AMD talks back on Intel's Conroe

[booyaah]

Hi, as you can obviously tell by my post count I am newly registered to the forums, but not to the TH site or the hardware enthusiast scene. I recently came upon this article the other day.

»AMD disputes Intel Conroe performance claims

"It's driven by the fact that they can't talk about their current products, because everybody knows their current products aren't very good," said Henri Richard, AMD's chief sales and marketing officer, in an interview with CNET News.com late Friday.
Richard was responding to a prediction from an Intel executive that the company's chips scheduled for the second half of the year will deliver a 20 percent improvement in performance over comparable AMD products scheduled for release in the same time frame.

Old news to some? Can you say bring it on Intel?

Source: http://www.firingsquad.com/news/newsarticle.asp?searchid=9344&filterLevel=1&up=&page=1

 

[Action_Man]

Old.

 

[ShortyNumber45]

I hope Intel won't blow it.

 

[booyaah]

2 weeks is old? /shock

j/k..

 

[bront]

2 weeks is old? /shock

j/k..

This is the internet. 2 hours is old.

 

[Ycon]

AMD said that they arent quite going to be able to match.

Of course they didnt exactly say that but thats the essence.

 

[mdd1963]

So AMD's position almost seems, "yes, Conroes are faster, but we have products out now"?

 

[apache_lives]

Hi, as you can obviously tell by my post count I am newly registered to the forums, but not to the TH site or the hardware enthusiast scene. I recently came upon this article the other day.

»AMD disputes Intel Conroe performance claims

"It's driven by the fact that they can't talk about their current products, because everybody knows their current products aren't very good," said Henri Richard, AMD's chief sales and marketing officer, in an interview with CNET News.com late Friday.
Richard was responding to a prediction from an Intel executive that the company's chips scheduled for the second half of the year will deliver a 20 percent improvement in performance over comparable AMD products scheduled for release in the same time frame.

Old news to some? Can you say bring it on Intel?

Source: http://www.firingsquad.com/news/newsarticle.asp?searchid=9344&filterLevel=1&up=&page=1

here are some 99% accurate facts:

conroe is here in ~4months
amd is faster for the time
intel will take over the lead in every area
AMD will lower prices to match
AMD might take 2+ years to catch (dont know the full details on the future for them but they always come back)

does that answer things for you?

 

[ShortyNumber45]

It doesn't seems to be a real "fact" but a fanboyish fact.

 

[Ycon]

Nope

 

[mitch074]

At one time, Intel had the lead on AMD: Pentium-II then Pentium-III compared to K-6, especially on FPU operations.
Frankly, you could hardly find more bang for the buck than a Celeron 300A with an Abit BH6 (or BX6) and a fast stick of SDRAM: you plugged the processor in, you instantly got a 450-MHz rock stable system.
Remember that the P-III architecture was a long evolution from the Pentium Pro design... and then led to Pentium-M, then Conroe.

P-4 was probably the biggest long-lasting POS to have ever come out of Intel's labs - right there with the MTH for i8xx chipsets.

Then, AMD took the lead with the Athlon: the design aged quite nicely and had no competition from Intel (I used P-4's at work, they felt sluggish compared to my half-clocked Duron )

Now, Intel decided that there might have been something more to squeeze out of the P-3/P-M design (took them long enough).

What I wonder about, is: how comes AMD couldn't innovate much over these almost ten years on a design level, while a small Israeli Intel lab could?

My guess is that Intel will enjoy some time at the lead with Conroe, Amd will test-drive new platforms and new manufacturing processes with a proven architecture, then get out a new architecture. After all, it's strange that Amd's roadmaps suddenly got so vague...

Intel's roadmaps are usually vague and announce paper product launches, but they're being incredibly open about Conroe: after all, you could already build Pentium-M based desktop systems yourself, so it is little surprise that a tweaked processor running on the same platform would get out eventually.

Right now, we have Intel being open about a refit product (a well refit one, mind you) and Amd being rather quiet about the matter: because they didn't see it coming, because they couldn't innovate enough, or because they're preparing something?

K6 design came out in 1995; K7 was out by 1998, K8 is a K7 with a 64-bit unit... Would it be unreasonable to think that Amd must have spent some resources exploring something else?

end rant.

 

[ltcommander_data]

Those comments are old because of the context. AMD made those comments to respond to Intel boasting they would have a 20% performance advantage over AMD's future chip. That article was made on February 28, and at that point Intel's claims were completely unproven and unfounded. Since then, Intel has demonstrated benchmarks that substantiate their claims. You could question the validity of the benchmarks but at least they now have numbers to back up their claims.

 

[booyaah]

here are some 99% accurate facts:

conroe is here in ~4months
amd is faster for the time
intel will take over the lead in every area
AMD will lower prices to match
AMD might take 2+ years to catch (dont know the full details on the future for them but they always come back)

does that answer things for you?

2+ years to catch? I thought AMD was just waiting to work out the bug with DDR2 ram for the AM2 before 'letting it out of the gates'. Indeed seems like you have many fans over at Intel.

 

[Snorkius]

...K8 is a K7 with a 64-bit unit... Would it be unreasonable to think that Amd must have spent some resources exploring something else?

In that sense, the vast majority of all CPUs are just improved 8088s.

Damn these companies for not exploring alternatives! :roll:

 

[sepheronx]

here are some 99% accurate facts:

conroe is here in ~4months
amd is faster for the time
intel will take over the lead in every area
AMD will lower prices to match
AMD might take 2+ years to catch (dont know the full details on the future for them but they always come back)

does that answer things for you?

