Sign in with
Sign up | Sign in
Your question

p4 2.8 Lookin good.(official article thread here)

Last response: in CPUs
Share
August 26, 2002 5:30:40 AM

Comments on the new article should go here I suppose.


one glaring flaw so far.

GeForce 4 Ti 4600 (MSI MS-8872) Version: 2.00
Memory: 64 MB DDR-SDRAM
Memory Clock: 650 MHz
Chip Clock: 300 MHz


NIce 64 meg ti4600 there frank.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
<P ID="edit"><FONT SIZE=-1><EM>Edited by Matisaro on 08/25/02 11:01 PM.</EM></FONT></P>
August 26, 2002 6:00:18 AM

grasshappa
"jeez.... Intel loves waiting for the right times dont they... They knew...they knew 2600+ XP was coming out. They can afford to wait since they have the fastest chip out there. But not AMD, AMD needs to get out his chips to be a bit more competitive. and just as we thought AMD got the throne... BAM! INTEL STRIKES BACK! Although we knew the new intel's were coming out soon, it was obviously not coincidence that it came out just a few days after AMD's flagship cpu's...Intel saw, came, and conquered. But some of the AMD overclocked chips didnt do bad...but I feel bad for AMD for having such a hard time...It happens every time.
But after all, price/performance crown goes to AMD... but still.... Its gonna be very hard for AMD. IMO, the only way that AMD can take over is as soon as Intel delays something or makes a mistake...that will be the AMD's chance to truly take the crown (at least for a few months) OR AMD comes out with something UBER COOL. We will see

Comments? Feelings? Let it all out here brother...

real philosophy of life: "do onto others what you dont want them do onto you" "

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 6:00:45 AM

Grasshoppa the 2.8ghz release was planned for some time before amd releasaed the 2600+

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 6:17:38 AM

"Although we knew the new intel's were coming out soon, it was obviously not coincidence that it came out just a few days after AMD's flagship cpu's..."

matisaro, you love to fight and compete dont you? you even bothered writing "(official article thread here)" so people would write in your thread instead of mine. Lol thats completely cool with me. i dont care. btw, intel planned to release it around this time, but i dont think they specified the EXACT date. have fun, try not to get mad this time. and yes, your thread is THE OFFICIAL ARTICLE THREAD.

btw, i didnt write what u quoted me in this thread, i wrote it in my own thread, so why didnt you reply on my thread? did you want to find some excuse to write something so your post bumps above mine?....lol

real philosophy of life: "do onto others what you dont want them do onto you"
August 26, 2002 6:26:19 AM

I decided to support grassapa's thread.<P ID="edit"><FONT SIZE=-1><EM>Edited by castle on 08/26/02 02:35 AM.</EM></FONT></P>
August 26, 2002 6:52:17 AM

Quote:
matisaro, you love to fight and compete dont you? you even bothered writing "(official article thread here)" so people would write in your thread instead of mine. Lol thats completely cool with me. i dont care. btw, intel planned to release it around this time, but i dont think they specified the EXACT date. have fun, try not to get mad this time. and yes, your thread is THE OFFICIAL ARTICLE THREAD.


This thread was posted first, I saw your thread and moved your post here to prevent a splitting of discussion, I added the official thread title to ensure people didnt split up posting.

If your thread had been first and I posted, I would have deleted this thread and posted in your thread, sadly you do not show the same courtesy.

It has nothing to do with me being competitve, and you thinking that just shows how insecure you are.

This thread was first, yours was second, we dont need multiple threads on the same subject, live with it.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 6:54:12 AM

castle
Quote:
Maybe this is a little harsh, but: it was well known that Intel would announce 2.8G around or before Sept 1, but T-Bred 2600+ was a real surprise to everyone. It is therefore pretty clear that AMD tried to steal some thunder from Intel, not the other way around. It is also obvious to me that AMD paper launched 2600+ 5 days before a real announment of P4 2.8G so as not to be embarassed by Intel's 1G lead ahead itself: P4 2800- AXP 2200+ (1800M) =1G.



Look poeple we DONT NEED 2 THREADS ON THE SAME SUBJECT, I MOVED GRASSHAPPAS POST HERE BECAUSE THIS THREAD WAS THE FIRST MADE, IT HAS NOTHING TO DO WITH ME JESUS.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 7:04:26 AM

calm down matisaro....everything is all right. we have different opinions so we need 2 different threads. Its ok matisaro...i am such a bastard, i have no courtesy. bad bad bad, i should go confess

real philosophy of life: "do onto others what you dont want them do onto you"
August 26, 2002 7:11:01 AM

Quote:
calm down matisaro....everything is all right. we have different opinions so we need 2 different threads. Its ok matisaro...i am such a bastard, i have no courtesy. bad bad bad, i should go confess


Dude what the hell is your problem man, I made a thread for discussing the article, then later you made a copy thread, we have enough multiple thread problems in this forum as it is, so as service I moved your content to this thread to harbor fair discussion.

and you throw a hissy fit, check the times posted, your copycat post does not belong, im sorry but if it was the other way around I would delete my post, this has nothing to do with me, but its about the community!

YOu need to seriously calm down and stop being so confrontational.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 7:12:14 AM

Back on topic, the 2600+ was a huge surprise to everyone, the 2800 date has been out for months, I agree with castle that amd released the 2600 on purpose to mess with intel, not vice versa.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 7:15:08 AM

matisaro, i didnt copy your thread because i was writing it at the same time you were...except mine was longer and it took me longer to write. plus we dont talk about the same stuff. so they are different threads

mine was posted at 1:31 and yours was posted at 1:30. check it for yourself.

real philosophy of life: "do onto others what you dont want them do onto you"
August 26, 2002 7:28:30 AM

They were not different threads, both are commentary on the new article, if you check your thg history you will see fatburger began the tradition of creating a single official thread for each new article, this thread(which was genenrally the first posted) got to bear the title official.

mine was posted first, and your thread while having a different initial post is the same thread, which is why I helped out and moved all content to the first thread to start conversation.

we are getting off topic, just discuss the article, if you want to make a competing thread thats your right, you are only cluttering the forum though.


:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 7:32:27 AM

Grasshoppa

Quote:
you are right on the customer (which is most of us) side point of view. but you gotta know that sometimes release dates can change quickly. intel's chip was scheduled around or before sept 1. thats true, but they didnt specify EXACTly what day. that gives intel a little room to work for. Intel is not dumb, so they had their ways to know what day the 2600+ XP was going to released. IMO, i dont think AMD would dare steal thunder from intel b/c AMD knows that Intel can just beat them (in this case, 2800mhz). AMD has no other option but to release earlier. Remember, All this is my opinion so im not trying to start an argument. but because I have seen this (AMD vs Intel) (ATI vs NVIDIA) this kind of stuff happens a lot (where the powerful waits for the weak to release first) I concluded that Intel did the same again. Again, i might be wrong but you gotta admit that intel has done it before, and IT could be a possiblity that he did it again. We will never know what intel and amd knows about each company and "REAL" release dates. I dont think matisaro appreciated your retirement from his thread because his is the "ORIGINAL ARTICLE THREAD" but you made the right choice.


Good post, however the last bit was laughable, you are just trying to cause argument because you were so thoroughly schooled in our 8500 versus 4200 debate, thats fine.

Please dont let your anger at me personally disrupt the forum grasshoppa.


On topic, yes both companies make it a habit to release things to screw eachother up, but your assertion intel released the 2800 to screw up the 2600+ is wacky because the 2.8ghz release has been planned for some months on this date while the 2600+ was a complete surprise to everyone.(even intel I would imagine).

If anything, castles assterion amd released the 2600+ to screw with the 2.8gha launch is more valid and more logical.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 26, 2002 1:22:18 PM

Goddammit people!! :mad:  It's a thread for god's sake! Now instead of bashing for which is better, R8500 or Ti4200, now we have to bash which thread was first... blah, blah, blah... There were planty of threads for the same thing. Post whereever you want, but this is just getting childish. It's like kids playing on swings..."But I called them first"... Just stop. Turn this thread around into a normal comments on the P4 2.8GHz!

Well, what I think:
He still makes it sound as though the XP2600+ wallops the 2.53GHz, although his wording isn't as strong, therefore giving me the feeling that he's gotten commentary on his previous review of XP2600+. As we can see, the XP3400+ doesn't really even reach P4 2.8GHz levels sometimes, and rarely beats the 2.9-3GHz o/cs. Also, IIRC, the XP3400+ was 2.8GHz and not 2.66GHz.

<i>Past mistakes may make you look stupid, but avoiding future ones will make you look smart!</i>
August 26, 2002 1:45:56 PM

Quote:

Also, IIRC, the XP3400+ was 2.8GHz and not 2.66GHz.

No, the Athlon XP 3400+ would be 2.66GHz because 2.13GHz (2600+) + 66.66 * 8 (3400-2600=800) = 2.66GHz.

Anyway, there will not be a Athlon XP 3400+ T-Bred. The 3400+ will either be a Barton or a ClawHammer.

Anyway, I'm not surprised that I was right that Intel will need a bit more voltage at 2.8GHz and above:

Quote:

With the previous variants of the Pentium 4 (up to 2533 MHz), the core voltage remained constant at 1.500 Volt. It's a different picture with the new CPUs with higher clock speeds: the P4/2800 requires a voltage of 1.525 Volt - that's an increase of 0.025 Volt. This fact alone leads to the conclusion that Intel wants to automatically guarantee better results by increasing the core voltage by 1.6%. It shouldn't be a problem for the motherboards to adapt to this, since the voltage values are contained in the CPU as data, and this is automatically read and adjusted by the motherboard logic.


Remember when the Inquirer said that there were issues with the P4 2.8GHz+ under WinXP or something. Everyone thought it was a software problem, but I felt the P4 will just need slightly more voltage to remain stable at 2.8GHz or higher.

BTW, I just found out that the P4 2.8GHz isn't merely a faster P4. It has significantly higher IPC in some cases. This is from HardOCP:

Quote:

Editor's Note:
I love it when our readers know more than we do (which happens a lot). Although this has not been fully confirmed, the contributor is very much on the right track and this is surely information you will want to have if you are buying a Pentium 4 with hopes of overclocking it.

Several notes/comments on your article.

According to your article, there is nothing really new with the new P4 2.8GHz. There are a few other changes, as Aceshardware will report later in the day.

The 2.8 is the first P4 to use the new C1 stepping, as you can see with CPUID. Moreover, look at the L1 cache info under CPUID (and compare it to the older P4 processors), and you'll see that Intel doubled the number of entries in the instruction TLB. If I recall correctly, AMD did something like this with the move from the Thunderbird to the Palomino.

Some have said this change was made to aid in future versions of the 2.8 that may have Hyperthreading enabled, but apparently this may also improve performance for certain other types of operations. Certainly, this seems the case from a few of your benchmarks, where scaling exceeds the 2.8/2.53=10.5% that should be the theoretical maximum, given no changes other than increased clock.

In Serious Sam, you saw a 23+% improvement with the 2.8 over the 2.53; in Commanche 4, you saw a 11.5% improvement; in Jedi Knight II, you saw a 13% improvement. In each case, the improvement exceeded the 10.5% theoretical maximum assuming simply an increased clock. Other sites/reviewers also found a 16% improvement in TMPGEnc, a 25% improvement in SuperPI, and a 29% improvement in PovRay. You might want to make note of this in your review, for anyone buying a P4 will want to make sure they get the new C1 stepping.



Intelligence is not merely the wealth of knowledge but the sum of perception, wisdom, and knowledge.
August 26, 2002 3:39:15 PM

There are a bunch of whiney kids in both threads.

I'm not going to post disucssion material in either. Hope someone else starts a new thread.

Cooj

I am ignorant no more!
August 26, 2002 4:39:33 PM

Alrighty then. Here by my comments:
<b>1)</b> Thread redundancy is not exactly going to end the world.

<b>2)</b> If the threads were time stamped a whole one minute apart in a server that has at times taken (as far as I've seen) <b>up to seventeen minutes</b> to store a post, then for all intents and purposes they were started at the same time.

<b>3)</b> It annoys me that they <i>still</i> suggest that the AXP 2600+ beats out the P4 2.53 when in fact the <b>majority</b> of benchmarks showed otherwise. I still even stand by my opinion that the P4 2.4GHz with PC1066 is equal to an AXP 2600+ with DDR333. Sure, they each have their niches where they kick each other's arse, but they also level out quite equally against each other in the end, I think.

<b>4)</b> At least the simple fact that no one in their right mind could consider the P4 2.8GHz as losing to the AXP 2600+ mitigates #3.

<b>5)</b> grassapa, no offense, but you're clueless. Plain and simple. AMD would most definately try to steal thunder from Intel because frankly, it's the only thunder they really have at the moment. I guess to them, anything is better than nothing. To me though, it's just playing into the hands of fanboys and no one else.

It went like this: Intel planned on releasing the 2.8GHz for months. AMD <b>did</b> try to steal Intel's thunder with a pre-emptive strike using their 2600+. Their production wasn't up to an official release yet, so they did the generally disgusting thing of making a paper launch.

And it worked. The <A HREF="http://www.amdmb.com" target="_new">AMD fanboy posterchildren</A> still think that AMD is the best. However, for those who actually face reality, nothing changed. Intel is still kicking AMD's arse in performance.

AMD knew that those who had a grip on reality wouldn't be affected one way or the other. They also knew however that they could rile up their fanboys into a rabid seething mass, and so they did.

Powers and Presences only know why AMD is so intent on marketting specifically to their lapdogs instead of actually trying to, oh, compete. Still, that is AMD's current plan, and they're exceeding everyone's expectations at their skill in continuing the insanity.

<b>6)</b>Matisaro, the AXP 2600+ was <i>not</i> a huge surprise to everyone. There had been long speculation that AMD was having some sort of problems with their Thoroughbred core. Further, with the rather clock-speed-limited launches of the Thoroughbred and then subsequent lack of any more launches, there was further speculation by those who still liked AMD that a massive core revision was being developed. Most figured that core revision would actually first make it's appearance in Barton. Most figured that would happen just before the launch of the P4 3GHz to stun and amaze us. A rare few however hoped that AMD would somehow managed to surprise us all either just before or just after (depending on how good their new core's performance actually was to either beat the 2.53 or the 2.8) the official 2.8 launch.

Admittedly, I was one of the majority of these hopefuls that figured AMD would use a core-revision to release a Barton at 3000+ in the coming months. However, that still left the AXP 2600+ paper launch as virtually no surprise to me, nor to several of my friends.

<b>7)</b> The thought of a 3400+ Barton still leaves much in doubt though. Will AMD continue to use the same clockspeed scale? Or will they account for the performance increase that Barton's extra cache will bring? After all, the PR system is <i>supposedly</i> based on the performance to a T-Bird. So by this (admittedly very unbelievable) logic, then it's the overall performance that determines the rating, <i>not</i> the clockspeed. And thus the Barton <i>should</i> have a lower clock speed at the same PR of a Thoroughbred.

Should AMD neglect this simple logic, then it would undoubtedly disprove their whole ratings basis and give yet more proof that AMD is just after your money like everyone else and is more than happy to tell bald-faced lies to customers (that is the job of marketting after all) if it will make them feel better.

<b>8)</b>I've actually been impressed that Intel hasn't gone to yet another new socket yet. Even with a minor voltage increase, they're still staying put. It makes me wonder if the P4 really will remain in the same socket until Prescott.

<b>9)</b>The doubled L1 cache of the 2.8 is very interesting. It makes me wonder just how many advances Prescott really will have now if Intel is already slipping core revisions like this into Northwood. Before you know it, the P4 might actually have a better IPC than an Athlon. (Ha! Then again, maybe not. Can you believe that I even thought that for a second? What is the world coming to...)

<b>10)</b>I think it's time I went and checked the temperature readings from various places in Hell. I think it's a good time to start watching for a cold front moving in...

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 5:00:05 PM

hehe... u tell them pheonix!!!

oh yeah... and stop fightin u 2 (mati and grass). Mati, you do your own threads and let grass do his own threads. if he doesn't like you copying his threads, then don't do it again!!! It's as simple as that!!

The greatest risk of all is not taking one!
August 26, 2002 5:02:08 PM

Quote:

Alrighty then. Here by my comments:

Alrighty then. Here is my Rebuttal :tongue: :

Quote:

1) Thread redundancy is not exactly going to end the world.

EXACTLY!!

Quote:

2) If the threads were time stamped a whole one minute apart in a server that has at times taken (as far as I've seen) up to seventeen minutes to store a post, then for all intents and purposes they were started at the same time.

Possibly, but who cares who posted first?

[/quote]

The Athlon XP 2600+ is neck and neck with the Pentium 4 2.53GHz. Just because the P4 is ahead in a couple of games, and bandwidth hungry apps, doesn't mean it wins the majority of benchmarks. More apps exist out there that take advantage of the Athlon's FPU advantage than the P4's bandwidth advantage. The Athlon XP will constantly win in FPU intensive apps and lose in bandwidth intensive apps. Games are less important factors because at higher resolutions, the graphics card is almost always the bottleneck. It's really a matter of personal judgement really. Both processors excel but in different fields. You can't say the P4 2.53GHz is faster than the Athlon XP 2600+ in the same way you can't say that a physicist is smarter overall than a medical doctor. It's really up to you to decide which you prefer based on the benchmarks. Most sites look to the typical power user's use to compare each processors. Obviously many of these review sites view the Athlon XP 2600+ as the better processor for people who fit the "power user" or "gaming user" category. Of course, there's also the audiophile or videophile and business/office user, etc.

