• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Toms Review: P4 3000 vs. XP 2300+

D

dowxp

Diamond Member
Speed Euphoria: AMD vs. Intel - Who's Ahead?

"In the benchmark results, the Athlon XP 2300+ cannot quite keep pace with the Intel Pentium 4/3000, but the values that we measured are very impressive. In order to outperform the Intel Pentium 4/3000, AMD must switch over to 0.13 Micron in order to enable higher clock speeds. According to our laboratory results, with a speed of 1933 MHz, AMD can reach the performance of a Pentium 4/3000 based on DDR SDRAM. This clock speed can only be achieved by the new Athlon XP with the Thoroughbred core, which is expected soon. "
 
And people said that RAMBUS was bad. I guess this really shows everyone how much AMD needs the .13 micron chips.
 
Comparing the synthetic benches to the real ones, one notices a disparity. Clearly, SSE2 is finally showing its muscle.
 
Just as I thought, a lower clocked Northwood on an i850 board is better than a higher clocked Northwood on an i845D board (ie 2.6GHz vs 3GHz). If you want the fastest setup for games there is no question about it, P4 Northwood pair with an i850 mobo is the way to go. 😉
 
1933 mhz vs 3 ghz. for such a big difference in speed i have to say the XP didn't do bad at all.
 
Someone should make a P4 DDR SDRAM chipset with dual-channel capabilities. With DDR333 support the theoretical maximum could reach 5312MB/s. That would be significantly ahead of even PC1066RDRAM in terms of bandwidth. It would be better in the area of latency as well.
 


<< 1933 mhz vs 3 ghz. for such a big difference in speed i have to say the XP didn't do bad at all. >>

Exactley. Totally agree,

<< Just as I thought, a lower clocked Northwood on an i850 board is better than a higher clocked Northwood on an i845D board (ie 2.6GHz vs 3GHz). If you want the fastest setup for games there is no question about it, P4 Northwood pair with an i850 mobo is the way to go. >>

Well, Frank from THG didn't tell what the memory was running at for the 2.6 850 test setup. If it was running at PC1066 speeds, then this proves what I have said all along how PC1066+533fsb will make a significant difference when compared to 533fsb+PC2100. I did email though Frank about what was the memory running at.

EDIT:

<< Someone should make a P4 DDR SDRAM chipset with dual-channel capabilities. >>

Intel is supposedly developing one, but the problem with such a chipset is that it will cost so much that in the end, 850 will be cheaper, and the performance will not be worth the cost of a Dual Channel DDR chipset. nForce, maybe it would work, but Intel and nVidia are on such bad terms that such a partnership is unlikely.
 


<< Someone should make a P4 DDR SDRAM chipset with dual-channel capabilities. With DDR333 support the theoretical maximum could reach 5312MB/s. That would be significantly ahead of even PC1066RDRAM in terms of bandwidth. It would be better in the area of latency as well. >>



What about the PC1200 RDRAM revealed this weekend?
 


<<

<< Someone should make a P4 DDR SDRAM chipset with dual-channel capabilities. With DDR333 support the theoretical maximum could reach 5312MB/s. That would be significantly ahead of even PC1066RDRAM in terms of bandwidth. It would be better in the area of latency as well. >>



What about the PC1200 RDRAM revealed this weekend? >>



pc1066 isnt even out yet and neither is tulloch
 


<< pc1066 isnt even out yet and neither is tulloch >>



Correct, but I fail to see your point. Bovinicus stated that someone should build XYZ. Then you state that XYZ isn't even out yet. DUH! If it isn't even out yet, then someone should build it.
 
Tom's suite of benchmarks is clearly tilted toward Intel. Why, you ask? Well...

1) He only shows the lightscape results of SpecViewPerf, and that's the one in which even "Willamette" P4 is faster than the Athlon XP with corresponding model number. For reference, see Anandtech's SpecViewPerf tests.

2) He uses Lightwave 7.0b. While this is a valid real-world application, it clearly has a bug which causes it not to use any SSE instructions on other than Intel processors. How do I know? Just look at this graph at Tom's. Athlon Thunderbird 1400MHz is faster than Athlon XP 1500+ (1333MHz), and both are eclipced by 1200MHz Tualatin P3 running SDRAM on 100MHz FSB.
 
The problem with the article on tomshardware is the following:
Most of the benchmarks only show the numbers for the Athlon XP2000+

Now, maybe they just screwed up the numbering, in any case, the article once again seems to be sub par...

 


<< Well, Frank from THG didn't tell what the memory was running at for the 2.6 850 test setup. If it was running at PC1066 speeds, then this proves what I have said all along how PC1066+533fsb will make a significant difference when compared to 533fsb+PC2100. I did email though Frank about what was the memory running at. >>


It was running at a 472MHz FSB, instead of the default 400MHz FSB. It doesn't matter if he is using PC1066 or PC800 (I'm sure he is using PC800), 472MHz FSB is a 472MHz FSB no matter what type of RDRAM you use. PC1066 just lets you run at a higher FSB (ie. 533MHz). Just like people use PC2400 or PC2700 to run there DDR based mobos at a higher FSB.
 
<<SSE2 is finally showing its muscle. >>

I agree, but AMD already told us SSE2 will be implimented with there next gen of processors so SSE2 wont be a factor in comparing Amd's best offerings with Intel's. How much further can Intel go with their P4?, personaly it dosent look like AMD has much of anything to worry about, less the P4 is able to hit 5ghz anytime soon. 2ghz should be easy for AMD with the .13u process in any case.
 
The P4 needs 1100 more mhz to overtake the Athlon. Funny

Actually if you looked through the P4 2.4ghz beat the xp 2300 in every test except 3dmark2000(old) and the synthetic sandra benches.

Edit: The p4 3ghz is weak because ofcourse the memory benchmarks are bad. 2000/2000. They are using the wrong board. I get 2400/2400 with the Asus P4B266 with a 2ghz northwood at 2.46ghz. The asus p4b266 has an option to allow 3:4 on the cpu and memory. Right now my DDR is running at 156FSB while the FSB is at 123. If they used the Asus board with that 3ghz cpu the memory benchmarks would have been in around 2600/2600. That would have easily caused the p4 3ghz to win everything. The cpu just got paired with the wrong board. PC2700 DDR goes real well with this board.
 


<< Just as I thought, a lower clocked Northwood on an i850 board is better than a higher clocked Northwood on an i845D board (ie 2.6GHz vs 3GHz). If you want the fastest setup for games there is no question about it, P4 Northwood pair with an i850 mobo is the way to go. 😉 >>



The only reason why the RDRAM outperformed the DDR was that the DDR couldn't provide the bandwidth for the dramatically increased FSB. If DDR could be OCed or PC2700 was used, then it would most certainly remain in competitionn.

-DocSmarts
 


<< 1933 mhz vs 3 ghz. for such a big difference in speed i have to say the XP didn't do bad at all. >>



The reason for this was the memory bottleneck. Intel needs DDRII and RDRAM 533.

-DocSmarts
 


<< 1933 mhz vs 3 ghz. for such a big difference in speed i have to say the XP didn't do bad at all. >>



Actually guys, Tom is up to his old Tricks.


What he didn't show you was some of these DDR mothabords like the Asus, have a 3:4 ratio that DRASTICALLY increases the bandwith to the point it is very very close with the i850 rambus.


And I have the proof.

My best score on 3dmark is 9522. Amused it right behind me at 9498 (or close to that). EDIT: Amsued has the i850 chipset

He is clocked at 2.4 I believe to get that score. I am at 2.7 and his card allows him to overclock to 265/602 whereas I can only run stable with my t500 at 265/590

So, we can take the numbers Tom gave us for the P4 at 2600 true but the P4@3000 is bugus because Tom did not overclock the Ram like MANY of the i845 motherboards can.

Sorry guys, Tom did it again.


 


<< The P4 needs 1100 more mhz to overtake the Athlon. Funny

Actually if you looked through the P4 2.4ghz beat the xp 2300 in every test except 3dmark2000(old) and the synthetic sandra benches.

Edit: The p4 3ghz is weak because ofcourse the memory benchmarks are bad. 2000/2000. They are using the wrong board. I get 2400/2400 with the Asus P4B266 with a 2ghz northwood at 2.46ghz. The asus p4b266 has an option to allow 3:4 on the cpu and memory. Right now my DDR is running at 156FSB while the FSB is at 123. If they used the Asus board with that 3ghz cpu the memory benchmarks would have been in around 2600/2600. That would have easily caused the p4 3ghz to win everything. The cpu just got paired with the wrong board. PC2700 DDR goes real well with this board. >>



I have to disagree. Even with the increased FSB and the lowering ration 3:4, the bottleneck is still in the memory no matter what speed your FSB is running.

By the way, does that asus board also offer PCI & AGP locking? If not, do you know of which do?

-DocSmarts
 
Why are people so caught up on mhz vs mhz? There is a reason that AMD went to a PR rating, mhz myth, which apparently almost everyone in these forums is caught up in. Mhz doesnt matter! Performance matters.

I think the thing you are missing is that the latest Intel chip out right now, will outperform the latest AMD chip right now. And it looks to continue in the future. Despite the mhz myth, the top chips from each company is what you should be compairing. Intel outperforms AMD...and has for awhile now.

Don't jump on me cause i own an XP 1800, but Intel chips are no slouch. About the ONLY thing you can pick on is the pricing.

But yet again, Intel does NOT price their chips for resellers and non OEM. OEMS get a much better deal and that's why you can get an equal performance P4 solutions for the same price as an AMD. AMD prices their chips so low that they hurt themselves in the long run (see this quarters loss). Intel doesn't do such a silly thing and they stay profitable. AMD HAS to buy their way into peoples hearts cause they are starting from a huge disadvantage.

All i'm saying is...quit worrying so much about mhz vs mhz. Think Performance vs Performance. You guys really don't need more excuses to prove AMD is a good chip. They have totally different architectures. AMD has a great chip...as does Intel.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top