Sign-in / Sign-up
Your question

AMD vs. Intel BATTLE

Tags:
  • CPUs
  • AMD
  • Intel
  • Product
Last response: in CPUs
January 7, 2002 7:18:37 AM

hi all. our intel fans have something to be proud of now that intel has its .13 micron cpu out. however i think the bench marks are not as fair as possible. the only thing is i think a 2200 intel should vs a 2200 athlon. i'm glad intel is doing well. this could only mean better CPUs coming out faster. i wonder why THG didn't use the overclocked athlon to compare it to the 2200 P4?

<font color=purple><b>I have a pc with a built-in house.</b> :wink: </font color=purple>

More about : amd intel battle

January 7, 2002 7:27:47 AM

Quote:
hi all. our intel fans have something to be proud of now that intel has its .13 micron cpu out. however i think the bench marks are not as fair as possible. the only thing is i think a 2200 intel should vs a 2200 athlon. i'm glad intel is doing well. this could only mean better CPUs coming out faster. i wonder why THG didn't use the overclocked athlon to compare it to the 2200 P4?

Number 1: This is a comparison between the top of the line processors of each company. If you overclock one, you must overclock the other to be fair.

Number 2: There is no 2.2GHz Athlon. There is not even a 2GHz Athlon. They currently max out at 1.67GHz (this would be the XP 2000 model.)

Number 3: Comparisons based on the same clockspeed are now meaningless because the two processors have vastly different designs.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
January 7, 2002 8:04:39 AM

ok the only thing i'll agree with you on this one is your number 1. now where the heck do you see me use mhz or ghz on my post? therefore a 2200 is a 2200 whatever. since your number 2 is meaningless your number three don't mean nothin. as for your number 1 i was talking about the benchmarks that THG sometimes does in the last part of the article. THG sometimes has overclocked cpu compared with normal CPUs. =P


<font color=purple><b>I have a pc with a built-in house.</b> :wink: </font color=purple>
Related resources
Can't find your answer ? Ask !
January 7, 2002 8:21:21 AM

WTF? That was rather uncalled for. Not once did Raystonn insult you, and you turn around and drill him.

Point #2 was probably a slight thinko on his part. BFD.

As for your other bit...perhaps Tom didn't test an AthlonXP 2200+ because there is currently no known way to overclock it that far. The AthlonXP's multiplier only takes it to 1.67GHz (2000+ in PR speak) unless you overclock the FSB.

Kelledin
[dave@discovery ~] kill -9 1
init: Just what do you think you're doing, Dave?
January 7, 2002 8:29:16 AM

yeah i think i went overboard. well as you seen the other post he made a dumb comment. i think i was a bit ahhhhhhhh WTF i'll change the post.


<font color=purple><b>I have a pc with a built-in house.</b> :wink: </font color=purple>
Anonymous
a b à CPUs
January 7, 2002 9:04:19 AM

With the reduced core size, the reduced voltage, and doubled cache, I'd be interested in knowing whether the Northwood's overall heat dissipation is less then that of the Williamette, it's got 2 variables in it's favour, though I wonder whether the cache on a processor(any) dissipates a good portion of the total heat output.

Now, where is the competitive Intel pricing?; They're using 30mm Waffers, with chips not much bigger then that of the Athlon, yet they're charging much more than that of the Athlon. It would be nice, though hopeless to imagine; If Intel would actually offer a platform that was equivalent in performance to the AMD at the same cost.
January 7, 2002 12:21:53 PM

I was disappointed with the bench results of Northwood. With the die shrink to .13 and the additional cache, one would think that the processor would show a significant improvement in performance, leaving the Palomino in the dust. But, the gains were not as great as I was lead to believe by the hype of Northwood.

Since the die shrink did not produce an Athlon killer, it makes sense to ASSUME that when AMD goes to the .13 die, the Athlons speed won't be significantly increased by the die shrink alone.

In the meantime, it looks as if price/performance is still with AMD, and that gap will increase in AMD's favor when the .13 Athlons are produced (much smaller die size then the P4). I would even speculate that that we will see again the hugh price cuts for the .13 Athlon that we have seen with the T-Bird, as AMD will produce many more CPU's per wafer then they currently do (and more then Intel can produce per wafer).

