ClawHammer: Everything Goes As Scheduled

[NFS4]

http://xbitlabs.com/news/story.html?id=1013575024


<< As our sources in AMD have reported, the closest partners of the company have already received the first samples of the ClawHammer processor for the uni-processor stepping A0 systems made with the 0.13-micron technology in Dresden. Thus, the introduction of x86-64 technology goes as scheduled: AMD was going to start ClawHammer sampling in the first half-year. The first ClawHammer release in the desktop variant is scheduled to the fourth quarter of this year, and there?s no reason to doubt the reality of these plans. Also, the first ClawHammer samples are reported to be showcased at CeBIT show, which is held in the middle of March in Hanover.

We shall remind, that AMD will introduce a number of improvements in Hammer directed both to support the 64-bit code, and to improve the performance. 8 additional SSE registers and 8 general-purpose registers will be added in Hammer. Simultaneously, all general-purpose registers will be extended to 64 bits. Also, AMD brings out support for SSE2 instructions system and improves prefetching algorithms. The processor will have an integer pipeline lengthened up to 12 stages and 17-stage FPU pipeline, with good scalability for further frequency increase. Also, the CPU will have a built-in one- or two-channel memory controller supporting PC2700 DDR SDRAM. >>

 

[AGodspeed]

I guess AMD is a little further ahead than just "first silicon". The fact that they are going to be sending ClawHammer samples to their partners signals to me that they've gott CH's running at full spec, with only months more of debugging and verifying to go....

EDIT: Looks like you beat me to it NFS4.

 

[Bluga]

<< the closest partners of the company have already received the first samples of the ClawHammer processor for the uni-processor stepping A0 systems made with the 0.13-micron technology in Dresden >>



that's adrealine pumping

 

[Bluga]

<< The processor will have an integer pipeline lengthened up to 12 stages and 17-stage FPU pipeline, with good scalability for further frequency increase. >>



how many stages does that count?

 

[Bluga]

<< Also AMD official confirmed that mass Hammer processors will be provided with an HIS metal lid improving the heat dissipation from the processor die and protecting it against mechanical damages. >>



finally.

 

[AGodspeed]

<<

<< Also AMD official confirmed that mass Hammer processors will be provided with an HIS metal lid improving the heat dissipation from the processor die and protecting it against mechanical damages. >>



finally. >>



Agreed.

 

[Soulkeeper]

i hope it is high quality copper or something so us enthusiasts won't have to crack it off

 

[WyteWatt]

<< << Also AMD official confirmed that mass Hammer processors will be provided with an HIS metal lid improving the heat dissipation from the processor die and protecting it against mechanical damages. >> >>




Does this mean it will not be as easy to break the amd chip when putting on and taking off the heatsink/fan ? And how much lower do you all think the temps will be?

 

[Mookow]

<<

<< The processor will have an integer pipeline lengthened up to 12 stages and 17-stage FPU pipeline, with good scalability for further frequency increase. >>



how many stages does that count? >>



12. The number of pipeline stages people refer to for a given processor is usually the number of integer stages.

 

[Lord Evermore]

I don't think a heat spreader will necessarily IMPROVE cooling very much, it'll just make it a bit more reliable. We already have solid copper heatsinks, a heat spreader will contact the heatsink more but it's still only got contact with the die the same as our current sinks have contact with the die. And usually the heat spreader doesn't have the same (not as good) compound and contact as we can make ourselves. Plus it'll help stop cracked cores.

 

[saltedeggman]

P4 has heatspreader and by removing the it, temp can drop a few degree C

 

[LukFilm]

Is ClawHammer = Hammer? If not, what features will ClawHammer have? Is this the desktop version? What FSB?

 

[Mookow]

<< Is ClawHammer = Hammer? If not, what features will ClawHammer have? Is this the desktop version? What FSB? >>



Clawhammer is part of the hammer line of processors. Its the lower end version of hammer. The high end version is called sledgehammer and, I think, features 2 cores per physical processor.

 

[Adul]

Hope we can still unlock the hammers when it comes out.

 

[BFG10K]

Looks good but don't forget that the longer pipes will decrease the processor's IPC.

Also I wonder if it will finally have the same thermal protection as Intel CPUs have.

 

[AGodspeed]

<<

<< Is ClawHammer = Hammer? If not, what features will ClawHammer have? Is this the desktop version? What FSB? >>



Clawhammer is part of the hammer line of processors. Its the lower end version of hammer. The high end version is called sledgehammer and, I think, features 2 cores per physical processor. >>



Hammer is AMD's 8th generation microprocessor architecture. ClawHammer is the desktop variant of the "Hammer" architecture. SledgeHammer is the server variant (2-8 way microprocessor systems) of the "Hammer" architecture.

