Sign in with
Sign up | Sign in
Your question

claw hammer or sledge hammer

Tags:
Last response: in CPUs
Share
March 6, 2002 1:49:04 AM

at the same clock speed, which is faster?

<b> There is no greater sin against oneself than the failure to believe in one's own worth. </b>
March 6, 2002 2:49:56 AM

Sledge should be.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
March 6, 2002 11:39:37 AM

Sledge will provide several enhancements over Claw

1. Increased cache. Estimates are around 1 meg of L2 cache, iirc. Claw will probably have around half that.

2. 128 bit memory bus as opposed to Claw's 64bit memory bus.

These two things combined should give Sledge at least a 5-10% performance boost clock for clock over Claw. BUT you'll pay for that performance boost. Sledge uses a different socket. It's around 940 pin CPU compared to around 750 pins (sorry for the rough estimate..I could look it up but I'm still waking up and don't feel like it;) )

Mark-

When all else fails, throw your computer out the window!!!
Related resources
March 6, 2002 1:27:43 PM

Sledge is targetted at servers while the Claw is made for workstations an desktops, they share the same relation as Celeron and Pentium-II/III/4, or Duron and Athlon!

Sledge has a dual channel memory controller while the Claw has single channel, Sledge will have larger L2 cache and Sledge will be more scalable for MP systems while the Claw may just stop at dual.

Only thing I wish both were socket compatible, one single 940 pin socket that would be the superset of 754 pin Clawhammer socket. they still have time, about ten months to do that, hope they would!

girish

<font color=red>Nothing is fool-proof. Fools are Ingenious!</font color=red>
March 6, 2002 3:45:40 PM

hmmm...

i see. so for performance its still better to go with sledge?

<b> There is no greater sin against oneself than the failure to believe in one's own worth. </b>
March 6, 2002 3:51:02 PM

Interesting so AMD will finally separate the server class processor from the workstation/PC class processor. Will you be able to run dual Clawhammers? Is the ClawHammer also replacing the Athlon MP or is the SledgeHammer replacing it.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
March 6, 2002 3:56:13 PM

I assume you mean run dual Clawhammers. The Clawhammer is for 1-2 way systems, the Sledgehammer for 1-8 way systems.

And I agree with Girish, a compatible socket would be very nice.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
Don't step in the sarcasm!
March 6, 2002 5:42:51 PM

So would the ability to set the FSB according to whatever your hardware can support.

(Note that this is different than unlocked as unlocked would also allow you to overclock the overall clock, where as just the ability to set your FSB would automatically set the multiplier to keep the CPU within the overall clock speed rating that it was sold at.)

<pre><font color=green>//error-proof coding</font color=green>
<font color=blue>void</font color=blue> main(){<font color=blue>return</font color=blue>;}</pre><p>
March 7, 2002 1:07:57 AM

BTW, does anyone here know whether the Athlon T-bred will be released in the U.S. in early/mid march or early april?

"When there's a will, there's a way."
March 7, 2002 1:38:56 AM

There was a rumor on HardOCP I believe, that was quickly quenched that Tbred would be released March 20. Still, at this point, AMD has kept the release date for Tbred quiet. It may well be that AMD will release the mobile version first, and then roll the desktop out a few weeks or months later like they did with the Palomino.

They continue to say that the .13 transition is on or ahead of schedule, so we'll see when AMD wants us to see I guess.

Mark-

When all else fails, throw your computer out the window!!!
March 7, 2002 2:37:15 AM

Quote:
BTW, does anyone here know whether the Athlon T-bred will be released in the U.S. in early/mid march or early april?





<font color=green>"No Thoroughbred for you! Come back, 2 weeks."</font color=green>
March 7, 2002 2:49:14 AM

Question......I haven't had time to read on the sledge as of lately, but also wasn't one of the feature going to be on the sledge hammer a dual cpu core? Based off IBM's own Powerpc 4 dual core technology?

MeldarthX
March 7, 2002 8:51:15 AM

The dual core thingy as far i know is not true. Both the clawhammer and sledgehammer have a single core with a register that supports both 32 and 64 bit operations.

The sledgehammer will have two hypertransport I/O links on die. These support the 2-8 way connections. Which is why it has so many damn pins. So I dont see the pin count getting lower. So needing a different socket will likely stay.

T-Bred? Who the hell knows. For a chip <i>supposedly</i> due out in a couple of months, they sure are quiet about it.