Quote:

4) At least the simple fact that no one in their right mind could consider the P4 2.8GHz as losing to the AXP 2600+ mitigates #3.

Can't argue with that.

Quote:

5) grassapa, no offense, but you're clueless. Plain and simple. AMD would most definately try to steal thunder from Intel because frankly, it's the only thunder they really have at the moment. I guess to them, anything is better than nothing. To me though, it's just playing into the hands of fanboys and no one else.

It went like this: Intel planned on releasing the 2.8GHz for months. AMD did try to steal Intel's thunder with a pre-emptive strike using their 2600+. Their production wasn't up to an official release yet, so they did the generally disgusting thing of making a paper launch.

And it worked. The AMD fanboy posterchildren still think that AMD is the best. However, for those who actually face reality, nothing changed. Intel is still kicking AMD's arse in performance.

AMD knew that those who had a grip on reality wouldn't be affected one way or the other. They also knew however that they could rile up their fanboys into a rabid seething mass, and so they did.

Powers and Presences only know why AMD is so intent on marketting specifically to their lapdogs instead of actually trying to, oh, compete. Still, that is AMD's current plan, and they're exceeding everyone's expectations at their skill in continuing the insanity.

AMD doesn't care if they don't always have the top of the line processor, they're close enough to continue to be considered producers of high-end processors.

Quote:

6)Matisaro, the AXP 2600+ was not a huge surprise to everyone. There had been long speculation that AMD was having some sort of problems with their Thoroughbred core. Further, with the rather clock-speed-limited launches of the Thoroughbred and then subsequent lack of any more launches, there was further speculation by those who still liked AMD that a massive core revision was being developed. Most figured that core revision would actually first make it's appearance in Barton. Most figured that would happen just before the launch of the P4 3GHz to stun and amaze us. A rare few however hoped that AMD would somehow managed to surprise us all either just before or just after (depending on how good their new core's performance actually was to either beat the 2.53 or the 2.8) the official 2.8 launch.

Admittedly, I was one of the majority of these hopefuls that figured AMD would use a core-revision to release a Barton at 3000+ in the coming months. However, that still left the AXP 2600+ paper launch as virtually no surprise to me, nor to several of my friends.

It didn't really surprise me either.

Quote:


7) The thought of a 3400+ Barton still leaves much in doubt though. Will AMD continue to use the same clockspeed scale? Or will they account for the performance increase that Barton's extra cache will bring? After all, the PR system is supposedly based on the performance to a T-Bird. So by this (admittedly very unbelievable) logic, then it's the overall performance that determines the rating, not the clockspeed. And thus the Barton should have a lower clock speed at the same PR of a Thoroughbred.

Should AMD neglect this simple logic, then it would undoubtedly disprove their whole ratings basis and give yet more proof that AMD is just after your money like everyone else and is more than happy to tell bald-faced lies to customers (that is the job of marketting after all) if it will make them feel better.

I disagree. The overclocked "Athlon XP 3400+" performs rather poorly for it's rating. AMD needs the extra 256KB cache to just maintain a decently accurate PR rating.

Quote:


8)I've actually been impressed that Intel hasn't gone to yet another new socket yet. Even with a minor voltage increase, they're still staying put. It makes me wonder if the P4 really will remain in the same socket until Prescott.

Keeping the same socket is rather useless if the 3GHz P4 might not be compatible with many (if not all) current motherboards.

Quote:

9)The doubled L1 cache of the 2.8 is very interesting. It makes me wonder just how many advances Prescott really will have now if Intel is already slipping core revisions like this into Northwood. Before you know it, the P4 might actually have a better IPC than an Athlon. (Ha! Then again, maybe not. Can you believe that I even thought that for a second? What is the world coming to...)

Hehe, it might just happen right before the launch of the Hammer with the introduction of Hyperthreading on the 3GHz P4. AMD is lossing IPC while Intel is gaining IPC, at the moment.

Quote:



10)I think it's time I went and checked the temperature readings from various places in Hell. I think it's a good time to start watching for a cold front moving in...


I'm sorry but I have no idea what you're saying here.








Intelligence is not merely the wealth of knowledge but the sum of perception, wisdom, and knowledge.
August 26, 2002 5:09:30 PM

first of all, you are not "OFFICIAL" of anything, why are you trying to fool the new members into thinking that?

btw, your AMDsuperfanboy (Van smith) is a wanker.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
August 26, 2002 5:29:33 PM

lol

<font color=blue>Unofficial Forum Cop</font color=blue>
August 26, 2002 5:41:09 PM

I'm impressed how the price of a P4 2.4 and 2.533 went up over $200 since the Athlon 2600+ article.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 26, 2002 5:53:49 PM

Quote:
3) It annoys me that they still suggest that the AXP 2600+ beats out the P4 2.53 when in fact the majority of benchmarks showed otherwise. I still even stand by my opinion that the P4 2.4GHz with PC1066 is equal to an AXP 2600+ with DDR333. Sure, they each have their niches where they kick each other's arse, but they also level out quite equally against each other in the end, I think.

Human nature prefers to belive what it wants to belive dispite the truth being out in front of it.

Quote:
Intel planned on releasing the 2.8GHz for months. AMD did try to steal Intel's thunder with a pre-emptive strike using their 2600+.

That's what I thought immediately at the release. I could point to the GPU industry where ATI and Nvidia do the same thing (Driver releases to preempt product releases, spec "leaks", ect). It shouldn't suprise anyone that it would happen in the CPU industry.

Quote:
Should AMD neglect this simple logic, then it would undoubtedly disprove their whole ratings basis and give yet more proof that AMD is just after your money like everyone else and is more than happy to tell bald-faced lies to customers (that is the job of marketting after all) if it will make them feel better.

What most people seem to forget in their AMD love fest is that AMD is in fact a company that is trying to make money. The PR rating is a publicity move to help to market their CPUs better, no matter what else they may say. I find it hard to believe they're doing it simply for the "Good of the Industry".

Quote:
9)The doubled L1 cache of the 2.8 is very interesting. It makes me wonder just how many advances Prescott really will have now if Intel is already slipping core revisions like this into Northwood. Before you know it, the P4 might actually have a better IPC than an Athlon.

If that happens, Prescot could simply be a second tier P4 release like the NW instead of the new P5. As for the IPC, I seriously doubt that simply because of the sheer wall they'd have to climb, but stranger things have happened.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 26, 2002 6:39:16 PM

Quote:
Possibly, but who cares who posted first?

Well, in theory it has meaning if you care about keeping the forum all neat and tidy. Since I've never seen that happen in <i>any</i> forum, THGC or otherwise, I don't think it has much meaning. Which is my point, that simply no one cares. :) 

Quote:
The Athlon XP 2600+ is neck and neck with the Pentium 4 2.53GHz. Just because the P4 is ahead in a couple of games, and bandwidth hungry apps, doesn't mean it wins the majority of benchmarks. More apps exist out there that take advantage of the Athlon's FPU advantage than the P4's bandwidth advantage. The Athlon XP will constantly win in FPU intensive apps and lose in bandwidth intensive apps.

I disagree. For every app that you can find an Athlon having an advantage from their FPU, you can find two apps that have the P4 showing an advantage for it's bandwidth and SSE2. That's been my experience so far anyway.

Moreover, the P4 2.4GHz actually scores a lot closer (yet usually still higher) to the AXP 2600+ than the P4 2.53 does. So my general opinion is that this is where the point of equality lies. Sure for some apps one or the other will do better, but I think that the 2.4GHz is the general point where they balance each other out, not the 2.53GHz.

At least when using PC1066. When using PC800 or DDR333, then it's the 2.53 if not even the 2.66.

Quote:
Obviously many of these review sites view the Athlon XP 2600+ as the better processor for people who fit the "power user" or "gaming user" category.

The funny thing is, the majority of 'recent' games would benefit more from a P4 than an Athlon. It's only really serious 3D rendering that's highly FPU intensive where the Athlon has a real shine. And even that is slipping these days.

Quote:
AMD doesn't care if they don't always have the top of the line processor, they're close enough to continue to be considered producers of high-end processors.

I agree in that AMD doesn't care if they <i>always</i> have the top of the line processor. However, they do push hard to jump slightly higher, even if only for a short time, and lately even if only on paper. It's the only way AMD can keep from losing it's fanboys who might drift off if someone could finally prove without a doubt that Intel is faster.

Quote:
I disagree. The overclocked "Athlon XP 3400+" performs rather poorly for it's rating. AMD needs the extra 256KB cache to just maintain a decently accurate PR rating.

**ROFL** That's a good point. The AXP does seem to rather be losing IPC lately instead of actually gaining on it like the P4 is. At this rate, AMD's AXPs might actually perform on par with an equally-clocked T-Bird soon. Wouldn't that be funny... ;) 

Quote:
Keeping the same socket is rather useless if the 3GHz P4 might not be compatible with many (if not all) current motherboards.

Tell that to AMD!

:o 

Heh heh. I couldn't resist poking that bit-o-fun at AMD, especially in light of their ever-changing heat sink requirements and old mobos that won't support new AMD chips either.

Seriously though, I think you underestimate the importance that it could have. If, say, I were able to purchase a new mobo that supported a 3+ GHz CPU, but could place, say, a Celeron 1.7GHz into it until I could actually <i>afford</i> a 3+ GHz CPU, then it has considerable value to me. :)  Especially if it were to also be able to support a Prescott CPU...

Quote:
Hehe, it might just happen right before the launch of the Hammer with the introduction of Hyperthreading on the 3GHz P4. AMD is lossing IPC while Intel is gaining IPC, at the moment.

While it is sad that it took Intel so long, it is nice to see Intel at least finally working seriously on their IPC. :) 

Quote:
In reply to:
--------------------------------------------------------------------------------

10)I think it's time I went and checked the temperature readings from various places in Hell. I think it's a good time to start watching for a cold front moving in...

--------------------------------------------------------------------------------

I'm sorry but I have no idea what you're saying here.

I'm saying that Hell might just be about to freeze over. Imagine a world where Intel has a better IPC than AMD, where AMD is manipulating benchmarks to show their CPUs in a better-than-realistic light, and where AMD forces consumers into a single memory type while Intel offers choice.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 6:41:20 PM

I do have to commend Intel. Unlike what AMD has done by generaly sitting on the core design and doing very little to it other than to reduce heat output to allow for higher clockspeeds, Intel has made marked improvements in their CPUs over the life of the core. The current P4 is so drasticly different from the origional P4, and it looks like it may continue to evolve.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 26, 2002 6:41:50 PM

The last time I used an Intel chip was 1994. Obviously for a few of those years I was only buying the AMD because it was cheaper and more than met my needs. Now I'm looking at what CPU i'll buy next and it's, rather obviously, a choice between the XP and the P4.
I still think i'll get the XP, simply because it's cheaper and I'm not needing anything like 2.8G. 2G would be enough for me. My choice has nothing to do with brand loyalty. Over the years I've had Intel, Cyrix and AMD processors. I think the biggest mistake I ever made was the Cyrix. Altho it was cheap as chips and I was only 14 at the time ;) 
Anyway, that's my 2 cents.
Really, everyone should forget about the performance and get into energy saving and buy a VIA...;)

AMD Is An Anagram Of MAD, Intel Is An Anagram Of INLET, Cyrix...Ah Who Cares?
August 26, 2002 7:07:21 PM

Quote:
Human nature prefers to belive what it wants to belive dispite the truth being out in front of it.

