Sign in with
Sign up | Sign in
Your question

Hammer Delayed....Again

Tags:
Last response: in CPUs
Share
January 31, 2003 4:11:14 PM

Got this off the enquirer:

AMD Athlon 64s delayed until September

Announces launch dates for Barton, Opteron

By Mike Magee: Friday 31 January 2003, 10:42


AMD WILL say today that it won't launch the Athlon 64 (Clawhammer) until September this year.
The firm also will confirm the release of a Barton XP at 3000+ on the 10th of February – another Barton at 3200+ will appear in the middle of the year.

But AMD will introduce its Opteron (Sledgehammer) chips at speeds of 1.4GHz, 1.6GHz and 1.8GHz in New York on April 22nd.

AMD, officially, will say that the Athlon 64 is not out until September because it is waiting for the introduction of a suitable 64-bit operating system. This, The INQUIRER believes, is the Windows 64 bit version specifically for the Athlon64.

If it is Microsoft that is behind the latest delays - they were supposed to be available in December last year - some will wonder whether that release will be, err, released on time too.

Even if AMD launches the Clawhammer in September, there are still logistical problems to solve - getting the chips into the distributor, system builder and into the retail channels on time.

The Opteron doesn't need to wait for Microsoft because there are suitable -- read Linux 64-bit operating systems -- ready to roll for the server based chip.

The Opteron will be a little under half the size of the Intel Itanium chip, which comes in at 400mm2, and it will push this fact like crazy at the launch.

Despite Intel's positioning, AMD still feels that the Opteron server chips have a fair chance of cracking the Itanium's still rather precarious hold on the X86 64-bit processing market.

Although the Itanic can run X86 code, AMD maintains it's not really an X86 chip at all, because of the extended instruction set it uses, based on the EPIC architecture.

The Athlon64 news will come as a blow to those who have waited and waited for the launch of the Clawhammer.

But, as Jerry Sanders III once said: "Together with Intel and Microsoft we are the Holy Trinity". And it cannot pre-announce the Windows OS for Microsoft.

The official line from AMD is that the Barton 3000+ and future speed grades, will give it sufficient clock speed and performance gains to compete with Intel's Pentium 4 offerings between now and September of this year. µ



Let's all settle our political differences with a LAN.

More about : hammer delayed

January 31, 2003 4:20:35 PM

The Opteron isn't though, and it's a good excuse to not release the chip w/o a viable OS solution (aside for linux). Hopefully AMD uses this time to make the chip better (DDR 400 ready) and so on...
January 31, 2003 4:59:24 PM

Does this wonder anyone? I mean, really? Looks again like it's going to be end of the year before anyone can actually buy an Athlon64 for real... like in 2001 and 2002. Q4 2003? Did AMD not want to release the 0.90 micron version of the Athlon64 at that time? Oh, all right, another delay then. How lucky AMD is that stockbuyers don't see this coming...
Related resources
January 31, 2003 6:23:29 PM

DDR 400 will never be use in Server XEON move to 666 FSB not 800 FSB so only PC2700 will br use in XEON base server on Nocana

Just next to the lab and the bunker you will find the marketing departement.
January 31, 2003 6:27:43 PM

good move from AMD they cannot beat P4 on desktop uless they are able to reach high speed on A64.Opteron will face XEON so be a Xeon killer.MP have give them a low end reputation also according to the inquirer retail channel are angry at the support that AMD give for there MP ligne.AMD will pay for that opteron reputation will take longer to built that expected expect in europe.

Just next to the lab and the bunker you will find the marketing departement.
January 31, 2003 9:27:50 PM

Well, their excuse actually <i>does</i> make a lot of sense.

1) We know already that the T-bred/Barton can scale farther than AMD has officially taken it. Even the Barton 3000+ is actually a bit under the highest speed T-bred MHz-wise. Sure it will run rather hot at those speeds--but hey, the Northwood already draws more power and runs even hotter just to achieve the same performance levels.

2) The advantages of 64-bit are negligible for the desktop, and without a 64-bit MS operating system, they're pretty close to nil.

So if the Barton can scale beyond the Athlon64's scheduled release speed, and most of the other advantages of Athlon64 will be unrealized anyways, why actually release the Athlon64? Not to mention which, even if Athlon64's yield is just outstanding right now, Barton's yield is probably still even better due to the more mature process.

Sure we'd like to see Athlon64 in April, but it doesn't make good business sense for AMD. It probably wouldn't have a lot of concrete benefit for us either--it would just appease our technolust a bit.