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.
March 7, 2002 9:54:25 AM

they dont, they share the same relationship as the pentiums and xeons

if in doubt blame microsoft...
March 7, 2002 1:38:25 PM

Correct Girish, but just because it's targetted for servers doesn't necessarily preclude it being used on some high end 1CPU workstations. And for those, the features I mentioned are the ones that count :) 

Mark-

When all else fails, throw your computer out the window!!!
March 7, 2002 3:51:15 PM

I'm a little light on knowledge of the future CPUs, so bear with me. I'm assuming after AMD releases the Thoroughbred and Apaloosa, they will continue to the Barton (.13 correct?). After that, does AMD have any plans for the future of Socket462 and the K-7 processor, or have we seen the end of that? On the other hand, the the end of the year, will we all be transferring to the Claw/Sledge-hammer K-8? So many new processors!

"When there's a will, there's a way."
March 7, 2002 4:40:47 PM

According to AMD's roadmaps from those cores you mention onwards, the K7 and Duron chips are marked "As Market Requires" or similar. So, I don't think AMD has firm plans to make further enhancements to Athlon or Duron. In fact, from the looks of it, AMD seems to be phasing these two chips out sooner than their roadmaps indicated just a couple months ago.

Hammer is apparently running ahead of schedule. They may even release it earlier in the 4th quarter instead of towards the end. But this is all just rumors. We'll have to wait and see what happens.

Mark-

When all else fails, throw your computer out the window!!!
March 7, 2002 5:07:37 PM

Actually I think I read that AMD wasn't going to release Barton. This is cuz Thouroughbred is .13 micron, and Barton was to be .13 micron SOI, and they were only going to put the Athlon on SOI to refine the technology for the Hammers. But, the Hammer is going so well that they won't need to bother releasing the Athlon at .13 SOI, as it wouldn't really provide a large benefit to the chip compared to Thouroughbred. AMD probably will try to phase the Athlon/Duron out in favor of the Hammers in 2003.

"Trying is the first step towards failure."
March 7, 2002 5:36:15 PM

What you *heard* isn't backed up by AMD's latest roadmap, It may be right, but presently, AMD still shows Barton as coming out in the 2nd half of this year.

Mark-

When all else fails, throw your computer out the window!!!
March 7, 2002 6:33:28 PM

Hmm...so the socket462 has a future! Is AMD even considering extending the market of K-7 processors all the way till next year? Would be interesting to see a 3000+ AXP running on SocketA. Heck, maybe we'll see that with Thouroughbred or Barton. BTW, is the "new" Duron Apaloosa going to be PR rating or Mhz?

"When there's a will, there's a way."
March 7, 2002 6:33:31 PM

Hmm...so the socket462 has a future! Is AMD even considering extending the market of K-7 processors all the way till next year? Would be interesting to see a 3000+ AXP running on SocketA. Heck, maybe we'll see that with Thouroughbred or Barton. BTW, is the "new" Duron Apaloosa going to be PR rating or Mhz?

"When there's a will, there's a way."
March 7, 2002 9:54:45 PM

The speedgrade roadmap/chart I have seen on the Inquirer indicates a speed grade of PR2600 in Q3. However, there is no speed grade indicated beyond the 3rd quarter.

The chart can be seen here:
http://www.theinquirer.net/20010203.htm


When all else fails, throw your computer out the window!!!
March 7, 2002 11:32:14 PM

The main difference between the Claw and Sledge will be the memory controllers on the CPUs. Two CPUs can easily share cache and memory using fairly basic logic. Extending that to 4-8 CPUs is a whole different story.

You can compare this to what SUN has with one memory architecture for 1-2 CPUs, another for the 2-8 CPU volume servers and another for the 4-128 CPU highend servers.

It is a good call by AMD to make the Claws a two proc architecture from scratch as MP is the what is required for all the VR & AI software which will emerge when you have this kind of horsepower.

<font color=blue> Smoke me a Chip'er ... I'll be back in the Morgan </font color=blue> :eek: 
March 7, 2002 11:36:44 PM

Hmm now that you mention it, I predict Clawhammer costs will be like MP ones not like XP. A little more, but still awesome for performance.

--
For the first time, Hookers are hooked on Phonics!!
March 8, 2002 4:07:15 AM

Has AMD stated the clockspeed at which they intend to release the ClawHammer? I believe I have heard something about a PR3400. What clockspeed do you think that would be?

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
March 8, 2002 5:05:22 AM

The latest rumors believe it to be 2.2ghz or 2ghz or thereabouts.

But no official clockspeed has been released.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
March 8, 2002 3:52:10 PM

Yeah even if that was the clockspeed, it proves nevertheless that they have made something more powerful than anything at such clockspeed. It means you'd need P4 twice the speed to equal it. And since they have 0.13m plus a 2 stage pipeline increase (12 total), this means ramping up won't be that hard! This could signify tremendous problems for Intel in competing. Why? Since it does so much per clock, it's like adding 50MHZ increments instead of the AXPs 66MHZ to equal a NW 533MHZ increment of 100-200MHZ. So if AMD wants to play rough, they'll bump speed by 100MHZ increments or 66MHZ, and the PR can jump as high as 500 points per step. They had rumoured to get a PR4400 by Q1 03, when PR3400 Q4 02 was just around the corner. I don't see any Intel 4.4GHZ or more ready to compete.
This all depends on how AMD wants to play and how their Hammer will start, but if they said on their roadmaps, PR3400, it means tremendous performance for the clock it has and that it will compete big time on Intel's Prescott 800MHZ FSB.

