News on Hammer

Toggle sidebar Toggle sidebar
K

Kemche

Distinguished
Oct 5, 2001
284
0
18,780
<A HREF="http://www.xbitlabs.com/news/story.html?id=1017377833" target="_new">X-bit</A>

Some interesting information here, nothing new other then there will be only 15% preformance boost for 32bit code per clock. I thought it was suppose to be 25%. If Clawhammer only gives 15% preformance boost then it's just not good because we know that Intel Compiler can give that much boost vs MS Compiler on P4 platform. 10% improvement on 64bit when recompiled for 64bit. But if you want to recompile your code you can also easily re-compile for HyperThreading and get more preformance out out P4. I know this is all speculation on P4 with HT enabled. But still AMD's number don't look that impressive. If you look at Current P4 and Athlon they are running neck and neck but when Intel releases 533Mhz FSB version of P4 it should give them 8-10% preformance improvement. If these numbers for Clawhammer are true then are we just waiting for something which will disappoint us?

KG

<b>"Hey! It compiles! Ship it!"</b>
 
This is certainly disappointing news, if true.

The Hammer is right now in an unenviable position, as it's been hyped up way too much. Most people think that Hammer will crush the P4 and Intel completely, usher in a new age of computing, and segregate between those with a Hammer and those without, with only those owning a Hammer able to attract a mate. Ok, I obviously made up the last part, but you see what I'm saying. 15% better per clock and running at 2GHz (going from the article) will be nice, but not enough to satisfy people that have put unrealistically high expectations on Hammer.

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft
 
Hmm, they also said the P4 will outperform anything in it's class before the Willamette was released. I think I'd rather wait and see. I mean, how can they possibly know unless they've done tests on the Hammer themselves? Even then, they're not using the final Hammer.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
 
Well I'll have to wait again from AMD. Remember, you also said it, you won't take anything seriously until real specifications from AMD and reviews. (possibly)
Personally we'd have to review what has been done of Hammer so far. Is it at full speed? Has AMD finalized all specs? Maybe their chipset is a low performer compared to what VIA might put up? Maybe the mem controller is not fully ready without real PC2700 CAS2 Ram?

Again I don't beleive it stops here at all.

--
For the first time, Hookers are hooked on Phonics!!
 
You gotta remember though, this isn't a release or final version of the hammer they are testing. They can still eke out more performance by changing the architechture. Doesn't really matter what the "numbers" are at right now, I think most of us are more intersted in the real Claw/Sledge hammer at (hopefully) the end of this year. No one can really say the benefits/advantages of 64Bit yet...Look at the Itanium...unaccepted by enthusiasts and realistic users alike, it is a major flop by Intel. Gotta wait till things fall into place first.

"When there's a will, there's a way."
 
it maybe only 15% faster per clock, but i assume that is only the initial model, further models will be tweaked and tuned, also, one of the advantages of the hammer is it has be designed to run at much higher clock speeds, remember that the athlon couldnt be infinitely increased in speed because it is limited by its architecture.

I need a 1.5 Ghz Athlon + 512mb ddr ram to write emails......honestly
 
Yeah but in this case the Hammer DOES outperform anything in its class.
It remains to see what else to tweak, check my post above to see why this isn't fully the Hammer.

--
For the first time, Hookers are hooked on Phonics!!
 
Here is some more stuff...If this is true then...

Hammer @ 2000Mhz = 2000Mhz Athlon * 15% = 2300Mhz Athlon XP Preformance

If you apply the current formula...67Mhz = 100PR then...

Current available Athlon XP's
XP 1500+ = 1333 Mhz
XP 1600+ = 1400 Mhz
XP 1700+ = 1467 Mhz
XP 1800+ = 1533 Mhz
XP 1900+ = 1600 Mhz
XP 2000+ = 1667 Mhz
XP 2100+ = 1733 Mhz

Future version of XPs
XP 2100+ = 1733 Mhz
XP 2200+ = 1800 Mhz
XP 2300+ = 1867 Mhz
XP 2400+ = 1933 Mhz
XP 2500+ = 2000 Mhz
XP 2600+ = 2067 Mhz
XP 2700+ = 2133 Mhz
XP 2800+ = 2200 Mhz
XP 2900+ = 2266 Mhz
XP 3000+ = 2333 Mhz

So the Hammer @ 2000 Mhz would be able preform like P4 @ 3.0 Ghz. But P4 is going to get the 533Mhz FSB so Hammer would preform lower then P4. This doesn't make sense does it?