Perhaps it's just often we are limited to a point-of-view that prevents us from seeing the whole and thusly weighing it properly?

Quote:
That's what I thought immediately at the release. I could point to the GPU industry where ATI and Nvidia do the same thing (Driver releases to preempt product releases, spec "leaks", ect). It shouldn't suprise anyone that it would happen in the CPU industry.

I suppose not. To a lesser extent it had been going on with price cuts for a long time. This though, using a pure white-paper launch simultaniously with a <i>very</i> debatable and unsupportable benchmarking 'scoop' is all very ... dirty. Intel was mocked for this kind of stuff. Yet now we get to AMD doing it, and it's all just <i>'Well, so? Your point is...'</i>. It's really rather hypocritical of the AMD lapdogs to not be just as outraged (if not more so) in my opinion. After all, it is their precious AMD <i>stooping</i> to such tactics.

Ah well. Like I should ever have expected the world to <i>not</i> be hypocritical...

Quote:
What most people seem to forget in their AMD love fest is that AMD is in fact a company that is trying to make money. The PR rating is a publicity move to help to market their CPUs better, no matter what else they may say. I find it hard to believe they're doing it simply for the "Good of the Industry".

That's been my sentiment for years. Even then though, I had always given AMD a moderately high regard for holding an air of integrity even when they were not doing the best. Now though, I can't even give AMD that much credit. All they are now is just another cheap CPU manufacturer trying to make a buck, no matter how dirty they have to play to do so.

Quote:
If that happens, Prescot could simply be a second tier P4 release like the NW instead of the new P5. As for the IPC, I seriously doubt that simply because of the sheer wall they'd have to climb, but stranger things have happened.

I've always seen Prescott as likely to be a midpoint between either. I figure it'll just mostly be enhancements to the Northwood core. Significant enhancements, but still just enhancements. A seriously new core just doesn't seem likely to me.

After all, if Intel just concentrated on fixing the flaws in the P4 core that keep it's IPC down so low (such as the P4's bitshifting, the single and highly overworked FPU, the limited L1 caches) they could massively improve the IPC with relatively simple core improvements. Another FPU might take a lot of core space, but then the shrink to 0.09 microns <i>should</i> make it possible, <i>especially</i> with another layer or two over Northwood.

Then make sure that the CPU has enough of each logic unit and cache to split it's power in half and run HyperThreading <i>correctly</i>, and it'll be even more dangerous.

And then drop in x86-84 instructions and double the caches and registers to support double-sized integers without any loss of performance from 32-bit processing as the cherry on the top, and you litterally have a Hammer-killer. And it <i>should</i> all fit into the same die size as a Northwood uses if done correctly.

**sigh** Of course, it'll never happen. Intel won't produce a die of the same size, I'm sure. On top of that, Intel doesn't want to crush AMD. They want the supposed competition so that they're in no way a monopoly.

Which means that Prescott won't be much. A faster FSB. A bigger L1 and L2 cache. Still debatably-useful hyperthreading. Still crappy at raw FPU. Still shoddy at bit-shifting. A few integer tweaks. A few improvements to the instruction sets. A few SSE2 tweaks to try and improve the crappy FPU performance. And probably silicon for X86-64 that'll probably be disabled by default, if not entirely unusable. It'll probably be there though, even if it can't be used.

Certainly the Prescott will be no P5. Intel just can't afford to advance that much that quickly.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 7:10:17 PM

Quote:
Really, everyone should forget about the performance and get into energy saving and buy a VIA...;)

Ugh. What a load of crock <i>that</i> is. I honestly don't know how VIA has managed to fool so many people with it already.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 7:50:03 PM

You mean people actually buy VIA processors? When the Register had a competition to win one I entered. The challenge was to think of something creative to get the VIA C3 to do. I suggested you could use it to weigh paper down...for some reason I didn't win. Hmm...

AMD Is An Anagram Of MAD, Intel Is An Anagram Of INLET, Cyrix...Ah Who Cares?
August 26, 2002 7:53:35 PM

You mean people actually buy VIA processors? When the Register had a competition to win one I entered. The challenge was to think of something creative to get the VIA C3 to do. I suggested you could use it to weigh paper down...for some reason I didn't win. Hmm...

AMD Is An Anagram Of MAD, Intel Is An Anagram Of INLET, Cyrix...Ah Who Cares?
August 26, 2002 8:07:21 PM

silver phoenix.
why would AMD try to steal thunder from Intel if they knew that Intel 2800Mhz was about to come out and AMD knew that 2800mhz was more powerful than the 2600+xp. AMD knew they couldnt take the crown for even a week. so (btw all i said was my opinion, so dont call me clueless) IMO, AMD didnt try to take any thunder, they were just desperate (and almost forced) to release 2600+xp before the 2800Mhz from intel released. tell me if i dont make sense at all

btw, coolj and chuck: if you read my posts carefuly, i havent whined and said anything bad... i tried to keep peace and be cool with matisaro at all times.

real philosophy of life: "do onto others what you dont want them do onto you"
August 26, 2002 8:12:33 PM

Can someone tell the administrators that in the Sysmark Benchmarks, they didn't have a result for their overclocked 3400+ athlon XP?
August 26, 2002 8:36:02 PM

Quote:
Alrighty then. Here by my comments:
1) Thread redundancy is not exactly going to end the world.


ROFL! Seriously.....

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>
August 26, 2002 8:45:38 PM

Quote:
You mean people actually buy VIA processors? When the Register had a competition to win one I entered. The challenge was to think of something creative to get the VIA C3 to do. I suggested you could use it to weigh paper down...for some reason I didn't win. Hmm...

I suppose some people must, or else VIA wouldn't sell them anymore. I still find it humerous though that Intel offers considerably better solutions than VIA as 'mobile' processors. On top of that, a Duron can be underclocked to perform better for virtually the same thermal output.

I remember the Register's competition. My suggestion for an ultra-small (and hence portable) spy-cam server with a wireless network to take voyeurism to the next level lost too. :(  The world will be such a more drab place without it.

Oh well. Soon there will be affordable PDAs that can be used instead and be much smaller. ;) 

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 8:49:54 PM

They're good for compact systems. Those tiny cases for tiny jobs are great for via. For example, if you want to build your own Tivo you might as well use a compact system with a via CPU. A celeron or duron might do the trick, but the via CPU's are cool, simple, and there's not much worry about needing a lot of CPU power for something this simple.

<font color=red>I have a computer and it does weird stuff. please help.</font color=red>
August 26, 2002 8:51:36 PM

Quote:
silver phoenix.
why would AMD try to steal thunder from Intel if they knew that Intel 2800Mhz was about to come out and AMD knew that 2800mhz was more powerful than the 2600+xp. AMD knew they couldnt take the crown for even a week.

Simple. AMD doesn't care about being the best in the long term. They have absolutely no desire for that market at this time. <i>However</i> they knew that they could really stir up their fans into an excited frenzy by beating Intel, no matter how briefly. Look at sites like <A HREF="http://www.amdmb.com" target="_new">http://www.amdmb.com&lt;/A> to see how much it's done.

It was a clearly intentional and shameless attempt to steal Intel's thunder in many computer circles by revving up their fanboys.

Quote:
btw all i said was my opinion, so dont call me clueless

There <i>is</i> such a thing as a stupid question, just as there <i>is</i> such a thing as a clueless opinion or opinion holder. Don't blame me for pointing out your own shortcomings or inability to see the scene from a larger point of view.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 26, 2002 8:52:56 PM

Quote:
The funny thing is, the majority of 'recent' games would benefit more from a P4 than an Athlon.


Who wants to play Comanche 4? It's a real world gaming benchmark, but I never found any gamer who plays Comanche 4. Lots of gamer even don't know it's name. Every polular game ever released is not cpu hungry like Comanche 4. Aquanox is another game that is used to benchmark graphics card. This is graphics card hungry and sucks like Comanche 4.

Quote:
While it is sad that it took Intel so long, it is nice to see Intel at least finally working seriously on their IPC


When P4 will have higher IPC?

Quote:
and where AMD forces consumers into a single memory type while Intel offers choice.


No AMD buyers complain about absence of Rambus platform for AMD. Intel buyers would be happy if P4 worked to its full potential with DDR memory.
August 26, 2002 9:19:20 PM

It's interesting. The way you put it, and all the evidence I've seen supports what you're saying, is that in theory, it's only a matter of time till Intel puts togeather a P4 killer.

Meanwhile, the Hammer is still subject to wide speculation as to if it will be the P4/Athlon killer it has been billed to be by AMD and it's Fan(atic?)s.

AMD realy better pull it's collective head out of it's crack and start wowing the world with a real chip, because the awe factor of Hammer has worn off with most.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 26, 2002 10:36:40 PM

On the part about Prescott. Intel said at the release of the P4 that the P7 core will be the last they will ever make. So no, it won't be a new core. There will be no new cores until IA-64 trickles down to the desktop market (starting with support in Teja supposedly).
As for the enhancements. x87 FPU power is something Intel would like to throw away like a bad cold. They had hoped SSE2 would replace it but such a thing didn't happen because of the limited scalar functionality. Perhaps they will finally release such an extension with SSE3. If not, hopefully they'll wise up and realize that they do indeed need a scalar FP function on their MPU's.
On the part of the small enhancements, here's a list that I remember off-hand:
1. Double-pumped, 32-bit ALU's for simple integer operations.
2. Double-pumped L1 data and possibly trace cache.
3. Hyperthreading.
4. SSE3
5. The improved FSB and cache.

It seems they are focusing quite a lot on improving integer performance here. I don't know what enhancements SSE3 will bring but the double-pumped trace cache (if that happens) should be able to supply the execution units quite well. Now the only question is whether they'll expand the decoding stage. While the trace cache is nice and all, there are some parts of code that simply aren't reused and some loops which are too big to fit in the trace cache.
It is still too far away to see what the exact enhancements are but the sheer boost of cache and memory bandwidth alone should allow it a pretty hefty improvement as far as IPC compared to the current P4.
Hyperthreading on the P4 currently is also in a rather crude state. Remember, you only have 1 trace cache with 12k entries fed by 1 decoding unit. That's hardly enough to contain and run 2 separate threads. Another limitation would be that of the FPU. Despite what is said a lot, it is not the lack of another FPU that prevents the P4 from performing x87 FP operations fast. I think the biggest limitation would be the fact that the dual FPU pipelines are not fully pipelined. The main pipeline can take in 1 add one clock, and 1 multiply the next clock. It can't handle 2 adds one clock after another nor 2 multiplies one after another. This is very hindering and unless you hand code your program specifically to follow these examples, you will not reach the full potential of the FPU. Hyperthreading could certainly help here. You have 2 threads of independent instructions in which you can call individual add and multiply instructions each clock. This could significantly increase per clock FP performance. Although whether it'll actually improve in reality would depend on how well Intel implements Hyperthreading. They did say that it was "enhanced" in Prescott.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 27, 2002 3:08:11 AM

I agree.
I think that AMD's tactic has changed and that they will continue holding the competition by improving Hammer. Who knows really, it might be and turn out a P4 killer, but for how long...
Yeah it is grim now, the hype is gone because there was empty hype. I think AMD is learning from the current P4s and the next ones coming. If they ever hope to regain our interest, Hammer should have 10% lead in almost any bench, otherwise it'd be just another speed grade.