Ray's with Intel so it's obvious he won't beleive it directly, neither from AMD and he claims he wants benchs to prove, but I think we all know how Hammer will do.

--
For the first time, Hookers are hooked on Phonics!!
March 8, 2002 4:29:33 PM

Quote:
They had rumoured to get a PR4400 by Q1 03, when PR3400 Q4 02 was just around the corner. I don't see any Intel 4.4GHZ or more ready to compete.

1) Actually, at the last Intel Developer Forum, Intel demoed a P4 already running at 4 GHz. Admittedly, this was probably done through some severe water cooling. However that still doesn't detract from the point that if they already can run the P4 at 4GHz, then it wouldn't take much effort for them to produce a product to compete with a PR3400. (Especially with the Northwood's 4x133MHz FSB ramp already planned to come out months before then.)

2) Prescott (the next-gen P4 @ .09microns) is slated to come out in 2003, leaving not much time for ClawHammer to make it's impact before Intel has it's next core. And it wouldn't surprise me at all if Intel released Prescott early if the ClawHammer was in any way threatening.

<pre><font color=green>//error-proof coding</font color=green>
<font color=blue>void</font color=blue> main(){<font color=blue>return</font color=blue>;}</pre><p>
March 8, 2002 4:31:01 PM

Quote:
I think we all know how Hammer will do

No, we do not. Until we see some real benchmarks from a 100% stable platform, we actually have no idea how it will perform. So far all we have are promises from AMD. Do you think any company would ever refrain from over-hyping its products?

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
March 8, 2002 4:44:04 PM

Quote:
Do you think any company would ever refrain from over-hyping its products?

Some do because they simply don't have the resources to hype their products. It is extremely rare however.

Just as interesting of a question though is if you change 'would' to 'should'. Do you think any company should ever refrain from over-hyping their products?

<pre><font color=green>//error-proof coding</font color=green>
<font color=blue>void</font color=blue> main(){<font color=blue>return</font color=blue>;}</pre><p>
March 8, 2002 5:02:29 PM

AMD has a tendency of delivering all their hype. So did Intel once, until they released the Willamette. The Northwood did however bring out most of the hyped performance.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
March 8, 2002 5:04:36 PM

The definition of 'over-hyping' is really subjective based on just how good the company believes its product to be. I believe a high self-esteem is very healthy, especially in a company. Without it you tend to fail. While 'over-hyping' may not be the best thing, there would be a struggle as to who thinks it really is too much hype. Who makes the call? It is best just to let those who will hype do so, and to realize it is going to happen without swallowing everything hook, line, and sinker.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
March 8, 2002 6:32:23 PM

I pretty much agree. Though I think that Intel has been overshooting a bit for a while now. It's nice to know the future plans of a company, but when you get so sick of hearing about something and it won't even come out for another six months, it makes one wonder if the hype could be toned down a litte. :)  In general though, I don't think hyping products is a bad thing. In fact, companies like AMD could probably stand to hype quite a bit more.

<pre><font color=green>//error-proof coding</font color=green>
<font color=blue>void</font color=blue> main(){<font color=blue>return</font color=blue>;}</pre><p>
March 8, 2002 7:51:03 PM

Quote:
it proves nevertheless that they have made something more powerful than anything at such clockspeed.


What exactly proves this? I must've missed an important post.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
Don't step in the sarcasm!
March 8, 2002 11:51:27 PM

Ray wondered at what starting clockspeed will the Hammer begin.
I told him regardless if it began at 1.5GHZ or 2.3GHZ, the tremendous work per clock will compensate for that, hence PR3400 will stand out no matter what, and will be able to compete against a ferocious 4GHZ from Intel.


--
For the first time, Hookers are hooked on Phonics!!
March 8, 2002 11:55:50 PM

Overhype?
Each time AMD overhypes us, they eventually come with something that does live up to the hype.
Last time Intel hyped us, was the P4, which when I read the sheets, they were awesome enough to blow me away, but in the end it was so disappointing I could not look at Intel for a while. Slvr you already explained much about the die size and all, thank you very much! But that will never make up for the fact Intel hypes us too much and doesn't live up to it, like AMD did.
The Palomino continued to give even more performance, with even less heat drawing from it. That was more than we could ever ask! The SSE made it jump higher than ever in multimedia, and we can easily see that. The only thing now that is needed, is more clockspeed, so it can continue ramping up until the Hammer goes. If it had the speed potential as a NW, it would easily tramp it. But alas, Intel is in the green now.