<i>I can love my fellow man...but I'm damned if I'll love yours.</i>
February 1, 2003 2:10:38 AM

It just makes me wonder how much of AMD's marketshare could erode, since Intel is planning all sorts of releases, keeping themselves in the limelight. Then again, only time will tell.



This post is guaranteed fresh for up to 7 days after the date printed on the box.
February 1, 2003 2:15:19 AM

except on Xeon MP market they will release maybe something that can hold on to opteron in 1H of 2004 in more that 1 year.I think intel plan to put this XEON mp obsolete to deerfield

Just next to the lab and the bunker you will find the marketing departement.
February 1, 2003 2:21:05 AM

I think Opteron to Xeon, Opteron will have the market for sure, and will dominate it for a long time, if not Intel ceasing Xeons for Itaniums.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
a b à CPUs
February 1, 2003 3:40:09 AM

But imagine the gaming performance that *could* be available if games were written to take advantage of 64-bit technology? With the latest video cards, maybe we could get 1000FPS in Q3!

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
February 1, 2003 12:58:39 PM

If one gets 300fps in Q3A with one PC, then you would get with the "same" 64-bit version of that PC approx. 290fps! not 1000fps.

Some people still believe that 64-bit will make everything faster but in the beginning (the next 1-2 years) no games and programs will see amazing performance increases simple due to being ported to 64-bit. The reason for this is because you will have much more dataoverhead with 64-bit games/programs which the first generation of Hammer CPU won't be able to manage fast enough to get a speed increase. It just doesn't work out for the moment. Therefore this whole 64-bit hype is hopeless, unless you're working in the server domain.

As Carmack already stated, there will be no Doom3 64-bit version because the performance would too bad...
February 1, 2003 1:07:15 PM

Epic contradicts your statement, they have already developped a 64-bit version of UT2003 and have demonstrated a ~15% performance increase.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
February 1, 2003 3:04:33 PM

1000FPS? Even rail gun maestros are probably perfectly happy with 100FPS. I think you're missing a sarcasm smiley there or something. Here you go: :wink:

<i>I can love my fellow man...but I'm damned if I'll love yours.</i>
February 1, 2003 5:08:44 PM

ok then ... rofl watever
February 1, 2003 7:14:17 PM

I'll definetly be getting a hammer system some time....I'm looking at 2 or 4 processors (man, if they find a tweak to the athlon 64's for multiproc support that would be all hotness!) and will be checking out how my simulations r doing on 64 bit vs 32. Hmm, this windows 64...will it run 32 bit apps just fine?

"If everything seems under control, you're just not going fast enough."
- Mario Andretti
February 1, 2003 10:10:25 PM

funny, in the interview I read, Mark Rein stated that the 64-bit performance of UT2k3 was somewhat worse than the 32-bit one.
February 2, 2003 2:09:36 AM

I also i have see that 64 bit result in slowdown most of the time

Just next to the lab and the bunker you will find the marketing departement.
February 2, 2003 2:52:05 AM

Without any real compiler right now, I doubt there is any benefit to use a very young new extension that barely has any support.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
February 2, 2003 2:55:07 AM

[frothing at mouth] me no sar happy with any amount of frames! Must have more!!!![/forthing at mouth]

"If everything seems under control, you're just not going fast enough."
- Mario Andretti
February 2, 2003 5:44:10 AM

Ok, just to clarify, x86-64 brings a lot more than just "64-bitness". It brings along 8 more general purpose registers and 8 more SSE/SSE2 registers. Something which is very welcomed in x86 and would more than account for 15% if not more of a performance increase with the proper compilers. The 64-bitness has nothing to do with any increases in games, the extra 8 SSE/SSE2 registers do (assuming the game utilitzed SSE/SSE2) and the 8 extra GPR's would help as well.
There's also the other enhancements of the Hammer design over the current K7, including the integrated memory controller, the extra micro-ops packing and bundling, the extra decoder in 64-bit mode (although I doubt this will help much). These will help in games or any other application when compilers are set to utilize x86-64, 64-bit or no 64-bit. And as we've seen in some of the leaked benchmarks, they may bring a close to 30% per clock improvement over the current AthlonXP. The benefits of utilizing a 64-bit processor when it is really needed (when 64-bit integers are needed or more than 4 GB of memory is needed) is well beyond 30% per clock. It could be close to 300%.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
!