AMD vs Prescott

Okay, people really need to get their heads removed from their rear ends. I am tired of hearing all of this crap about the prescott's thermal output. I must have seen at least a hundred articles criticizing the prescott, calling it an oven, too hot, blah blah.

To put an end to this crap from these AMD geek boys, here is some nice hard evidence.

From AMD's site, the Athlon 64, which includes models FX51, FX53, 3000+ through 3800+, all have a maximum thermal output of 89.0 watts.

source:
http://www.amd.com/us-en/assets/cont...ec_Rev_ENG.pdf
http://www.amd.com/gb-uk/assets/cont...on_Rev_ENG.pdf


From Intel's site, the Intel 2.8E, 3.0E, 3.2E, (prescotts) all have a maximum thermal output of also 89.0 watts. The higher end chips, including the second revision 3.2E and 3.4E have a maximum thermal output of 103.0 watts.

source
ftp://download.intel.com/design/Pent...s/30056102.pdf (page 64)


So what was that you moron prescott critics? Oh you didn't know that your worthless 64 bit processors are exactly the same as the prescotts? And since the people who would be bashing the prescott's thermal output couldn't do the math, 103 watts is only a 15% increase for the higher end prescotts. (Which also have the performance increase of far more than 15%)


Another interesting point, AMD says flat out that any one of their crappy processors are going to die within seconds of being powered on without a heat sink. I wonder what Intel did about this; ignore the problem like AMD just did?

source:
http://www.amd.com/us-en/assets/cont..._Rev05_ENG.pdf (page 4)

Of course not, how could you forget about the Thermal Monitor Feature from Intel. The PROCHOT# signal, in both the prescott and new LGA socket processors, is a feature that specifically prevents the chip from overheating. This feature throttles down the frequency of the chip to prevent any and all damage from occurring, even in a catastrophic failure of a heat sink. (it falling off)

source:
ftp://download.intel.com/design/Pent...s/30056401.pdf (page 24)
ftp://download.intel.com/design/Pent...s/30255301.pdf (page 30)


For now, I only have one question for all of the prescott heat bashers out there, what are you bashing morons? Processors get hot, they always have. AMD processors are exactly, to the watt, as hot as prescotts are. If any of you have any evidence to refute what is here, I would very, very gladly like to see it.

Intel is a monopoly... People doesn't love monopoly...

Linux is to Windows what AMD is to Intel.

The only difference is that Intel is good and makes a better, faster product then AMD. Let's think for a second then, intel is better, faster, bigger, AMD is slower, smaller. . . so if Intel is faster and more efficient, why isn't intel linux, and AMD crappy microsoft.

Different strokes for different folks....

Since I'm not using autocad, maya or any other processor intensive apps on a regular basis I'm not going to see a performance gap between Intel and AMD. Given that fact, I see no reason for me to shell out the extra bucks for A NAME.

As for the 64 bit part of the discussion, you have to admit that this is one circumstance where Intel has to play catch-up with AMD and you should be GRATEFUL...things like this keep Intel honest and on its toes. After all, AMD has made YOUR Intel machine cost less and that makes us ALL winners....no matter which chip you prefer.

1. I have yet to see someone indulge in Prescott bashing at AO.
2. If you notice, AMD never disputed what you are saying, it's misguided zealots who did.

Quote:

---

Oh you didn't know that your worthless 64 bit processors are exactly the same as the prescotts?

---

Ummm... correct me if I'm wrong here, but the AMD 64's came first. That would make the Prescott a clone of the AMD 64 and not vice-versa?

Cheers,
cgkanchi

Quote:

---

Originally posted here by The3ntropy
The only difference is that Intel is good and makes a better, faster product then AMD. Let's think for a second then, intel is better, faster, bigger, AMD is slower, smaller. . . so if Intel is faster and more efficient, why isn't intel linux, and AMD crappy microsoft.

---

I really think you need to read a lot of review of AMD/Intel war... Intel processor are better in some test while AMD is better in other test (Mostly gaming).

If I take your idea, any compagy that are big do better product that small compagy?

Tim_Axe was poking fun at prescotts here posting this link about prescott heat.

I am grateful for competition and understand that it makes everything better. The 64 bit chips that run standard x86-32 applictions though, are going to continue to do only that. I have still not seen any significant server applications that are ready to benefit greatly from a migration to 64 bit. I understand the benefits of the jump, but my dual xenon 4GB ddr servers are still handling thousands of queries a second on the database and webserver, and are spitting out results without problems. No games take full advantage of 64 bit code. So until there is some reason software side for me to upgrade my hardware to 64 bit, all of you with your AMD 64's still look to me like people who drive military issue Hummer's in a city.