<font color=blue>This is a Forum, not a playground. Treat it with Respect.</font color=blue>
January 7, 2002 1:04:15 PM

<blockquote><font size=1>Svar på:</font><hr><p>With the die shrink to .13 and the additional cache, one would think that the processor would show a significant improvement in performance, leaving the Palomino in the dust.<p><hr></blockquote><p>No one has ever claimed that a die-shrink in itself would boost the P4. You've got it all wrong; the die-shrink only paves way for a new series of clockspeed increases. It's the 2'nd level cache which was hoped to increase performance, and it does, between 5-9% depending on the application. Not that much, but every little bit helps.

I was actually surprised to see that Intels new offering was able to pass the Athlon XP2000+ in the majority of the benchmarks. Now Intel has to turn the speed-dial a bit more, so that they have a clearly better version (with a price-premium, that OK with me) so that the 2.2GHz Northwood can be priced like the Athlon XP2000+. This could be done together with the general price-adjustments later this month. But I doubt very much it'll happen. Intel are to greedy.




/Copenhagen - P4 Willamette 1700MHz@2109 MHz, Vcore 1.75V@2.20V on Abit TH7II-RAID.
January 7, 2002 1:36:02 PM

Quote:
Since the die shrink did not produce an Athlon killer, it makes sense to ASSUME that when AMD goes to the .13 die, the Athlons speed won't be significantly increased by the die shrink alone.


Die shrinks alone offer MINIMAL performance gains, like nearly unmeasurably minimal, due to the shorter trace paths etc etc, however northwoods performance boost came from the cache doubling, not the die shrink.


If the tbred does not have any core changes(and some sites claim they may have 512k of cache) Then a 1.5ghz tbred will perform pretty much EXACTLY the same as a 1.5ghz palomino.

The difference is in the cores newfound ability to be clocked higher due to the die shrink.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 7, 2002 1:38:09 PM

Yeah, anandtech has the 2.2ghz and xp2000 neck and neck, really too close to call.

One major difference though, about 200 dollars worth.
The 2ghz northwood which intel has priced to compete with amd's 2000+ is not faster than it, in fact it loses most benchmarks versus the axp2000+, so price/performance is still in amds court.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 7, 2002 1:54:59 PM

Yes, unless you get a P4 tweakers mobo, like the one I'm using, and get the 2.0GHz Northwood going; something like 2.4GHz and PC960 is definately within reach using the stock heatsink. I don't believe it will be easy, if at all possible, to get an Athlon XP2000+ up to similar performance. In that case the Northwood 2.0GHz will be giving the biggest bang per buck.


/Copenhagen - P4 Willamette 1700MHz@2109 MHz, Vcore 1.75V@2.20V on Abit TH7II-RAID.
January 7, 2002 2:22:57 PM

Unless the overclocked p4 can offer more than 1.5 toimes the performance of the maximum overclocked performance of the amd, the northwood is NOT more bang for the buck.


You pay 1 and a half times more for the p4 than you do for the amd, ONLY if the performance of the p4 is 1 and a half times that of the amd is the p4 the better bang for the buck.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 7, 2002 2:46:41 PM

I have an AMD Athlon and my opinion on this AMD vs Intel battle is this:

The bottom line is that both companies has provided their flagship processor for review, the P4 2200 performs better than the AthlonXP 2000+. If AMD had the technology/capabitlity of producing a faster chip by whatever method they wish either up the MHz or make their chip do more per MHz, they would have done it. But they haven't.

So for now we can conclude that Intel make faster chips than AMD.

If we talk about value for money then that's another issue.

Anyway to be honest I wonder if the higher power usage of the Athlon would eventually even up the money you spend on the system.

Just my tuppence worth.

-----------------------------------
Don't extract the urine.
January 7, 2002 3:05:35 PM

Quote:
where the heck do you see me use mhz or ghz on my post?

