Poll: Is ClawHammer worth waiting for?

[winterlude]

Poll: Is ClawHammer worth waiting for?

In the coming year, AMD is coming out with thoroughbred (.13 version of Palomino) and Barton (not much hype at all about this one) for the desktop market. But they also plan to launch the much ballyhooed Hammer which is aimed at all markets (desktop, server, workstation, mobile) and is certainly the future direction of AMD.

During that time, Intel won?t be standing still, with the Northwood pending, and the mysterious Banias to enter the market to compete with ClawHammer and Barton on the 32 bit front also a year from now.

VIA?s Richard Brown has stated in today?s article that the Hammer platform?s cost will be ?competitive,? but competitive with what? Banais, Barton, or Xeon($$).

In the original AMD roadmap (well, one from a year or so anyway), Hammer was supposed to be debuting right about now, but for a number of reasons, AMD put it off. But even a year from now, will applications in the desktop market be mature and restricted enough to start porting to the 64 bit arena? But lets face it: desktops now are more powerful than servers and workstations a few years ago. As a result, people are doing with desktop machines what servers and workstations used to do such as rendering, encoding, and video editing. But that?s still a small segment of the market (although it is growing). So what it all boils down to, in other words, is do you think you would pay more for the performance and forward compatibility of the Hammer as opposed to sticking to the Thoroughbred, Barton, or (assuming, probably falsely that Intel?s 32 bit solution would be cheaper than AMD?s 32/64) Banias?

 

[Agent004]

Is it worth waiting for?

Well, is it a bit early to assume it will be great and everything works upto expectation? I think I reserve my judgement until I have some more info on it.

 

[Buzzman151]

Also.. are any of these going to have higher FBS's to untilize the increased memory bandwidth on boards like the Nforce. I'm sure that more boards like that will be soon to come. It seems to be even with the theoretical doubling to the bandwith to the ram, the Nforce's memory bandwidths tests we're flying past that of the kt266a which is of course due to this FSB restriction. A chip like the P4 would eat up that bandwith and acually be able to utilize it. Hopefully AMD won't be falling behind.

 

[AGodspeed]

Also.. are any of these going to have higher FBS's to untilize the increased memory bandwidth on boards like the Nforce.

ClawHammer is rumored to debut with an 800MHz FSB.

Is it worth waiting for. It's too early to tell really.

 

[Buzzman151]

So is the Clawhammer suppose to be more of server, workstation, or desktop use? Plus with that great of and increase in the FSB, won't there be a need to increase the cache or am I just smoking crack?

 

[Demonicon]

I'm sure there will be an increase in cache. I can't wait for it myself.


Is it worth waiting for? For me it is, but I have a very good system right now, and don't see the need for an upgrade until then. HTH

 

[AGodspeed]

So is the Clawhammer suppose to be more of server, workstation, or desktop use?

The "Hammer" series of processors is going to include two different versions:

1. For server, workstation use (SlegeHammer).
2. For desktop, mobile use (ClawHammer).

Plus with that great of and increase in the FSB, won't there be a need to increase the cache or am I just smoking crack?

I don't think it's really necessary to increase cache when the FSB scales up, so I'll agree with you by saying you're smoking too much crack.

 

[Bovinicus]

I agree with the other posters. It is far too early to tell.
Here are some improvements the core will feature:

- Improved branch prediction unit
- Integrated memory controller
- Support for SSE2 instructions
- .13 micron SOI manufacturing process
- 800MHz FSB (Possible)

The pipeline length of this CPU will also be increased to 12 from the current 10 of the Athlon, but the branch predictor should help alleviate any performance problems this might bring about. All in all I think some very reasonable decisions have been made from a performance standpoint. For more information check out Anand's article about the architecture. I think it looks promising, but Intel is no slouch. They have a damned good R&D department. I think they will be spitting out some competitive solutions.

 

[winterlude]

On paper anyway, Clawhammer will be able to dominate Thoroughbred since the hammer line puts the Northbridge and the memory controller in the CPU. Whether or not Barton will do this too, who knows? Unless Barton is much cheaper than the clawhammer solution (i.e. becomes similar to what the duron is to the athlon), I don't see the point in AMD persuing it.

 

[winterlude]

I think it's ineveitable that Intel will, in this case, follow AMD's lead and produce their own 32/64 hybrid solution. AMD is creating a potentially very profitable niche market by creating a series of hybrid CPU's and a lot of hardware and software companies are conferring with them like VIA in today's article and MS, which is building a 32/64 OS--probably still under the XP umbrella.

There's no way the Intel could sit around while getting hammered by Hammer. The big question is how quickly will 64 bit apps start to crop up that appeal to people who do most of their work in a 32 bit environment? If the explosion of apps after win 95 came out (which bridge the 16/32 bit gap by incorporating DOS) then AMD will clobber Intel if Intel doesn't offer the public a direct alternative i.e. their own hybrid. Right now, the IA64 "can" run 32 bit applications, but it does so at a huge performance cost. Intel may simply work on improving the IA64's 32 bit performance to make it viable as a hybrid alternative.

Intel had planned to take their sweet time in porting the tech world at large from 32 to 64bit, but AMD's Hammer will certainly jump start the process and may very well catch Intel with their pants down.

 

[Athlon4all]

Hammer is an exciting product and it will definately give Northwood a run for it's money, but is it worth putting off building a Athlon XP system now? No, not at these prices on XP's.

 

[AGodspeed]

There's no way the Intel could sit around while getting hammered by Hammer. The big question is how quickly will 64 bit apps start to crop up that appeal to people who do most of their work in a 32 bit environment?

For Hammer to be a bigger success than the Athlon (which on paper it certainly seems that way), the above must occur. If successful 64-bit applications are released to the general public next year, Hammer will likely be a huge success, only because Intel doesn't have a x86-64 type processor in the works (well, at least not in 2003 it seems).

If I was a betting man, I'd take Hammer over Northwood, but it's way to early to make much of a logical guess anyway.

 

[jeffrey]

click edit
select add poll
title your poll
create choices
edit message
= your poll