I never called either processor a clone of the other, just saying that they both had the same thermal heat dissipation.

Quote:

---

while AMD is better in other test (Mostly gaming)

---

Really? Oh you must be referring to Directx8 or OpenGL gaming, since that is the only place the Athlon 64 takes a lead over the P4. The Athlon 64 and P4 both ranked within 1% on all of the Directx9 tests but the P4 took a strong lead on almost every other benchmark they did. If the Athlon 64 pulled ahaed, it was by 1% or 2%.

source:
http://www6.tomshardware.com/cpu/200...n-fx53-17.html

One other thing to mention as well, before someone attempts to target price comparisons, I bought my 2.8C M0 for $200. Being an M0 SL6Z5, it was produced with a batch of northwood gallatin extreme editions (3.2EE, 3.4EE). I have got it up to 3.8 with only air cooling (a $50 swiftech MCX478-V), right now it is at 3.5Ghz. It doesn't have the additional cache activated, but it is running perfectly stable at 3.5 with a 1000Mhz FSB at 1.600 volts. The Athlon FX 53 is ~$700, mine was ~$250. If I even neglect the 25% FSB gain and higher core speed, I could call my chip the same as the 3.4EE in the previous link. Considering the 3.4EE did the same or better in every benchmark, except the old Directx8 and OpenGL tests, going dollar/peformance, I would score well over 200% of what the FX 53 would get.

This in my last answer to this thread because you cleary hate AMD and trying to make you understand AMD and Intel good/bad point would just be a waste of time.

But I have time to waste.

Xeon cost 1000$ and Athlon 64 cost 200$. Is Xeon x5 faster that an Athlon 64?

Well,

I take the view that the whole discussion is academic, unless you really have a use for high end equipment, and the majority of users do not.

Secondly, the processor is only one of the factors affecting performance...............it is important that all components are "balanced", so with a high end processor you are going to fetch up with a very expensive system.

Let's face it "bragging rights are expensive" :)

I think the whole heat issue is irrelevant provided that you use properly specified heatsinks, fans, thermal paste and a good case design. If it were that much of an issue then the manufacturers wouldn't make the processors because they wouldn't be able to sell them?

My feeling is that AMD have been slagged off for making hot processors for a long time, so now that Intel have made one the AMD boys are having a bit of revenge :D

just my thoughts

SDK > If you can provide any proof for your statement, I will gladly adapt my opinion to match the facts. You clearly said

Quote:

---

while AMD is better in other test (Mostly gaming)

---

I did not start this thread to have it become some AMD vs Intel thread, but if you can provide a link to a reputable site or some kind of benchmark that shows how 'AMD is better in gaming', I would gladly read it and like to see it. I understand that AMD has taken a lead in Directx8 and openGL architecture, but games that require these instructions have been outdated for years and any 1Ghz proc and generic video card can handle them fine (When was the last time you played a game of Quake and were worried about fps?). When I see that P4's and 64 FX's both get the same scores, with only a few percent variations, on Directx9a/b/c architectures, I call them the same. But since the P4 shines on damn near anything else you can throw at it because of its higher clock, I will continue to recommend it and use it. But again, if you can show something to refute what I have stated and posted here, I would be glad to read it.

The fact what you're oposing Xeon, 64 FX and P4 Extreme Edition suck for me. Those processor are for regular user machine. Those are just technology showoff.

Most of my reference website are in french but here some.

http://www.hardware.fr/articles/496/page12.html

http://www.materiel.be/cpu/2004cpu/page8.php

The graphic should talk for themself.

Quote:

---

Xeon cost 1000$ and Athlon 64 cost 200$. Is Xeon x5 faster that an Athlon 64?

---

The xeon is used in 2+ cpu systems. You cannot compare the xeon to a p4 or 64 FX since neither of these processors can run in a system with more than a single processor. If you want to do a price comparison, compare similar processors. The price of an atlon mp that runs at 2.2 Ghz (2800+), and the cost of a xeon that runs at 2.2 Ghz is the same, $200.


I fail to see your point from the graphs. I understand french enough to get the gist of what the article was saying, but I do not understand how you get that the Intel processors are so much worse for gaming from those two graphs. In all of the graphs shown, the processors were all within a few percent performance region, with the P4 ahead some time and the FX 53 others.