Now let's take the 10% preformance Gain for re-compile for 64bit. This will give hammer 25% preformance per clock vs. Athlon XP.

Hammer @ 2000Mhz = 2000Mhz Athlon * 25% = 2500Mhz Athlon XP Preformance - on 64bit apps.

XP 3000+ = 2333 Mhz
XP 3100+ = 2400 Mhz
XP 3200+ = 2467 Mhz
XP 3300+ = 2533 Mhz
XP 3400+ = 2600 Mhz

So hammer @ 2Ghz would preform like Athlon XP @ 2500Mhz which turns out to be like 3300+Mhz for re-compiled application. Re-compiled application for P4 get's more preformance then 10%.

KG

<b>"Hey! It compiles! Ship it!"</b>
 
Opps, I forgot to mention above for Hammer 3000+ PR will equal to ~2300 Mhz in Frequency. Wow the forumula to calculate PR will be totally different since there will be 700Mhz difference between PR and actual clock speed.

KG

<b>"Hey! It compiles! Ship it!"</b>
 
Intel themselves for one.
Then people asked them why it isn't, they started spreading FUD by changing subject.

--
For the first time, Hookers are hooked on Phonics!!
 
Do you know if these percentages that are being thrown around right now include the architectural improvements for the MB's, like hyper-transport?
 
I don't know but as for now all this is is just a speculation and nothing else. We won't know for sure until THG or AnandTech does the review.

KG

<b>"Hey! It compiles! Ship it!"</b>
 
When you said "They said the P4 would...", who did you mean by "they"?

EDIT:
We won't know for sure until THG or AnandTech does the review.
Click to expand...

Yeah, I can't wait for THG's review. Benchmarks showing Hammer up by 5%, and the conclusion is "AMD's new technology simply blew the P4 out of the water..."

<font color=blue>If you don't buy Windows, then the terrorists have already won!</font color=blue> - Microsoft<P ID="edit"><FONT SIZE=-1><EM>Edited by FatBurger on 03/29/02 12:44 PM.</EM></FONT></P>
 
Id take a 15% improvement. WHens the last time anything gave 15% over the previous generation?
Besides, thats initial numbers... it will ramp up.

I dont see a problem or disappointment either way

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.
 
WHens the last time anything gave 15% over the previous generation?
Click to expand...
When the Pentium 4 transitioned from the Wilamette to the Northwood...

-Raystonn



= The views stated herein are my personal views, and not necessarily the views of my employer. =
 
Exactly! There were only two times when newer generations processors didn't provide appreciable performance from older generations, the Pentium Pro, and the Pentium 4 Willamette. Otherwise, all other processor upgrades have significantly improved performance one way or another. Even the P3 Katami added SSE without hurting the performance of the processor in other places. Later, the Coppermine provided a huge performance boost!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
 
I don't know how X-Bit got the 15%/25% figure. AMD has given a significantly higher increase in their estimate. I would hope that AMD knew what their processor performance was by now, based on IPC.

AMD said that 32bit apps should see 20-25% performance boost and 64 bit apps should see an additional 10-15%. This is on average 10-15% higher than X0bit says, goiong by your above statements Kem.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>
 
While I wonder if AMD can execute fully and on time, so far it appears they are on schedule and on target, from what I have read. Of course, there's still 6 months to go and anything can happen.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>
 
I assume by previous generation he meant P2 to P3, K6 to K7. So your point doesn't make sense anymore. If going from the P3 to P4 from scratch gave less performance, then where's the boost? Now going from the Athlon to Hammer, new architecture, gives you 15%, that is friggin nice for an early stepping. But it's early, so just you wait!

--
For the first time, Hookers are hooked on Phonics!!
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom