PDA

View Full Version : Nforce is keeping AMD afloat. What if P4 got Nforce?


DXnfiniteFX
01-02-03, 09:40 PM
It is a given that the performance of the Athlon XP is slipping behind the P4s. However, they are still managing to keep up mainly because of the incredible boosts attributed to the Nforce.

However, I have heard that the P4 will EVENTUALLY get the Nforce platform under its belt. . . I think it was sometime this year. If that is true, will AMD be able to keep up? Or will Intel continue their performance gap? A P4 Nforce would likely mean a Prescott platform as well. . . and with the advanced hyper-threading of the Prescott with the naturally hyper-threading friendly nature of Nforce performance optimizations, how will Nvidia affect this power struggle? I honestly think they hold the trigger.

thcdru2k
01-02-03, 09:51 PM
i thought intel didn't want nforce because they wanted to make money off their own mb chipsets and nvidia wanted to use intel for cpu but intel said no.

FastM
01-02-03, 11:35 PM
Originally posted by thcdru2k
i thought intel didn't want nforce because they wanted to make money off their own mb chipsets and nvidia wanted to use intel for cpu but intel said no.

I heard Intel wants to much money to allow nVidia to distribute chipsets for intel cpu's so nV said screw it, but i also heard it will eventually happen.

StealthHawk
01-03-03, 03:15 AM
i believe that nothing concrete has ever been stated. in other words, don't hold your breath.

Firelord-OCHW
01-03-03, 05:36 AM
Originally posted by DXnfiniteFX
It is a given that the performance of the Athlon XP is slipping behind the P4s. However, they are still managing to keep up mainly because of the incredible boosts attributed to the Nforce.

However, I have heard that the P4 will EVENTUALLY get the Nforce platform under its belt. . . I think it was sometime this year. If that is true, will AMD be able to keep up? Or will Intel continue their performance gap? A P4 Nforce would likely mean a Prescott platform as well. . . and with the advanced hyper-threading of the Prescott with the naturally hyper-threading friendly nature of Nforce performance optimizations, how will Nvidia affect this power struggle? I honestly think they hold the trigger.

Ummm the only reason the P4 does so well is the extreme speeds it can reach, not becuase its a great performer.

The XP does far more per clock than the P4 and put them at THE SAME CLOCK SPEED ie MHz for MHz then the XP would wipe the floor with the P4.

Lets not forget the P4 has 512KB cache vs the 256KB of the XP.... When Barton hits the stores then it will be evened out greatly.

Not only this BUT Oppainter at Xtreme Systems has a score in the Top5 with an AMD XP roughly 1.2GHz SLOWER than it P4 counter parts. That shows something.

Its not all about the chipset the CPU has a lot to do with it as well

ricercar
01-03-03, 01:48 PM
nothing concrete has ever been stated

I thought the NVIDIA CEO said that Intel wants too much money for the P4 FSB license, and NVIDIA isn't willing to pay it. IIRC Intel wanted something absurd, like 33% of what NVIDIA could charge for the entire nForce chipset.

DXnfiniteFX
01-03-03, 03:13 PM
Didn't they say that eventually Nforce would make its way to the P4 playform one day? I seem to recall reading an article about the P4 platform sometime this year. Likely the end, I assume.

StealthHawk
01-03-03, 04:39 PM
Originally posted by ricercar
I thought the NVIDIA CEO said that Intel wants too much money for the P4 FSB license, and NVIDIA isn't willing to pay it. IIRC Intel wanted something absurd, like 33% of what NVIDIA could charge for the entire nForce chipset.
yes. so like i said, anyone telling you that nvidia will make a mobo for P4 is giving you nothing but speculation or rumor.

i still think it would be advantageous for nvidia to get into the Intel market eventually. so i can see it happening. but as i said earlier, i wouldn't count on it.

lagadu
01-03-03, 07:25 PM
the problem would be that the nforce would hardly be a value chipset, but rather a performance one, and in the perforamance segment, intel rules and they indeed make impressive chipsets (no wonder, after all they know their cpu inside out :) ), so we'd have nvidia competing with intel for the performance chipset market, and when it comes to this intel clearly has the "name", it's known for both performance and extreme stability, and this is the market the p4-nforce would aim. i'd say this is one of the main reasons we'll never see a p4-nforce, it'd be a very hard upstream battle with probably little financial gains...

just my 0.02...

StealthHawk
01-04-03, 03:12 AM
Intel was a long time in coming with a good DDR motherboard. it was just a few months ago with the i845PE. competitors like SiS filled the void before that.

on a side note, how can anyone claim that nforce2 doesn't significantly increase performance over previous motherboards, and this IS the motherboard that put AMD back in the running performance-wise. every review i have ever seen has shown this to be true(over the previous champ, the KT333/400)

Riptide
01-04-03, 09:46 AM
Originally posted by Firelord-OCHW
Ummm the only reason the P4 does so well is the extreme speeds it can reach, not becuase its a great performer.
The XP does far more per clock than the P4 and put them at THE SAME CLOCK SPEED ie MHz for MHz then the XP would wipe the floor with the P4.
OMG, we haven't seen this BS argument before.

As a consumer, I don't care if it takes 500 more MHz to make it run faster. Big deal. Whoopee. You think that matters. It doesn't. Noone cares but people out there who are, if you'll excuse the term, fanboy zealots.

It's all about price/performance ratio and whether they get there with more MHz or a different architecture matters absolutely, positively, diddly squat.

DXnfiniteFX
01-04-03, 10:33 AM
Yes, that arguement is weak. I don't care if the Athlon can do 10 times more work than the P4 at the same clock speed, if it doesn't outperform the P4, then it's slower. The whole point is that you don't buy because of which does more work per clock cycle, you buy which does work work, period. As of now, the P4 is reigning champ in nearly all benchmarks by a very large margin. This arguement is flawed and pointless.

Anyways, back to the topic. Toms Hardware themselves said that the Nforce performance increases should have an even bigger effect on the P4. The first reason is obviously bandwidth the P4 can make use of. The second is the chipset optimizations are very friendly to hyper-threading, likely increasing hyper-threading performance.

I also remember that although Nvidia scrapped plans last year and the year before, an article said they would eventually get back to it. I forget where I read that. However, it was sometime this year. I do know, however, that the article should be reliable overall and not some "Inquirer" rant.

lagadu
01-04-03, 11:25 AM
but then there's one thing to consider, the dual channel is nothing new to the P4, not only does rdram do that, but also the i7205 (granite bay) so i think nvidia would need some other selling point...

StealthHawk
01-04-03, 08:17 PM
Originally posted by lagadu
but then there's one thing to consider, the dual channel is nothing new to the P4, not only does rdram do that, but also the i7205 (granite bay) so i think nvidia would need some other selling point...

except that Granite Bay's dual channel performance is weak.

see this thread for more detail: http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=4755

lagadu
01-05-03, 10:33 AM
yeah, it's still a bit behind the 850E that is already months old but it already beats easy the 845PE with ddr333 using 266 and it's one of the best overclockers out there, plus i wouldn't be surprised that if needed intel would allow manufacturers to release a bios that allows asyncronous operation of the memory...

when i first saw benchies of it i was *Very* dissapointed with it, but come to think of it, it's not that bad.

[edit] and the nforce2 dual channel VS single channel also has a rather small performance difference, smaller than the one seen with the 7205.

savyj
01-05-03, 12:24 PM
I'd have to say that the nForce 2 would give the P4 little or no advantage over their current chipsets. Why? Because, myself and some other believe that the last several generations of chipsets for AMD Athlon products have been aimed at the value market, not the high end market. VIA has officially claimed as much. Many (including myself) believe that the chipsets for the Athlon have never truly taken advantage of the performance available in the Athlon package. Until now. The nForce 2 is a "high end" chipset, capable of really flexing the Athlon's muscle. Intel seems to always have a good "highend" chipset available for their processors, therefore adding a P4 nForce2 wouldn't really gain them much.

As much as I hate to point to anything Tom's Hardware has done, in his last VGA roundup it clearly shows an nForce 2 Athlon XP 2700+ ON PAR OR AHEAD OF a similarly equipped i845PE Intel 3.06HT system. Well, well. Imagine that. Decent chipset and AMD's already back in the game.
http://www.tomshardware.com/graphic/20021218/index.html

As for Intel and NVIDIA working something out.
I believe NVIDIA said that if Intel was willing to pop their head out and come up for air on the licensing costs that they would build a P4 nForce2 chipset. Guess what the chances of that happening are?

My question is...will the barton be out before the next ramp up of the P4? Could get interesting if so.

StealthHawk
01-05-03, 06:38 PM
Originally posted by lagadu
yeah, it's still a bit behind the 850E that is already months old but it already beats easy the 845PE with ddr333 using 266 and it's one of the best overclockers out there, plus i wouldn't be surprised that if needed intel would allow manufacturers to release a bios that allows asyncronous operation of the memory...

where are benchmarks showing Granite Bay with dual DDR266 beating i845PE with DDR333? i have not seen any and have not heard anyone claim otherwise until now.

edit: and i should make clear that it beats it in a wide range of tests. i have seen tests where it is slightly faster, the same, and slightly slower. but it should be dominating, as the theoretical bandwidth is higher.

lagadu
01-06-03, 08:24 AM
here it's only second to the 850E, and sometimes beats it in ut2k3 and specviewperf goes 50/50, divx encoding is faster than the 850E

http://www.anandtech.com/mb/showdoc.html?i=174

this is why i say that the GB is #2, but very close to the #1 (850E) and also close to #3 (845PE) but closer to #1 as it sometimes can overtake it...

StealthHawk
01-06-03, 04:48 PM
different benchmarks at different sites which show a more wide variety of tests show a much closer picture, but i guess i will leave it at that.

edit: i guess i should end by saying that in most tests the i845PE and Granite Bay are very evenly matched. in some cases i845PE pulls away by a little, and in others Granite Bay pulls away. you would think Granite Bay would always be at least equal to i845PE as well surpassing it. especially for the cost of the mobo. and that is what my biggest gripe is. if it were more reasonably priced, it would be another story.

lagadu
01-06-03, 07:11 PM
i agree with you, in the end all 3 chipsets are almost the same :)

about the price, yes it's a little too much to pay, the extra 50 are hardly worth it, but this would depend on the user :cool: