Sign in with
Sign up | Sign in
Your question

AMD's Official Response to Intel's HT Technology

Last response: in CPUs
Share
June 20, 2003 11:31:57 PM

<A HREF="http://www.amdmb.com/news-display.php?NewsID=6479" target="_new"> Click to read</A>

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new"> My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new"> My Rig</A></b>
June 20, 2003 11:53:15 PM

Interesting! But is it true, and what exactly is a "industri standard test"? And what about gaming? Everything we se, well atleast most tests, including THG, reports that the Intel CPUs with HT is slightly faster than the AMDs. But then again, most Intel CPUs with HT have a FSB of 800 which AMD is still to catch up with.

Anyhow I find that it is difficult to read something useful from CPU comparison tests, because the different softwares used are generally optimized for either Intel or AMD. Maybe the tests should be done with three (who has the time or the resources?) different versions of the software? One nonoptimized version, and one intel optimized and one AMD optimized. Perhaps its not that difficult to make benchmarking software in three different versions?

But then again, the OS the prg are running on is perhaps more ore less suited for one brand (may it be Intel or AMD). And how well does benchmarking programs reflect the software we actually buy?

Well, my conclusion is, try to figure out which software one is going to use, make a budget, and get the platform that gives you the best ratio/invested$.

Anyway those are my thoughts on the subject.

Turk
June 21, 2003 1:23:54 AM

This is a response to HT and not to the 200 fsb chips. What I find most interesting is the list of benchies at the bottom of the page. They seem to be using a lot of Bapco tests that predate Bapco's dumping progs where intel did not outperform Amd. Kind of fitting dont you think? These are the same tests that showed the willy P4 1800 on par with the palomino xp1800. Just the same, I dont think any of the Athlons quite measure up to the 200 fsb P4 3 gig. at stock clocks.
Related resources
June 21, 2003 1:24:57 AM

Any statements released by an official company, be it Intel, Amd, Nvidia, ATI, whatever, one has to take with a grain of salt. The company will always try to make themselves look better, while downplaying the other companies products (in this case, hyperthreading). Amd says something like their processors are 23% faster or something? Are they making a fair comparison? (xp3200 vs 3.0ghz, compared to xp3200 vs 2.4ghz). Anyways, just my thoughts
June 21, 2003 1:46:01 AM

If you want to comment on an article sarge, I think it's a good idea to read it first. The comparison is stated as being between an xp3000+ and a 3.06 P4. Read the fine print at the bottom and it will tell you what benchies were used. There is no doubt that the tests can be reproduced though there is a question of the benchies being dated.
June 21, 2003 2:26:58 PM

----------------------------------------------------------
There is no doubt that the tests can be reproduced though

there is a question of the benchies being dated.
----------------------------------------------------------

Exactly, which brings us (me?) back to my point of more or less optimized software. One could even argue for a test of CPUs in a nonoptimized elementry OS. Then again, this would also be missleading since this wouldn't accurately reflect "actual" conditions.

Anyone seen any CPU benchmarks in Linux lately?

Turk
June 21, 2003 2:40:49 PM

Here's <i>my</i> response to AMD's Official Response to Intel's HT Technology:

:lol:  heheh :lol:  heheheh :lol:  heheheheh

You weren't expecting a serious response, were you?
June 21, 2003 3:09:21 PM

Just watch how in a few months AMD will have a public annoucement with all the joy: We are proud to announce HyperThreading technology is coming into future K8 cores!

That is typical nVidia behavior lately as well.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 21, 2003 3:23:07 PM

AMD has benchmarks that show not just the 3200+, but also the 3000+ as being significantly faster than the P4 3.0, so definately take this "response" to HT with a grain of salt.
June 21, 2003 7:10:04 PM

My Trust on AMD and nVidia is going down day by day. But then, who knows who else is cheating.

<b><font color=red>The statement below is true.</font color=red></b>
<b><font color=blue> The statement above is false.</font color=blue></b>
June 21, 2003 11:17:14 PM

AMD is cheating now?

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 22, 2003 2:23:40 AM

Their benchmarks are sort of cheating, because the information presented in them is intentionally misleading. A customer looks at the AMD benchmarks and thinks a 3200+ easily beats a P4 3.0.
June 22, 2003 3:37:39 AM

Say it isn't so! A manufacturer artificially inflating and biasing benchmarks to make their products seem more favorable?! No! It can't be! Only the Evil Intel (tm) does that!

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
June 22, 2003 8:11:06 AM

One thing you must all remeber, this is the end of the line for the SOCKET 462 processors. Hmm, jeez, come to think about it, that socket has been around for how long? How many YEARS? How long did it take Intel to finally reach AMD? Or even pass them? But if we keep argueing the 'AXP vs. P4HT' debate forever, why don't we start arguing about the P2 and the K6-2?

No, where the arguemnets should be taking place is whatever Intel has on the line for the future (what is it?) and the Hammer. Weathewr you think it or not, I firmly believe the Hammer is gonna whoop any current P4 out of the water. But that would be like the old P3's vs. Athlon or something, so we need to look at what Intel has to offer. And to tell you the truth, I don't know, because I forgot to care.

Yeah, I come off sounding like an AMD fanboy, but it is pretty darn close. I have never personally been felt like I was screwed by AMD, however I can say I was screwed on more than 1 occasion by Intel, which I will not go into because I don't want to spend another 5 minutes on this post. I stick with who I 'trust.'
I will never like Chevys, I've had 2 of them, nothing but problems and leaks. My father has had 3 Fords in the last 17 years, and not one leak, not one major problem (only normal wear things like U-Joints and Wheel-Bearings). Same goes for MOPAR (Dodge, Chrysler, Plymouth [screw Jeeps]) in my experience. So with that experience, I will say the next truck I get will be a Dodge or a Ford.

I too once had money, then I got into computers...
<A HREF="http://service.futuremark.com/compare?2k1=6607283" target="_new"> 3DMark </A>
<A HREF="http://www.anandtech.com/mysystemrig.html?id=24383" target="_new"> My Rig </A>
June 22, 2003 8:54:29 AM

Haha...


They used a 3000+ with A7N8X nforce 2 mobo+ 2X256MB Corsair CAS2,400MHz while the Intel system had a 3,06Ghz P4 + Intel D845GEBV2 the RAM running at 333Mhz.

That shouldtell us everything about the quality of the benchmarks.Add to this the 2-3 years old software they used.

Long live Intel!

From the darkside...you know!
June 22, 2003 1:10:59 PM

Quote:
I don't know, because I forgot to care.

Quote:
I don't want to spend another 5 minutes on this post.

So I guess you'll understand me if I don't spend more than 30 seconds writing a reply. :smile:
June 22, 2003 2:09:54 PM

It is not cheating or lying if it is possible!

Cheating would be if they used a software that would, say, like in the graphics cards, render a scene with less image quality to raise performance.
I just woke up, my head has little processing power to find a similar analogy.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 22, 2003 2:10:53 PM

Hammer did not kick any P4 out of the water.
Opteron lost 90% or more of the Workstation benchmarks to Xeon P4s.
Prescott WILL contend the Athlon 64 very well.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 22, 2003 3:40:17 PM

I must agree. I think Prescott is enough to keep A64 at bay. It should run 32-bit code excellently. Come on, it has a full meg of cache, completely redesigned core, extra instructions (SSE3 or something), improved HT functionality, and good chipsets that support it (right NOW). Plus, if it is indeed 90nm tech, then it can easily go up to 4Ghz and beyond. And that's really not something the A64 can compete with. My thoughts here.

You see, Skligmund, this is what you were forgetting to care. Some roadmaps show that, by 2004, we'll be seeing the grantsdale chipset, which tops out at a DC DDR-II 533 and 1066Mhz FSB, depending on availability of DDR-II and competition from AMD. Even when grantsdale isn't around, a ~3.4 or 3.6Ghz Prescott isn't a sloppy contender! And AMD hasn't got this kind of roadmap for 2004 either.

Now picture this: AMD having troubles with wide support for x86-64 code, without which its new A64 is just a "glorified and overpriced XP" (thanks for that quote, slvr_phoenix), and that probably will be the case for the rest of 2003, with 2004 looking a bit better. Intel is then ready to go 1066Mhz and DDR-II 533 by next year, if needed. Plus, Prescott can scale up to 4+ Ghz. And I'm ignoring the possibility that Yamhill (secret 64-bit instructions in Prescott, as a last resort) is actually a true rumor. Are you still so sure that A64 can kick the new Prescott P4's *ss, for the better part of 2003 and 2004? I wouldn't be. What do you guys think?

Maybe I'm downplaying AMD's moves here, but that's just how I see things.
June 22, 2003 7:15:39 PM

Quote:
That is typical nVidia behavior lately as well.

I'd like to now coin a term.

