I'm disappointed, especially being from a site like Anand's. Game benchmarks were great. Then they used WorldBench and RetardedMark for EVERYTHING ELSE. Damage... Tech Report... please, pleeaaasee do the quality benchmarks using REAL applications that you usually do...

Wow, I wonder what is the first order effect that is responsible for the score differential. It could be a bunch of tiny changes but for >20%, there must be some key feature?

I can't wait to see the numbers for the $316 Core 2 Duo E6600 compared to that of the $1031 Athlon FX-62. :-)

AMD better get that K8L out soon ;). Altough i want to see the test with optimal timings for both AMD and Intel to.

As Conroe is definitely Intels next gen CPU, AMD has got some time now to make sure its next gen CPU gets better, maybe wel now for sure how they perform against each other this fall.

BTW, which are the sockets respecitvely for Conroe and K8L?

Funny with the fear minimum framerate, I wonder what the cause might be.

But a 316$ E6600 for me thanks, and then to 3Ghz or above :P

But I don´t get running the X6800 with 800Mhz DDR2. Since the memory system is 12.8GB/s vs the FSBs limit at 8.5GB/s. Core 2 Duo don´t need 800Mhz like AM2, low latency 533 or 667 will do fine unless you add IGP.

I dont get it. A great improvement is made, and some of you seem angry or defensive.

Please tell me there's some of you out there that just plain buy the best given the price at the time?

My last chip was the pd 805 - no way was i getting a single core for a similar price. Prior to that was an athlon64 3400+. prior was an xp-M when it had fallen to about $70. prior was a northwood 2.4C.

Are there still some of you here that just dont give a sh!t who makes something as long as it's the best thing that suits your pocket?

Seeing the benches is far from a waste. Not all of us have endless pockets - it's one reason why we are here at techreport, or look at stuff like e3 or computex. We want to know what's coming and we plan accordingly. for something like conroe, That might include saving up and waiting for those few months of press releases. I fail to see how looking at a graph - seeing one bar longer than the other consistenly, then judging the chip it relates to as being faster "biased". Dont believe it? FINE! wait till it's out, but dont act as if you have to defend a company.

I sense dissapointment from Conroe when it comes out retail.

It will be faster, but just OK faster, nothing spectacular like 50% or what some previews were indicating (the worst was I think the comparison here at TR at server end) ... hope I am wrong though, as I was preparing to buy one when it comes now... hmmmm... let's wait and see the real thing in a month or two.

i too the older reviews showed much better results, the one with a amd fx 60 overclocked to 2.8 and the conroe at 2.6 or summin had a bigger increase in peformance than the reviews that are comming out today, some said coz it was coz of crossfire, but the amd fx 60@62 had crossfire to and still was much slower

but anyway, im looking to buy a new pc in a month anyway so do i

1. get conroe

2. get AM2 with a cheap sempron and upgrade to the K8l when it comes out, is K8l on AM2 by the way ?

so do i do option 1 or 2

i got a feeling that the K8l will just beat conroe like by %100 in everybenchmark when it comes out and i will be like "why didnt i get that instead of conroe"

I remember conroe's game performace was 20% greater at 2.67 ghz vs 2.8 ghz athlon fx in fear and now it seems to be 17.5% at 2.93 ghz vs 2.8 ghz athlon. Somehow from the performance preview to now AMD processors caught up 10% of the performance? Seems to me conroes lead is more like 15-18% of equvalent AM2 chips at this point in games and that only at 1024x768, a resolution most users dont even run their systems on so tell me exactly how is this such a gloom and doom scenario for AMD? From the woodcrest to the conroe benchmarks I dont see a lead that gives Intel much of a headroom for the months to come. AMD has been making small steady increases in performance since rev D of thier athlon but Intel processors have actually worsend in performance in their newer revisions so this new core's amazing increase in performance is about 3-4 years in the making. At this point waiting for AMD's answer just seems so much more appealing. Q4 2006 through 1H 2007 will be a great time for microprocessor enthusiasts.

What I really want to see is a comprehensive review. All the new Intel procs and all the most current AMD X2 and FX procs in a big chart filled review to tell us where the performance is at.

Add in the 'real' world pricing and that will be the pudding we are all looking for.

Nice, but we all know it's not going to be cheap.

20% faster, probably 50% more expensive - I'll file it into the "wait until it's good value" category that so many other things I don't own fall into.

So, could some AMD fan tell us how Conroe doesn't really exist and this was all a sham by Intel? It's kind of weird to read one of these discussion threads without such a post.

Where is the NDA now....

A Hotel room, eh? Wonder if the audio tests had pr0n music goin on?

Considering the Intel processor has a 1066 MHZ FSB and is running dual channel DDR2-800 it is not a shock that it doesn't have the memory bandwidth that the new dual channel DD2 AMD procs have.

It's nice to see that the early benches were fairly accurate in showing off Conroe's performance. Now Intel just needs to launch them in usable quantities (read: Not Just Dell), and at an attainable price.

*shhh* Don't tell 'kitty about this, it might make its head explode ... ;-)

Ultimatelly it will all boil down to the price. Especially at the midrange. So I am far from crossing AMD just yet.