It was inferred by your post because obviously it makes no sense without this. What possible reason would you have of increasing the PR rating of the AMD processor to equal the clockspeed of the Pentium 4 when A) AMD's PR rating has nothing to do with Intel's clockspeed, B) AMD has stated its PR rating is based on the T-Bird, not any Intel processor, and C) AMD can set whatever PR rating for their processors that they like based on their own opinions.


Quote:
THG sometimes has overclocked cpu compared with normal CPUs.

Yes, and they refused to overclock the Intel-based system in the same manner in which they overclocked the AMD-based system, using the FSB; and this even though you can do it on both systems without moving the AGP or PCI out of spec, and you can even do it on the Intel system without running the RDRAM out of spec as well. Was there any reason for what they chose not to do? Either they knew it would work and they did not want to do it, or they did not know much about the hardware they were reviewing. I am choosing to go with the latter to give them the benefit of the doubt for the moment.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =
January 7, 2002 8:01:02 PM

well i don't know what you got out of my post, but i was not thinking about mhz or ghz. and the PR rating is based on northwood not Tbird. i'm a member of the AMD PRO team and we also talked about this. for a while now people have been saying different thing and i never made a comment, but now people can see that the PR rating is about the same as the northwood. you can see this on most benchmarks.

i won't make any assumptions as to why THG didn't overclock.
i have to bee 100% sure before i make that comment.

<font color=purple><b>I have a pc with a built-in house.</b> :wink: </font color=purple>
January 7, 2002 8:14:17 PM

Quote:
the PR rating is based on northwood not Tbird

Well it would seem we have reached the limit of your knowledge on this subject. You can even ask Matisaro, the person who usually disagrees with just about everything I have to say. Heck, ask this entire forum. The PR rating on Athlon XP processors is measured to give performance relative to the T-Bird. AMD has stated this themselves.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
January 7, 2002 8:29:25 PM

i don't know you or Matisaro. i'll see if more people agree with you. then i'll do some research and see if that is true. it wont bother me to say you are right, but thats once i do some research. i've seen different people say its based on different CPUs so i'll email AMD and make sure. if i'm not mistaken even THG said its based on the northwood, but that probably don't mean nothin to you.

<font color=purple><b>I have a pc with a built-in house.</b> :wink: </font color=purple>
January 7, 2002 9:12:24 PM

Scamtron, you're wrong! The PR rating AMD uses are supposedly based on the T-Bird. They even claim that on their site! How could AMD possibly know the performance of the Northwood months before it's released unless they have spies at Intel, lol.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 7, 2002 9:33:24 PM

I have to agree with Raystonn. PR is based on the tbird. Always have. Dumb people who think otherwise. Read More.
January 7, 2002 9:40:26 PM

indeed. PR is refrenced to the T-bird.

i have a question for you Ray, about northwood OC-ing.
whats the setup on the 850 to deal with elevated FSB and ram speed. does it have a fixed pci/agp as seems to exist on the 845D or do u use a higher divider?

- Know thyself
- Seek
- Love
- Cherish Life
- And do not Yield!
January 7, 2002 9:40:39 PM

Well read the past post Matisaro and Ray and amd say its based on the frist tbird. So a Amd athlon 1800 is just like a 1.8 tbird
January 7, 2002 10:10:04 PM

Quote:
whats the setup on the 850 to deal with elevated FSB and ram speed. does it have a fixed pci/agp as seems to exist on the 845D or do u use a higher divider?

The PCI and AGP buses depend on the motherboard. The Abit TH7II allows you to fix the PCI/AGP at 33/66 regardless of how the FSB is currently set, or use a divider. The P4T-E offers a 1/4 divider.

The memory speed is set using a selection of either 3x, 4x, or 'auto'. If you want to keep your memory within PC800 specifications while you use a 533MHz FSB, simply use the 3x multiplier. 133.3MHz x 3 = 400MHz. It comes out to the same 400MHz as you would have gotten on the normal 100MHz x 4 setting.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
January 7, 2002 10:44:11 PM

Does raising the bus speed without touching anything else, (RAM speed, AGP bus and PCI bus and processor speed) improve performance? I mean, will a 1.8GHz P4 with 533MHz bus and PC800 RAM be significantly faster than one with a 400MHz bus?