<b>nvidiate</b>, v.
<b>1.</b> to stretch the truth or make a false claim
<b>2.</b> to make ones' self appear to be better than someone else even though one is not
<b>3.</b> to skew test results in one's favor

I think AMD likes to nvidiate when comparing their processors to Intel's.

Okay, brain. You don't like me, and I don't like you, but let's get through this thing and then I can continue killing you with beer. -- Homer Simpson.
June 22, 2003 7:40:35 PM

Quote:
AMD is cheating now?

not in the exact sense. but its kinda sad to see the latest PR numbers with their dropped performance. I am just a little upset that my fav processor is no longer the best. :smile:



<b><font color=red>The statement below is true.</font color=red></b>
<b><font color=blue> The statement above is false.</font color=blue></b>
June 22, 2003 11:21:46 PM

Oh my I agree.
But it still isn't cheating really, there is no play of data so that it actually screws something up instead of helping (like the image quality drop for more performance, in graphics cards). Actually I dunno what cheating is, in CPU world! (use of SSE2 when prohibited?)

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 22, 2003 11:23:37 PM

Very nice new verb, methinks we should use it! (as a sig or promote it!)

However add one more which I initially meant:
<b>4.</b>Demean a current feature then later on appear all joyful promoting it.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 23, 2003 12:17:05 AM

Well, I would think "cheating" would be secretly using a patch that the consumers will never get their hands on (or most consumers won't get their hands on) that'll boost your performance but never mentioning it. If you went out of your way to point out that you were using a patch and even provided a link, that'd be fine I guess. Any modification of software without mentioning it would pretty much constitute "cheating" and I think both AMD and Intel in their "benchmarks" have done this in some way or another. I wonder if AMD used that patch to enable SSE on the Athlons in that Sysmark benchmark. There's no mention of it in that presentation. How many think they didn't?

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
June 23, 2003 4:15:53 AM

Ahh I see and yes, that was the thing I was looking for in CPU cheating.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:
June 23, 2003 8:32:27 AM

Quote:
Now picture this: AMD having troubles with wide support for x86-64 code, without which its new A64 is just a "glorified and overpriced XP" (thanks for that quote, slvr_phoenix), and that probably will be the case for the rest of 2003, with 2004 looking a bit better. Intel is then ready to go 1066Mhz and DDR-II 533 by next year, if needed. Plus, Prescott can scale up to 4+ Ghz. And I'm ignoring the possibility that Yamhill (secret 64-bit instructions in Prescott, as a last resort) is actually a true rumor. Are you still so sure that A64 can kick the new Prescott P4's *ss, for the better part of 2003 and 2004? I wouldn't be. What do you guys think?

Correct, 64-bit won't matter this year. If AMD wants to beat Intel, their Athlon64 needs to be faster than any Pentium in 32-bit! If they don't achieve it than their 64-bit support can be as fast as they want but the Athlon64 will be failure for sure.

The Yamhill technology is indeed no longer a rumour. It has been confirmed that the Prescott core has been enhanced to support 64-bit like the Athlon64s but using a different architecture (not AMD64). But the Yamhill support won't be enabled on Prescotts, it will arrive later after a core change...

<A HREF="http://www.chip-architect.com/news/2003_04_20_Looking_a..." target="_new">chip-architect</A>

Concerning AMD's statement that their CPUs are faster than Intel's with HT support is as pathetic as it gets. Using outdated benchies for an obsolete rating system and then trying to convince the world that HT is something like a simple clockspeed increase shows that AMD is really struggling for the time being. HT can be extremely effective when the software is optimized for it and there is no denial possible that HT's importance and need in the CPU sector will grow every day. (It's comparable to hardware T&L invention, most people don't know that it exists because every graphiccard supports it, but without hw T&L, games like HL2 and Doom3 would come out in 2008 and not in 2003/2004.)
June 23, 2003 8:40:17 AM

