Monday, May 22, 2006

AM2 party time. Frags readied for CONroe

DELL (DBA Alienware), HP, Lenovo, Microsoft, VIA, MSI, ASUS, NVIDIA, ATI, ECS, FUJITSU SIEMENS, GIGABYTE, KINGSTON, CORSAIR, SiS... all the big names you can find, are in for one big party for AM2 launch. The FX60 was 20-40% faster than Intel's fastest chip, the performance gap just got bigger.

One June 1st, AMD will demo some next generation futuristic technologies that will frag Intel amateurs like hell. In July, Intel is going to launch the Conroe. My suggestion to AMD is this: at the time of Conroe launch, release a quadcore AM2 chip built on 65nm with the K8L core. It doesn't matter if 65nm yield is only 10%. Just produce 5000 such chips. Send 100 to review sites, make the 4900 forsale at $3500 each. The goal is simple, show what AMD's true quadcore is capable of and push Intel further down the performance curve. This will create a halo effect on the whole AM2 product line.

15 Comments:

Anonymous Anonymous said...

I think that the best idea is to lower X2 3800+ prices to 200$ and introduce X2 and FX at 65nm.

10:23 PM, May 22, 2006  
Anonymous Anonymous said...

"My suggestion to AMD is this: at the time of Conroe launch, release a quadcore AM2 chip built on 65nm with the K8L core."

uh.... what makes you think AMD is capable of doing that??

10:53 PM, May 22, 2006  
Anonymous Anonymous said...

Here's one at http://www.techreport.com/etc/2006q2/woodcrest/index.x?pg=1 Intel's Woodcrest processor previewed
The Bensley server platform debuts. Conclusion is Intel's Woodcrest "fragged" AMD's Optys, for GOOD this time (in EVERY TEST including power dissipation)..

12:48 AM, May 23, 2006  
Anonymous Anonymous said...

AM2 brings a lot of benefits to end user:
- better quality chipsets (ATI, NVidia - see TechReport Benchmarks
- better upgradability)
- etc.

However, it appears that Conroe will be a formidable contender, as shown by the FX-62 versus Conroe benchmarks.

6:32 AM, May 23, 2006  
Anonymous Anonymous said...

AMD is going to demo next gen chips on June 1? Where is this taking place? Do you have more info than what's been posted? I'm interested. Thanks.

6:45 AM, May 23, 2006  
Anonymous Anonymous said...

June 1st? How do you know that??

7:39 AM, May 23, 2006  
Anonymous Anonymous said...

I'm still trying to figure out why the great Anand and INTELanites are saying AM2 is only 3-5% faster than 939 when I've seen 3 websites today reviewing AM2 and it's consistently 10-20% faster w/ RAM Timings worse than Conroe's Performance.

If you look at the good mobo, the RD580, it's up to 30% faster than the NF5, yet only ONE website is using the RD580 and everybody else is comparing Conroe to the NF5, even though it's a slower motherboard. Also, the NF5 FX-62 w/ RAM of 4-4-4-15 2T Timings? Are you kidding me? They also include 4 512 sticks of RAM in 1 test vs. 2 in Conroe...it must be national Rig Benchmarks day (either that, or somebody is DUMB).

The truth is: AM2 is 10-20% faster than Socket 939 w/ Bad Ram and a Good Mobo, now I'd like to see an FX-62 @ 2.8GHz w/ 3-2-2-8 1T DDR2-667 like Conroe, and I guarantee Conroe will not be the king.

I agree, Conroe is a nice chip from Intel, but we had to wait 3 years for Conroe? We got K8L right around the corner and you could cut off HALF of it's features and it'd STILL whoop Conroe, and yet Intel expects us to get it? What about Intel's Quad-Core? Wasn't it getting beat by a 246 .13nm and a 3000+? WTH? You gotta be S-M-R-T to buy Intel.

7:48 AM, May 23, 2006  
Anonymous Anonymous said...