The 'CPU test' portion of 3d Mark has always been a joke. It has never reflected anywhere else. The 3dmark processor test showed that the FX 53 was ahead by 5 fps (90 vs 95) on one 'cpu' test and then on the other, it fell behind by .6 fps. (13.2 vs 13.8)


A much better way to benchmark processors for gaming is to bench two systems both with the exact same video card. Then compare the results when both systems had the exact video card.

From the link I posted previously, all systems were fitted with the exact same Radeon 9800 XT, and then 3dmark 03 was ran on all of them. Across the board there was a 10% range in graphic performance, with both the FX 53 and the P4 3.4 at the top within less than a 1% variation.

source:
http://www6.tomshardware.com/cpu/200...n-fx53-17.html (second graph)
*Note in this link there is also a black bar that shows the results of an FX 53 overclocked to the max, the article was written specifcally for the FX 53 and results of overclocked P4's are not included*

Same Review... Game Wise!

http://www.hardware.fr/articles/496/page10.html

I have 8 AMD Processor who are better or egal score in Unreal Tournament 2003 that the best Intel Processor.
I have 6 AMD Processor who are better or egal score in FarCry that the best Intel Processor.

http://www.hardware.fr/articles/496/page11.html
I have 6 AMD Processor who are better or egal score in Warcraft III and in IL-2 Sturmovik that the best Intel Processor.

http://www.materiel.be/cpu/2004cpu/page12.php
I have at least 4 AMD Processor who are better or egal score in Unreal Tournament 2003 that the best Intel Processor.

I am not familiar with IL-2 Sturmovik. Unreal Tournament 2003 is a game that I know well. As I previously admitted, the Athlons, especially the FX-64's, are better with Directx8 and OpenGL instructions than the Intels. Unreal Tournament 2003 and WarCraft 3 are both Directx8. If you find a benchmark done for Quake (1, 2 or 3), Half-Life 1 or Counter-Strike, I would agree and be willing to bet that an AMD chip will beat the Intel for those games as well. Looking at those scores though will show us that we our comparing in the hundreds of frames per second.

It may be just my opinion, but I believe once a game is well over 30-50 frames per second, and then even excedes the refresh rate of the monitor, it does not matter what it's exact score is.

For all of the games that are out now, and the games that will be out soon, Unreal Tournament 2004, Half Life 2, Doom 3, Painkiller, Manhunt, X2 The Threat, and recent benchmarking programs 3dMark 03, PcMark 04, et cetera, both AMD and Intel score very similarly.

Far Cry is a very interesting game. It is Directx9 based. There are 2 major versions of it out now. The original was written for Directx9b; the patch that was relased included Directx9c instructions, which mainly just included Shader Model 3.0.

As for the site that you posted, it showed that 6 AMD processors beat out Intel for the FarCry 'benchmark'. I thoroughly read through that entire article, as well as several other on that site and was unable to find any more information on this 'benchmark' that they did. The interesting point that I am attempting to bring up is that there is no official way to benchmark FarCry. That is why it is rarely used as a benchmark program. There are no applications that use its engine to reproduce the same run over and over. The only thing that I can imagine is that possibly they used the FarCry demo intro that shows no gun fire or fighting what so ever, as well as an extremely small amount of any textures, sand, water, lava, jungle, and no vehicles. Even for this demo, I have seen people use cards such as Radeon x800 XT's and Geforce 6800 Ultra's, and 80fps+ is not common.

Not only is it uncommon to achieve those high rates that they showed, I find it hard to believe that a few hundred Mhz difference in processors affected game performance that substantially in images per second. When I overclocked my processor's core by 35% (from 2.8Ghz to 3.8Ghz), I only noticed an increase of about 5% maximum on any games and benchmarks that I used, (no where near the 40% change that they show). Any game benchmarks hardly changed for me until I changed the core clock and memory clock of the graphics card.

If you could find out any more information about how they did this 'benchmark' on FarCry, I would be very grateful, as I have never seen such results anywhere else. I will continue to search their site and references to their site, but I am doubtful in my chance of finding anything else.

Ding! Dong!..............end of round three :D


I have a few comments to make, that I hope will slightly introduce other aspects than the purely technical?

Firstly, my position is very crude and very simple................if it does the job I go on cost..£$€, that is the deciding factor, surely it has to be. I use both AMD and Intel, and must say that I have NEVER had a bad experience with either. Surely, if they were that bad, they would not sell? If one were so superior to the other, the "other" would go out of business.

I still say that the secret lies in balanced components, and the purpose for which the system is required.

In my view, it is a bit like intelligence tests............they mostly show how good you are at intelligence tests, not how intelligent you are. In a way it is a bit like those crossword puzzles (motes croisses?)....................you may be very good at them, but can you articulate yourself like a confidence trickster with only a grade level education?

I am typing this on a machine with an Intel MoBo and PII/266 processor.............why?.............because it does the job. On the other side of the room I have an AMD 1900+ and an IBM with a PIII/533...............they have 1024 and 512 of much faster memory than this one's 128 Mb of 72 pin EDO RAM :D

I use this machine for general internet browsing, and typing letters etc.

I have a Digital Venturis PI/133 that will work as well............except that my ADSL provider's software won't accept it.............yet it benchmarks to within 12% of this one (OK, I admit that I did that as a "project").

Let's face it, gaming and graphics are pretty specialist, and in my opinion (based on serious commercial graphics Apps) are less dependent on miniscule processor performance differentials, as memory and the actual graphics card.

Let my epitath be:

"He loved the technology, but he hated the price"

:D

Quote:

---

if it does the job I go on cost..£$€, that is the deciding factor

---

Usually, AMD are cheaper that Intel.

See the USUALLY?

AMD are usually cheaper than Intel for similar performance. I know that if you compare the processors side by side, Intel will be more expensive most of the time.

However, there are great exceptions on either side. Which is why I never compared prices, because I believe them to be very similar as well.

My 2.8C M0 SL6Z5 for instance, which is a gallatin 3.4EE just labled as the 2.8, only cost me $200. With the stock cooler I could hit 3.2 before it got too hot, with the new swifty MCX478-V I put on there, I can now hit 3.8 with it. A 35% increase in performance from stock is great, especially considering how cheap the processor was. Now it is higher than even the newest LGA Intel processors that only hit 3.6 with an 800Mhz bus. A 3.4EE still costs rougly $1000. Mine is perfectly stable, I have been operating it for months and have had 0 crashes related to hardware, even though I push it and try to get it to. It is stable running Prime95 for 24+ hours any test.


So sure you can say that AMD are usually cheaper than Intel. But with great exceptions on either side, I have read several articles on athlons that clocked up very high very easily, it is hard to purchase either one based solely on price or make your decision on one brand because of price.

There is not a lot of user that overclock their machine. I don't consider that fact when buying a CPU.

Hey guys,

This is a really great thread, and I am enjoying it. Now I have a problem.............I would tend to prefer an Intel for over AMD for an overclocking competition......it has been a while but I took a P4 /2.26 and stoked it up to 3.066...................I wanted to break the 3 Gig barrier.

At the same time I built the Wife (notice the capitalisation) a Duron 1.3............I spent the money on software, because that is what she will and does use.

I have great affection for the 1.3 Duron........solid as a rock............I am not well impressed by Celerons though, as I have seen too many problems with them.

Some of us are lying in the gutter and looking up at you guys playing with the stars (appologies to Oscaar Wilde)

Cheers, and a good week-end (fin de la semain)
:D

I was not happy with the Prescott core. I bought a 3 GHz with 1 meg of cache and had a horrible heat problem. I bought an AMD 64 3000+ and never looked back. nuff said.

Quote:

---

There is not a lot of user that overclock their machine. I don't consider that fact when buying a CPU.

---

How can you not consider overclocking a factor when it can make that much of an enormous difference? I always squeeze all of the performance that I can get out of my hardware. When there is the potential to increase the speed by 30-50% how can you ignore that and then go compare the results of the processors running at stock speeds? I pushed an additional 35% out of mine which is very noticable, with only a minimal amount of modified cooling (buying a new heatsink), it is as stable as it is at stock speeds, and now I never even need to bother with it (unless I want to), it just works.

Why would you not want your processor to run at its fastest possible speed?

Did you read? I never said anything about overclocking. When I bought mine, it ran hot out of the box.

http://www.anandtech.com/cpuchipsets...spx?i=2026&p=3

Unless it has been improved.... I never bothered to look back.

alittlebitnumb > If you refer to the fact sheets that are from both intel and amd, that I have posted previously, you will see that both processors have the same thermal output of 89.0 watts. This thread was not about the temperature of the processors under load or at idle, the discussion just drifted off with me and SDK.