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 7, 2002 11:08:19 PM

<font color=blue>"Die shrinks alone offer MINIMAL performance gains"</font color=blue>

How about that! I always thought that the smaller component construction was a major factor in speed, as traces are shorter, transistors are smaller and switch faster, less inductance to overcome in the component. But, as demonstrated (to my surprise), the die shrink was insignificant.

I understand how the smaller construction allows for faster clock speeds, but this (according to my level of comprehension) will only allow for a few hundred more MHZ of speed if no redesign of the chip is introduced.

So the Thoroughbred will probably top out at about 2.2ghz (not sure of the PR rating). But even with the smaller die, and especially if more cache is added, this will be one hot chip!

Looks like AMD will be depending on the Hammer series to compete with the Northwood beyond 2.2ghz (not PR).

<font color=blue>This is a Forum, not a playground. Treat it with Respect.</font color=blue>
January 7, 2002 11:41:58 PM

Actually about the PR rating on Tbird.
At first it sounded real, but due to 66MHZ increments and 100 points per new grade, it seems to me that it clutches.
First of all, the AXP 2000, which "supposedly" is a Tbird at 2GHZ, would have gotten better than a P4 2.4GHZ at least before it is outrunned. Unfortunatly it seems otherwise as most benchs are tilting over the P4 2.2GHZ, in which 200MHZ difference between Tbirds and P4s is not enough before Tbirds are beaten. So at XP2000 the value is still going down due to an unsynched follow-up in clock cycles.
What this comes down to, is:
The more you increase the PR Rating, the more the MHZ difference between it and a Tbird, and the less the PR performance defines, and the less it will beat P4s like a Tbird did.
If they had done 100MHZ increments, this would all make sense so damn well, and an XP2000 would indeed not be beaten down by a P4 2.2GHZ.

--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...
January 7, 2002 11:51:22 PM

If each 100 PR was 100MHz then there would be no point to the PR ratings. The PR ratings should be the average IPC of AXP/ average IPC of the T-Bird. And this should remain constant. Assumming the AXP is 20% faster per clock than the T-Bird, then the AXP's PR rating should always be 20% higher than it's clock rating, however, this not the case. So, indeed, AMD's PR rates are making less and less sense unless AMD is secretly improving the IPC of the AXP ever so slightly with each release.

Anyway, you can't compare a 2.2GHz Willamette to a 2.2GHz Northwood in relation to the T-Bird because on average, the Northwood has a higher IPC than the Willy. Anyway, IPC is a relative term as we do not know the exact # of IPC each processor is capable of but I'm willing to bet that will all the latency we are seeing, both the AXP and Northwood are doing under 1IPC average.



AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 8, 2002 1:10:32 AM

hmmm. a most excellent answer.
wish that happened on more athlon chipsets.
*frowns at my partly dead IDE controller*

The only loyalties i have is to Performance, Cost
Reliability and the Truth.
January 8, 2002 5:13:08 AM

Quote:
Does raising the bus speed without touching anything else, (RAM speed, AGP bus and PCI bus and processor speed) improve performance? I mean, will a 1.8GHz P4 with 533MHz bus and PC800 RAM be significantly faster than one with a 400MHz bus?

Raising the FSB (Front-Side Bus, aka System Bus) on a Pentium 4 system will increase the clockspeed of the processor. The processor's clockspeed is determined by the multiplication of the FSB and the 'multiplier'. The multiplier is locked at a certain number. So, when you increase the FSB, you increase its clockspeed. There is no way to raise the FSB speed on a Pentium 4 system without also increasing the processor's clockspeed.

Now then, if there was a 533MHz System Bus version of the Pentium 4 released (aka Northwood B), it would have more bandwidth and less latency when using memory. Many applications are memory limited, and many others are bottlenecked by memory fetches (latency). All of these applications would see improvements of up to 33% if all other system components remained the same.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
January 8, 2002 9:26:25 AM

Heh, hes right, the pr ratings is in relation to the tbird, not the northwood.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!