--
When buying an AthlonXP, please make sure the bus is at 133MHZ, or you will get a lower speed!
August 27, 2002 3:13:09 AM

but if hammer comes out in time (this winter?) then it might take the crown for a while until prescott comes out i guess. but prescott aint comin out till next summer (right?)

real philosophy of life: "do onto others what you dont want them do onto you"
August 27, 2002 3:16:30 AM

It definitly won't come out this winter, I am sure of it. And if it did, the crown it'd take would be pretty sad IMO. It's just not the K7 days anymore, I fear. I would certainly like AMD to come back with something that can hold of the P4 at bay for a while, or compete very well.

For 3 years AMD continued to get a lead 80% of the time with the K7, now it's Intel's turn and they are doing it aggressively. If AMD cannot regain that, they'll lose who they used to be known as, enthusiasts who aim for performance. So if Hammer is a bubble that bursts too quickly, I'd say that's sad.

--
When buying an AthlonXP, please make sure the bus is at 133MHZ, or you will get a lower speed!<P ID="edit"><FONT SIZE=-1><EM>Edited by Eden on 08/26/02 11:17 PM.</EM></FONT></P>
August 27, 2002 5:36:11 AM

"Keeping the same socket is rather useless if the 3GHz P4 might not be compatible with many (if not all) current motherboards."

Well, AMD did that exact same thing when they went to the 266 FSB Socket A's.

Knowledge is the key to understanding
August 27, 2002 5:44:06 AM

Seems to me the nice balance has been restored.
got a bit lopsided there with the tbred stuck on 2200+ for a while... but back to normal now.
And what do i mean?

1. P4 in the lead by a significant amount, depending on which benchmark you look at, with the Tbred overtaking on benchmarks that favour floating point.

2. AMD cpu's cheaper by a significant factor.

so there you have it. P4 is the performance king!, but at a hefty price including PC1066 RDRAM.

And i predict it will keep going like that, later tbreds/bartons being balanced by even higher P4 speeds.

<b>Due to Customer Complaints, this sig has been witdrawn from public use. Thankyou. :lol:  </b>
August 27, 2002 5:51:09 AM

Go with a SiS 648 along with DDR 400. That would give you the same (if not better) performance than a comparable PC1066 based system. It also brings the cost down alot because the chipset itself and the RAM is less expensive.

Knowledge is the key to understanding
August 27, 2002 12:54:10 PM

Quote:

Go with a SiS 648 along with DDR 400. That would give you the same (if not better) performance than a comparable PC1066 based system. It also brings the cost down alot because the chipset itself and the RAM is less expensive.

People, do you realize that PC1066 is only 2-3% faster than DDR333 or PC800? PC1066 is a mere 1% faster than DDR400.

Intelligence is not merely the wealth of knowledge but the sum of perception, wisdom, and knowledge.
August 27, 2002 2:07:39 PM

Quote:
People, do you realize that PC1066 is only 2-3% faster than DDR333 or PC800?

As I mentioned before, this depends completely on the app being benched. Check out the P4 2.8 review at <A HREF="http://www.hothardware.com" target="_new">Hot Hardware.</A> It compares the P4 2.8 on both PC800 and PC1066 RDRAM, at 533 FSB. Here are the percentage increases for each app they tested:

Business Winstone 2001 : 1%
Content Creation Winstone 2002 : 5%
FlaskMPEG DivX 5.02 : 8%
Xmpeg DivX 5.02 : 7%
PCMark 2002 (CPU) : 1%
PCMark 2002 (memory) : 12%
3DMark 2001SE : 4%
Comanche 4 : 6%
Quake 3 (640x480x16bpp) : 8%

As you can see, for any memory intensive operations (streaming media type stuff etc) PC1066 is noticeably faster than PC800 and may be entirely worth the cost delta to some people. Naturally games are currently more limited by graphics cards at typical resolutions, but with ATI's new R300 even 1024x768x32bpp is now limited by the CPU. Clearly, the additional bandwidth can be extremely helpful.

Ritesh
August 27, 2002 2:30:07 PM

I switched from Samsung PC2700 DDR to Samsung PC800 RDRAM on my P4 and overclocked it to PC1066 just for the heck of it. To my surprise I don't <i>feel</i> the difference. I can't for the life of me tell you there's a difference between DDR266 to DDR354 to PC800 to PC1066 in performance. Sure, with every jump there's a slight improvement in benchmarks, but I'm not sure if it's worth the extra $$$$.

Intelligence is not merely the wealth of knowledge but the sum of perception, wisdom, and knowledge.
August 27, 2002 7:17:20 PM

LOL this is the funniest [-peep-] i've come acrossed in this forum in a while!

"Opinions are like assholes, everyone has one"
August 27, 2002 7:23:05 PM

If we're looking for a benchmark that tests real life programs that people use, then slash off the first two because even if they are, it's not worth.
Flask, ok I'll grant you that.
Xmpeg, another worthy incraese.
PC Mark and 3dMark, forget the two, they are synthetic, and whatever increase there is, it's not worth noting, because it won't be realistic.
Commanche 4, don't care, the increase still remains in the 50 FPS, something so sucky by Nova Logic to program such game.
Quake 3, lol don't get me started on whatever improvement there is!

So IMO the price for PC1066 is really not that worth. For movie editing you're best off getting a 2.26GHZ P4 and OC the PC800 to PC1066, free performance and even more by the FSB.
Canadian pricing will make an even more dramatic difference in buying one of the two RDRAMs, I think AMD_Man or Chuck232 can vouch for that. PC1066 will obviously cost a fortune here, especially if today's usage is best off with 512MB, AND especially in WinXP.

--
When buying an AthlonXP, please make sure the bus is at 133MHZ, or you will get a lower speed!<P ID="edit"><FONT SIZE=-1><EM>Edited by Eden on 08/27/02 03:25 PM.</EM></FONT></P>
!