The temperature of the processors at idle and load will be very different, that will depend upon the hertz rating of the processor and the heatsink that is being used. If these factors were kept the same, then the temperatures would be the same. With different heatsinks being used, and also the different clock speeds of the two processors, you cannot compare the temperatures that they are at during idle and load tests.

I was comparing the actual thermal output of the two processors, which in this case is the same. Both companies show that the processors both output 89.0 watts.

I run a P4 3.2E (yes prescott @ 800mhz FSB adn 1MB cache), and thers not a whole lot of difference with the heat generated by the processor from my old 2.8ghz machine.

I prefer the Intel chips over the AMD's. Sure AMD's are cheaper, but I get more preformance out of the Intels than what I do with the AMD's. And yes I do game a lot but I don't really notice a difference with the AMD processor.

I don't think human eyes could see difference between AMD and Intel Processor at all if they are supposed to be equal. (Exemple, P4 3.2 vs Athlon XP 3200+)

You have to remember, that the P4 3.2 runs @ 3.2ghz, the XP 3200 runs at what? 2.1ghz?

Some thing like that but their core architecture is very different! Intel was always marketing MHz but now, they stop. Something that AMD learn a while back! :)

Sorry I came to this thread late, but I've been away on a trip for the past week or two...


There is a problem with your standing on your source material. If you take a very close look at your Intel source, you would see that 103 watts is the TDP, or average thermal dissapation. Intel's maximum thermal output is not listed anywhere in that document!

On AMD's source, they list the maximum. You cannot compare AMD's maximum heat dissapation to Intel's average heat disspation and claim that they are the same temperatue.


Yes I run AMD. But this is because I wanted an affordable Dual Processor system. I would also happily build a prescott system and overclock it like a bat out of hell. But that would mean single processor, and I enjoy two processors too much to go a single processor route.

I'm not trying to sound mad, but I'm trying to show that I understand both sides of the issue, and want to point out that you are standing on something very unsteady. We can talk about whatever you want in this thread, but I just wanted to show that your primary source documents are being interperted inaccurately.

Quote:

---

Originally posted here by cheyenne1212
You have to remember, that the P4 3.2 runs @ 3.2ghz, the XP 3200 runs at what? 2.1ghz?

---

Cheyenne ... that's the whole point,
Intel's focus is on Mhz, GHz nowadays, AMD on actual performance.
AMD cpu's do more in one clock cycle then a similar Intel cpu, that's way they can have lower Mhz rates and still perform very well.

Next, I see people here comparing Athlon64, Athlon FX, Xeon, P4's, those cpu's are all very different and built for other purposes.
Like stated before by people overhere is the Xeon primarly a cpu used in multiprocessor systems. If you want to compare cpu's and price tags then compare cpu's that aim at the same market share. For the Xeon this would be the Opteron (also a member of the 64bit Hammer architecture family). The Celeron compares with the Duron, the P4 with the Athlon.
The Athlon64 has the advantage that it has 64bit and 32bit compatibility while no other processor for the home market has this onboard, so you have to say AMD is ready for the home pc future.

The prescott is probably the fastest cpu at the moment but probably also the most power hungry and hot one.

If you realy want impressive oc results and cool temperatures without spending a lot of money I would go for an Athlon Mobile with Barton core with the right batch ID, those run at low stock voltage, meaning they run cooler and have a lot of head room for oc'ing, and as some sort of bonus they come all unlocked. This is cause the power management in laptops need to influence cpu speed in order to save battery time.

conclusion the AMD / Intel struggle is more about emotions and what kind of stuff you want to do with your pc.

Quote:

---

Originally posted here by VictorKaum
If you realy want impressive oc results and cool temperatures without spending a lot of money I would go for an Athlon Mobile with Barton core with the right batch ID, those run at low stock voltage, meaning they run cooler and have a lot of head room for oc'ing, and as some sort of bonus they come all unlocked. This is cause the power management in laptops need to influence cpu speed in order to save battery time.[/B]

---

SO TRUE!


I have two sweet Mobile Bartons in my current rig -- Dual IQYHA-stepping 35watt 2400+ Mobile Bartons. They're running at 2.4GHz / 1.7v. Temps are in the upper 40's to lower 50's. Passed Prime95 without trouble. 2.4GHz is ~3400-3500+ish or so. And running them dual is like a dream. Very nice indeed. Overall the processors, motherboard, and cooling cost less than $450. I had HDD/CDROM, and getting everything else (minus video card) ran me up to $650. Not bad for this mean of a machine. :cool:

AMDs are the best for gaming and Intels are the best for multi-tasking and multimedia work.