To my knowledge, SledgeHammer will not have 2 cores per die until smaller geometries are reached in the next few years (.09u, .065u, etc.).

 

[Adul]

30% faster then an athlon XP!! Good GOD!! :Q

 

[sandorski]

Yum. Hopefully the prices will be comparable to XPs, if they are, I sure wouldn't mind getting 1 in a year or so.

 

[Athlon4all]

<< Also AMD official confirmed that mass Hammer processors will be provided with an HIS metal lid improving the heat dissipation from the processor die and protecting it against mechanical damages. >>

Yahoo!!!! That's great! No more crushed cores!

<< ClawHammer will be about 30% faster than Athlon XP working at the same frequency. >>

Hot Dog!!!!With that, a Hammer at 2GHz should be able to beat a P4 clocked as high as 3.5GHz!!!I cannot wait for HAMMER!!!!!

 

[m2kewl]

Wow - that's awesome info.

*wipes the drool and coffee off the keyboard*

 

[johndoe52]

<< ClawHammer will be about 30% faster than Athlon XP working at the same frequency >>


So is there going to be a new PR system for the clawhammer too? Silly marketing.

 

[NFS4]

<<

<< ClawHammer will be about 30% faster than Athlon XP working at the same frequency >>


So is there going to be a new PR system for the clawhammer too? Silly marketing. >>


Anand said that although he didn't like the PR rating, that it is actually working...that's all that matters

 

[KenAF]

john,

<< So is there going to be a new PR system for the clawhammer too? Silly marketing. >>

Nope...that's why AMD projects a performance rating of 3400 for the first Clawhammer. While a 2.0GHz Athlon has a PR rating of about 2500, a ~2.0 Clawhammer will have a performance rating of about 3400. If you increase that 2.0GHz Athlon 2500+ rating by 33%, as indicated by the Xbit article, you just about hit the 3400 figure that AMD has said will be the rating of the first Clawhammer. At its presentations, AMD has always maintained that the Clawhammer would significantly improve its performance per clock (IPC), such that it would attain greater performance at identical clockspeeds.

Clawhammer is essentially an [substantially] improved version of the Athlon, as you can see right here; it does not contain multiple cores. Hammer increases the number of integer pipeline stages from 10 to 12, and the FPU pipeline stages from 15 to 17. All else being equal, increasing the pipelines on the Clawhammer would reduce IPC, but AMD has made so many other improvements that they more than offset the disadvantage caused by branch mispredicts. Of course, AMD will also make other layout improvements; the Clawhammer should scale better down the road, but the processor itself is 30% larger than the .13m Athlon, and produces substantially more heat [than .13m Athlon], so that may limit it somewhat as far as clock frequency (which it will make up for in IPC).

By comparison, the P4 is able to achieve such high clock speeds in good part because of its 20 integer pipelines, 29 FPU pipeline stages, redundant circuitry to get high speed signals across the chip (that's part of the reason why its so damn big!), and asynchronous architecture where parts of the processor run at double clock speed (i.e. 4.4GHz on a P4 2.2), and other parts run at half clock speed (1.1GHz on a P4 2.2). No one should expect Clawhammer to reach anywhere near as high as the P4 on a given process--it won't, but it should make up the difference through improved IPC (relative to the P4's advertised clock). Indeed, as we move to .09u, the differences in clock speed between the P4 and Clawhammer should be even greater as they are now.

Assuming the 30-33% figure for the performance advantage over the Athlon at a given frequency, that would put the initial 3400+ Clawhammer (Jan-Feb 2003) at between 2.0GHz and 2.1Ghz, and the 4000+ Clawhammer (mid 2003) at between 2333MHz and 2500MHz. The 4000+ is the last Clawhammer listed on AMD's roadmap for .13m. According to AMD's roadmap, the 4400+ is set to debut the end of 2003 on a new .09 process. Assuming the same ratios, the would put the 4400+ Clawhammer at about 2666MHz.

Imagine...AMD Clawhammer at 2666MHz while Intel is at 4.4GHz on the exact same .09 process...yet both offer identical performance.

 

[Adul]

Thanks for that insight kenAF

 

[Ausm]

<< 30% faster then an athlon XP!! Good GOD!! :Q >>



Woohooo


Ausm