2+ years to catch? I thought AMD was just waiting to work out the bug with DDR2 ram for the AM2 before 'letting it out of the gates'. Indeed seems like you have many fans over at Intel.

DDR2 will only give AM2 a boost of around 5% over 939, Conroe will have 20% over 939, so Conroe still wins. And no they dont have a bug, they tested DDR2 800 with AM2 and it seemed about right to what amd stated. So there is no bug and stop with that nonsence, if it had a bug, we would see some serious issues. Not to mention AMD is not bandwith hungry on memory, it likes low latencys, and that is what DDR2 doesnt have over DDR.

 

[conroe]

Here are some nice old comments by the head of the beast,
"We'll work our buns off to beat [Intel]", said Hector Ruiz
http://www.dailytech.com/article.aspx?newsid=418

Now why did he not say "to stay ahead of Intel?" He knew. It's no suprise he did, alot of people did.

So when Conroe is out Intel has a good lead and will keep it atleast untill AMD goes 65nm and maybe even then. After than Intel goes 45 nm and gains even more. In 2008 Intel will change more and leaves too many questions to even guess.

 

[mitch074]

I haven't said that; what I said was the K-7 design was tweaked and saw a 64-bit unit added to it. However, the general design, the integer, SSE, 3Dnow and FPU units are still mostly the same, the crossbar system too...

The 8088 was using 20-bit addressing, there was deep architectural changes made to it to address 24-bit (286) then 32-bit (386) memory spaces; then, if went from a completely CISC architecture to an 'almost' RISC architecture (486 - also saw desynchronized CPU/system bus + integrated FPU in DX), then the memory bus was widened from 16-bit to 32-bit (386DX) to 64-bit (Pentium). Pentium pro saw the integrated 2nd level cache and a specifically 32-bit architecture (until then, processors were still heavily 16-bit processing oriented)

Would you consider the addition of an SSE unit (128-bit floating point instructions!) to the Pentium II a big architectural change? So then, what's so exciting about a 64-bit unit? Keep in mind I do use 64-bit operating systems (Linux power!) and it's certainly interesting, however for mundane stuff I don't see much difference between 32-bit and 64-bit performance - as it was supposed to be.

AMD's big strenght lies in its initial decisions: it IS quite impressive for the K7 design to scale so well: it has seen SSE, multicore, memory controller, 64-bit capabilities added to it, while going from 600 MHz to 3.00 GHz.

The K6 was a very good improvement over the K5, and the K7 left the K6 in the dust. The K8 doesn't whip a K7s ass that much.

 

[fainis]

intel will take over the lead in every area ...ironically he said....
get real my friend ...what`s the matter with you....are you affraid to see the reality.......some areas he said ...haa.....if you carefully watch the tests done ... some 2,66 conroe against the fx60...all the tests ...and i mean all the tests ..favour Intel....
and what is the conclusion we can drawn from that test....a 2.66 conroe against the fastest amd on the market.......it is as simple as it can get .... that particular cpu and i mean conroe blows to pieces the fastest amd on the market....
you boys want to escape from reality ....... lock yourself in your little room and isolate from the rest of the world ..... it`s ok you can cry ... it helps sometimes

 

[mdd1963]

"DDR2 will only give AM2 a boost of around 5% over 939, "

The latest benchmarks I saw using an M2/DDR2-800 equipped x2- 4800+, gaming framerates were still 1-2 frames/sec slower.....

Perhaps with DDR2/1066 and a few more BIOS tweaks they will reach parity with less expensive standard 939/PC3200 rigs?

 

[upec]

I would not be suprise that Intel was able to catch up.
Intel is a much larger company and have a lot more resource than AMD.

 

[gOJDO]

I guess they will try to optimize the memory controller in the way of Intels Core architecture Smart Memory Access. They will add more L2 cache if Z-RAM technology for real is implemented in the 65nm SOI. Maybe HTT will be improved to 1600MHz full-duplex and more links will be added. I would like to see 4 cores on single chip linked in the way like 4 single-core Opteron 8xx are linked between with direct connect architecture. Maybe some more multimedia power will be added, the SSE4 and some extra clock, + 400MHz for quadcored and +800MHz more for dualcored chips.
But that will be after Conroes lounch as we can expect now. And maybe we can see some two processor systems with quadcore Clovertown processors(8 cores total in system) in the same time.
We can only guess and talk now. We don't have any kind of simillar chip in our hands, and we can only hope that the articles with informations, benchmarks, overviwes and etc. about them are not lies.
Let's see, what we have available now and what we got from both Intel and AMD. Lets see what are we using that for, before we judge who is better or worse.

 

[admiral25]

Correct me if im wrong, but the a64 and opteron are 90nm cpus while the conroe are 65nm(like the presler)? Why not wait till amd drops there new 65nm chips with new core setup b4 we say either one is better than the other.

 

[conroe]

If you want to wait untill next year, go ahead. AMD may win some benchmarks with K8l, but conroe will still be ahead in many more. There is not much info on K8l, just that it has a very good FPU.

 

[choknuti]

Funny how I really don’t want Intel to dominate AMD. I prefer a situation like the GPU market. Better competition leads to better prices for us. When/if Intel totally dominates AMD then I feel we wont be seeing sweet deals like the 805 anymore

p.s. I think the Conroe will be a great processor but am waiting for the release (and the release of the AM2) before making a final decision. Specially good for me since I hope to retire my old rig sometime November or December this year.

 

[mdd1963]

"while going from 600 MHz to 3.00 GHz. "

from 500 Mhz to 2.8G, officially....

They just can't seem to crack that 3G nut!