--
For the first time, Hookers are hooked on Phonics!!
March 9, 2002 3:58:42 AM

If anything, AMD NEEDS to do some more advertising and overhyping. If the market saw real benchmarking and not fancy catch phrases like "cutting edge" or "the power of an Intel PentiumIV", AMD would have a bigger share of the market. The only people who are disappointed are us, the hardcore overclockers, techie experts, and better informed. I don't think AMD needs to worry about overhyping their products, they already have a substansial share which threatends Intel. Biased, me no. But Intel (Whom I admired before Rambus and the PIV I820 and I840) has lost customers and faith with poor marketing decisions. Who knows, maybe one day, AMD and Intel will switch places.

"When there's a will, there's a way."
March 9, 2002 4:51:11 AM

I'm still missing what proves that the Hammer will have a higher IPC than any other processor ever made.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
Don't step in the sarcasm!
March 9, 2002 6:59:56 AM

You sure? sorry took a little while to respond, but the last tech docs I was looking at had one of the features of the sledgehammer with two cores. Maybe it won't be the first sledgehammer released, but it was a definitally a feature. AMD did licence that technology from IBM and I don't really seeing AMD just not using it.
In the tech docs, it was showing how AMD was going to keep the speed up 32 bit coding and 64 bit coding fast. Both cores would have 32 bit and 64 bit logic in them, but maybe this proved to be too difficult at the time to do, so we have this sledgehammer. Or maybe its still in the works, but being kept quiet......:)  I'll look around when I have time to see what happened to that part of the hammer

MeldarthX
March 9, 2002 7:05:25 AM

That wasn't water cooler.....that was vaper cooling....the cpu was frozen down to -50 degrees to get it to run at that speed. But you also have to take it with a grain of salt because it only had the speed grade sitting up their and nothing else. Not even bouncing boxes....... :smile:

MeldarthX
March 9, 2002 7:13:16 AM

nothing... yet. watching the bouncing ball program on a A0 core revision running underclocked aint the proof i like :smile:

P.S. my athlon 1200C is overhyped to 1400 :wink:

I love helping people in Toms Forums... It reinforces my intellectual superiority! :smile:
March 9, 2002 7:14:44 AM

Big mack

IA 64 running at half the speed of hammer and still king

cheap, cheap. Think cheap, and you'll always be cheap.AMD version of semi conducteur industrie
March 9, 2002 7:48:34 AM

Quote:
So far all we have are promises from AMD. Do you think any company would ever refrain from over-hyping its products?

AMD worked very hard not to overhype the AthlonXP with its PR rating. As we've seen from benchmarks, the PR rating accurately reflects an AthlonXP's performance in relation to a Northwood's MHz. It's even rather conservative in favor of the Northwood.

I think it's a pretty safe bet that if AMD releases a 3400+ processor, it will meet or exceed the performance of a 3.4GHz P4. Considering that Intel barely has a 3GHz P4 scheduled for release at the time of the Hammer 3400+, AMD will give Intel another run for its money.

<i>If a server crashes in a server farm and no one pings it, does it still cost four figures to fix?
March 9, 2002 12:02:31 PM

While Eden didn't specify X86 processors when making his comment on fastest IPC that's probably what he meant. And yes, it looks like Hammer may very well be the highest IPC X86 production processor to date, if AMD's statements thus far are to be believed and I see no reason why they shouldn't...unless AMD is planning to do to Hammer what Intel did to P4 prior to release.

Mark-

When all else fails, throw your computer out the window!!!
March 9, 2002 12:13:31 PM

Quote:


IA 64 running at half the speed of hammer and still king

The Hammer and the current Itanium won't compete for the same market because they won't be compatible. In that sense, you can't compare them. After all, the high clock speed is there for it to perform well in 32-bits.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
March 9, 2002 1:33:10 PM

Maybe juin didn't put the two names at their place and means Hammer running at half speed...lol...
Talk about a non-informed troll!

--
For the first time, Hookers are hooked on Phonics!!
March 9, 2002 2:34:04 PM

Claw/SledgeHammer will KILL mackingly in integer preformance with esstimated 1300 Spec int score for a 2Ghz Sledge and a esstimated 1000 Spec int score for a 1.2Ghz itanium2... at around one 4th the cost and half the power consumption...

Integer preformance is what matters for Servers/Workstations... so dont excpect itanium to start selling there...

though the Itanium will be better at FPU preformance... at 4times the cost and 2times the power and heat...

and eden - the Alpha EV7 is actually alot faster per-clock then the Hammer... and so are other RISC microprocessors (power 4, ultra sparc etc...)

This post is best viewed with common sense enabled
!