"If you look at the good mobo, the RD580, it's up to 30% faster than the NF5, yet only ONE website is using the RD580 and everybody else is comparing Conroe to the NF5, even though it's a slower motherboard."

Why not 666% faster? It is well known that in K8 world chipset means nothing, as memory controller is integrated in NB.

8:20 AM, May 23, 2006  
Anonymous Anonymous said...

"Why not 666% faster? It is well known that in K8 world chipset means nothing, as memory controller is integrated in NB."

According to you (and every Intel fanboy) a motherboard has nothing to do since the memory controller is gone? Wow...that's great logic. Let me lay some truth on you since you're a little lack in that department (might try hirin some AMD peeps, might actually be PROFITABLE for a change):

There's a few other things in a motherboard than a MC, want to find out what?

1) HDD Controllers
2) The big cheese - The Memory
3) Peripheral Controllers
4) VGA Bus's
5) A Southbridge (sometimes inside the NB)
6) HTT Speed

I would assume 1, 3 and 6 are okay, but 2, 4 and 5 are ridiculous.

2) You'll say the Memory Performance doesn't matter or pertain per motherboard? Oh okay, I guess we all can run Foxconn motherboards OH WAIT, NO! LET'S RUN PCCHIPS!!! HELLZ YA!!

4) IDF proved this Tru64

5) You have to be a Intel fanboy to not know the truth about this one

8:41 AM, May 23, 2006  
Blogger TheKhalif said...

I'm hoping that they release K8L as a dual core for desktop and Quad ONLY FOR 2xx/8xx. They have managed to get good increases from AM2 as the benchmarks on AMDZone show and DDR2 1200 will be awesome.

10:02 AM, May 23, 2006  
Anonymous Anonymous said...

"The truth is: AM2 is 10-20% faster than Socket 939 w/ Bad Ram and a Good Mobo, now I'd like to see an FX-62 @ 2.8GHz w/ 3-2-2-8 1T DDR2-667 like Conroe, and I guarantee Conroe will not be the king."

you guarantee, are u god?

10:20 AM, May 23, 2006  
Anonymous Anonymous said...

"you guarantee, are u god?"

A God, not THE God.

10:49 AM, May 23, 2006  
Anonymous Anonymous said...

Anand also says that AMD will have a preview on June 1. He states that it is not K8L, just a stop gap for the high end systems. Sounds interestingly vague....

11:44 AM, May 23, 2006  
Anonymous Edward said...

"We got K8L right around the corner and you could cut off HALF of it's features and it'd STILL whoop Conroe, and yet Intel expects us to get it?"

uh... I don't think K8L will whoop Conroe. It might win but marginally. The improved 128b SSE might just get even with Core 2's SSE impl.; same with the out-of-order load & improved branch prediction. I'll be glad if I am wrong, though.

My belief is that future CPUs will see diminishing performance gain from generation to generation. One excpetion is media/image (ie. FP) processing, where branches are few and load latency can be hidden. With faster FSB & larger cache/prefetch it's relatively easy to drive up these benchmarks (just look at how Conroe improves over Yonah on these area, while both have essentially the same architecture).

However, AMD's on the right track other things: its architecture scales better to multi-cores, and its ccHT links enable unlimited varieties of coprocessors. These might not increase the absolute benchmark values, but they improve functionality and usability. Right now Intel doesn't seem interested in either - ok, it might be working hard on its CSI, but it definitely doesn't want coprocessors from any other company (imagine an "Intel Inside" box with an Opteron as coprocessor). Ie., Intel's anti-competition attitude is its biggest obstacle here. A big shame I'd say.

3:17 PM, May 23, 2006  
Anonymous Anonymous said...

AM2 is far from fragging anyone, 3 to 5% clock for clock performance increase is not much. I think the real "fragger" will be K8L; intel is far from retaking market share in the Server space, especially when you start adding cpus.

5:15 PM, May 23, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home