Pentium 4 SUCKS at every programs.. BUT QUAKE 3!!!.. haha... from TOm's.. I knew it!

[j@cko]

from the update review, it shows P4 is slower than both P3 and Athlon in encoding Mpeg 4.... and any other tests.. but Q3..
what a shame..

 

[Dulanic]

#1 Include a Link
#2 This was just posted 2 other times if you looked at all.
#3 I would have no damned clue what you where trying to say if I had not already read the review.

 

[Wingznut]

Jack = Hardware's long lost brother?

 

[PliotronX]

So does this mean everyone who plays Q3 must pay the $1200 for a new P4?

 

[erub]

o hell yes it does - how else are ya gonna get 220 FPS at 640x480?!? Uhmm..anybody have an idea of how many fps the human eye can reigster? Well, if your spending that much on hardware, you probably wont be playing at 640x480, and at 1024x768 or greater the video card will be the bottleneck...so...

 

[Noriaki]

I'd just like to point out that Quake3 is highly SSE optimized, and therefore it shows the potential of the Pentium4, if they can convince everyone to optimize heavily for SSE.

I still think it's too damn expensive and I don't buy chips based on "potential future performance from this core", so I think the P4 is useless...but I wanted to point out the optimizations none the less.

 

[brian_riendeau]

Would someone, anyone, please show me where it says Quake3 is SSE optimized? The NVIDIA OpenGL drivers have SSE optimization, but Quake3 itself?

 

[BFG10K]

Erub:

Uhmm..anybody have an idea of how many fps the human eye can reigster?



Noriaki:

I'd just like to point out that Quake3 is highly SSE optimized, and therefore it shows the potential of the Pentium4, if they can convince everyone to optimize heavily for SSE.

Quake 3 is not SSE/3Dnow! optimised. The reason why the Pentium 4 does so well is because Quake 3 loves a high memory bandwidth and in this case it's being provided by the P4s high clock speed and quad-pumped 100 MHz FSB.

 

[Noriaki]

Quake3 isn't SSE optimized? Hrmm I've heard that many times. Though I suppose that doesn't make it true. I stand corrected.

 

[Leo V]

Any professional application that I've used or plan to use requires a strong floating-point unit; it's little wonder an Athlon 600 was about twice faster than my P3/450 in several of them.

Extra flavors like MMX can speed up simple tasks that are easily parallelized. Videocard analogy: your videocard accelerates simple, strictly limited pixel operations very well; but for any expandable image processing, you're stuck with software mode, whose versatility is unmatched. Likewise, parallel MMX only suffices for basic tasks, perhaps doing software T&L (which videocards are taking over anyway); for any versatile application, you're back to your FPU.

Intel's inclusion of their MMX2 playtoys is no excuse for their unacceptably weak FPU. If it's as weak as Tom suggests, I entirely agree with his conclusion--it's more of a fashion fad than a workstation chip.

 

[ragiepew]

If q3a is not SSE optimized then why didnt the Athlon spank the p3 in q3a since you state its it depends greatly on a fast bus and memory clock. If it was all about the bus, then the ev6 + pc133 ram should on the athlon should give it an advantage over the p3 and its pc100-133mhz bus.

alin

 

[Leo V]

By the way, I remember John Carmack was interviewed long ago, saying that Quake3 does have P3/MMX or SSE (?) optimizations. The P3 was fairly new then, and the Athlon practically nonexistent--so we cannot blame ID for this.

[Edit:] I could be wrong from what Jukka says. But I do remember reading the above.

 

[jpprod]

I'm pretty sure it's the L2 cache which makes both P4 and P3 exceed (equally clocked in P3's case) Athlons in Quake 3 despite Athlons much better x87 floating poing performance. Quake III is very bandwidth-hungry, but performance in it doesn't solely depend on bandwidth as Tbird Athlon with it's DDR FSB doesn't edge out a Coppermine P3. Data sets in Quake 3 though huge, could be extremely uniform and well-organized thanks to Carmack's careful coding so that there are very few cache misses.

Now... all we'd need to prove this theory would be a Athlon variant with low-latency L2 cache on a 128 or 256 bit bus. Are you listening, AMD

 

[Whitedog]

Pentium 4, the most Advanced paperweight in the world.

LMFAO!

 

[JackMDS]

By the spirit of Thanksgiving.

It is relay a ?shame?.

Think of all the hungry, and the sick, that could be ?saved? if only P4 will do another 10FPS while playing Q3.

 

[jpprod]

Quake III is not SSE-optimized according to sources I saw at Ace's Hardware - some guy (see, I've got a foolproof source ) had scanned through the Q3 EXE binaries without finding any SSE opcodes. Only SSE optimizations playing Quake 3 come from display drivers.

[edit: please completely ignore this remark, but I just noticed that I've made 1990 posts so far - just ten more and I've neffed my way all the way up to an platinum member! I feel like such a big man now!]

 

[pdo]

Leo is a girl?

 

[Whitedog]

<<I'm pretty sure it's the L2 cache which makes both P4 and P3 exceed (equally clocked in P3's case) Athlons in Quake 3>>

Uhmmm, No. You ever hear of SSE? Quake III is coded to take full advantage of these 3D enhancment instruction sets, which Athlons don't have. Plus, Quake III has poor, if any 3Dnow! coding into it.
Intel slid a few $$$ to Id I heard for this. But don't hold me to that.

[edit] Well, I guess you Have heard of SSE , but I don't buy that bit from Ace's. Id said right out that Quake3 was optimized with SSE. [/edit]

 

[Diesel21]

Ok, Here is what the main man Carmack has to say about P4 and Quake3!

quote (from www.dailyrush.dk)

If I had to hazard a wild guess, I would say that it is because much of Q3's structure is broken up into nice separate phases, like game, cgame, front end, back end. The P4 relies on a trace cache for best performance, and it would have effects in some ways like a smaller instruction cache. It may turn out to be something unrelated, like a better bus interface helping the graphics driver more.

Carmack does not even mention SSE! If he does not mention it I am not so sure its in there...

 

[Leo V]

paldo, don't worry. I'm just a straight guy exploring my femininity--and what better place for that than Anandtech? (OK perhaps the Sharkyextreme forums, but I don't swing that way!

 

[Priit]

SSE or no SSE in Quake3, P4 sucks anyway. I just wonder how long it takes to calm down and not bash P4 5 times in hour

 

[Dulanic]

Don't you guys research before you make comments? Tom clearly showed almost all of the P4s speed in Q3 came from the massive memory bandwith. Once we see Crush 128bit dual channel memory bus and DDR2100 I expect the Athlon to be close to the P4 in Q3.

 

[j@cko]

HERE IS THE LINK>> IF YOU WANT THE LINK

 

[Rectalfier]

P4 1.5 Ghz is the fastest Quake3 CPU out there, however considering it is running at 1.5Ghz it still does not compete clock for clock. A 1.5Ghz Palamino would wipe the floor with poor P4.

 

[Leo V]

Speaking of Palamino, I'm quite dissappointed by its continuing to use the .18u process. I thought AMD would shrink the die, finally putting an end to the infernal heat. Instead, they sold out to the Devil and plan to make Palominos just as satanically hot/power-hungry. The shame, the shame!