Yeah... we all did... Amd was the Angel, come to duel the mighty Intel devil. And they did won, a time back, but lately AMD's done the same thing we all loathed Intel for doing, playing the users for fools. AMD's been stringing along their fans by making up funky numbers on their processors trying to ramp up sales by stating that a 3200+ processor is equivalent to an intel P4 3.2ghz, and for the unknowing user, they think it for real. Where as the real world benchmarks show the xp3200+ just above the 2.6ghz and mostly beneath the 2.8ghz P4. AMD have rested too long, going with the stream. Besides someone earlier stated that they've never been disapointed over an AMD system, but with an Intel. The reverse is for me, but I own 2 AMD systems none the less, I don't care who's worse than who. The fact is that AMD is no worse than Intel and vica versa. AMD will do the same "bad" things as Intel does, unfortunately it makes us unable to trust any of them. Regardin the Opteron numbers, it's not quite the good stuff that everyone were told it was, same with Itanium. When are they going to learn? Never, never never never never. As long as there's only 2 contenstants for the crown, there'll come heaven and hell before either one behaves. But I just wish AMD would release a new processor that could compete with Intel's P4. In Norway where I'm from, AMD is more expensive than Intel ... Go figure, less performance for more money?!?!?!?.. Makes no sense, but it used to be that way with Intel, now it's AMD. And the charts that spurred this discussion is the same that Intel released on it's P3 launch, when AMD was shuffeling K7 processors. Nothings changed, nothing is going to change. Damn, we need Cyrix back up and running (not that their current cpu's is in any way good enough to compete with either AMD or Intel) with a revolutionary new CPU, or better even a completely new player in the market..... But I guess the costs of doing such a feat is not worth it, or?...

HRD
June 24, 2003 3:29:29 AM

Not from you when it comes from AMD. I know how your a big fan of AMD PR right? LOL{cool}
June 24, 2003 1:24:33 PM

Quote:

... and good chipsets that support it (right NOW). (Mephistopheles, 06/22/03)

I would tend to agree with these comments, specifically.

That is the primary thing that always irked me about Intel vs AMD: memory and HDD throughput. Intel chipsets easily acheive ~25% to 75% (in Sandra. 24%-63% in PC Mark 2002) better memory performance and a somewhat less stellar percent increase in HDD performance using standard, on-board IDE controllers.

I cannot think of a logical reason for that discrepancy, other than legal constraints, such as pantents. Otherwise, it means that Intel's chipset designers and engineers are better than anyone elses.

Also:
Quote:

AMD's been stringing along their fans by making up funky numbers on their processors trying to ramp up sales by stating that a 3200+ processor is equivalent to an intel P4 3.2ghz.... (hasse, 06/23/03)


If Intel chipsets performed as badly, or, conversly, AMD chipsets performed as well as Intel's, perhaps a 3200+ WOULD come a lot closer to the peformance of a P4 3.2

Regards,

Dave

Intel giveth and Microsoft taketh away.<P ID="edit"><FONT SIZE=-1><EM>Edited by dwellman on 06/24/03 09:35 AM.</EM></FONT></P>
June 24, 2003 1:46:11 PM

This conversation is going great so far. :)  I don't think I even have anything to add.

:o 

Well ... other than one minor suggestion. ;) 

Black_Cat, nvidiate was great. :)  But in light of <A HREF="http://www.amdzone.com/articleview.cfm?articleid=1296" target="_new">this</A> I think that the term might have to be re-coined something like <i>Appleate</i>. (Not to be confused with appellate, which is something entirely different. Heh heh.) I don't think that any company has gone to such extreme lengths to stretch the truth than Apple. I'm amazed that to this day they haven't been sued into nonexistence by both their competition and their customers.

"<i>Yeah, if you treat them like equals, it'll only encourage them to think they <b>ARE</b> your equals.</i>" - Thief from <A HREF="http://www.nuklearpower.com/daily.php?date=030603" target="_new">8-Bit Theater</A>
June 24, 2003 5:48:47 PM

Quote:
is godly in FPU calc's

Exactly. I am very interested in the new Itanium (not for me, for the physics institute in which I work), which will be out within 6 days. Hopefully, we'll see something on that. The current Itanium (1Ghz) is godly in FP ops, so what about a 1.5Ghz Itanium with more cache and revised architecture? :cool:
June 24, 2003 8:41:44 PM

Quote:

FastStream64 Technology
VIA FastStream64™ Technology uses an expanded array of prefetch buffers to reduce latency in the memory controller, extending performance at a lower transistor cost than dual channel implementations. With a 266/333/400MHz processor bus and supporting DDR266/333/400 memory, the VIA Apollo KT600 provides lightning fast access to system memory and can reach a peak bandwidth of 3.2GB/sec, enabling the full performance potential of the fastest AMD Athlon™ XP processors to be harnessed.

If only it were true.

<A HREF="http://www.via.com.tw/en/apollo/kt600.jsp" target="_new">http://www.via.com.tw/en/apollo/kt600.jsp&lt;/A>


Intel giveth and Microsoft taketh away.
!