Wow @ P4's performance (lack of)

[MisterMaya]

Hi, I don't post to the CPU guide much, but I just had to post and say *wow* I just experienced first hand how bad the performance per clock of the P4 sucks. I've been running an Athlon 850 since they were new, and my grandpa just unloaded his old P4 1.4GHz on me, so I got a Quadro4 750xgl in it and anxiously tried some Maya rendering of equivalent scenes between the two. The P4 barely inched ahead of the Athlon. Granted it wasn't a well-rounded perfect THG-quality benchmark, but I was kinda expecting some performance gain since the clock was just about 2x, but there was no performance to be had whatsoever.

I'd estimate the performance gain from an Athlon 850 to a P4 1.4GHz in Maya renders with raytracing and refractive obj's at about 10%. Absolutely pitiful, IMO.

Just my little rant of disappointment out to people who care 'bout this stuff... if anybody complains about me doing inaccurate benchmarks, I'll never buy Jello from them again. ;-D


-- Monkeys? What does this .sig have to do with monkeys? --

 

[lhgpoobaa]

yep. the first P4's were dogs.
hot, not much faster clockwise than athlons and slow.

things have improved alot since. the architecture of the p4 core allows speed to scale up alot.
the northwood is much better. faster, cooler (thanx to the dieshrink) and has more onboard cache that helps significantly.

also... what motherboard was that P4 on? i bet it was a SDRAM system. P4+sdram is NOT a good combination. ever.

<font color=green>Proud member of THG's</font color=green> <font color=blue>Den Of Thieves</font color=blue> :lol:

 

[eden]

You have got to have read some THG reviews before ending up in this forum. If not, then I suggest you do RIGHT NOW, to see where the P4 is, and why a 1.4GHZ P4 is like the worst thing to ever buy...

Oh wait, Dell has 1.3GHZ+SDRAM, scratch that out!

--
Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

 

[FatBurger]

I just experienced first hand how bad the performance per clock of the P4 sucks

Per clock doesn't matter when you're at 2.5GHz.

Granted it wasn't a well-rounded perfect THG-quality benchmark

So it was actually accurate?

I'd estimate the performance gain from an Athlon 850 to a P4 1.4GHz in Maya renders with raytracing and refractive obj's at about 10%. Absolutely pitiful, IMO.

I agree, but you can't judge an entire processor on the first one released. The performance is quite a bit better now.

<font color=blue>Hi mom!</font color=blue>

 

[eden]

Actually jumping from a 1.6A to a 2.4GHZ, did 50% higher performance, the scaling on the P4 is now very realistic.

--
Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

 

[FatBurger]

I'm sorry, I'm not sure what you're saying.

<font color=blue>Hi mom!</font color=blue>

 

[eden]

I was saying how much now the performance really goes up in a good scale. So a 3.2GHZ P4 on 533MHZ FSB might be twice the performance of a 1.6A!
I am not sure if it will continue that way, but since the 2.4 is 50% more performing than the 1.6, that's pretty encouraging.

--
Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G!

 

[FatBurger]

Oh, right. I just hope that memory and bus speed can keep up with the clockspeed, or else it won't scale that well for long.

<font color=blue>Hi mom!</font color=blue>

 

[MisterMaya]

Eden, I've been reading THG since I came to college four years ago. I read the old reviews about voodoo2 SLI vs. TNT2. Like I said, it was a *rant* not a *review*

-- Monkeys? What does this .sig have to do with monkeys? --

 

[MisterMaya]

dunno what motherboard it was on, just a gateway 2000 system my grandpa unloaded on me when he upgraded. I told him I wouldn't buy it from him because it wasn't worth the market value to me, and I already had my 850, so he didn't feel like selling it and just gave it to me.

-- Monkeys? What does this .sig have to do with monkeys? --

 

[MisterMaya]

>> Per clock doesn't matter when you're at 2.5GHz.
(how do you do your fancy quoting thing Fatburger?)

Well, you gotta pay for that extra clock, that's my beef... aren't p4's still slightly more expensive than same-clocked athlons? Like you guys said, the scaling is better now, but just think if you bought a 2.2GHz proc and then found out it was *really* a 1.4GHz, and I could have had a 2.0GHz that performs like a 2.0GHz for a couple hundred less from AMD. (granted of course again that the scaling *is* better now)

The big thing that had me so disappointed is that I was hoping I'd have some actual proc power to crank out frames with, but in fact just had another obsolete system just like my 850. I did read the p4 vs. Athlon reviews when it first came out, but it was still a kick in the pants to see it firsthand.


-- Monkeys? What does this .sig have to do with monkeys? --

 

[FatBurger]

My 1.6A was $140 (now around $120) and it's running at 2.5GHz. If you don't overclock, then Athlon is by far the better deal. If you do, then Intel is currently a bit better deal.

And click on the FAQ link on the left to see how the markup works

<font color=blue>Hi mom!</font color=blue>

 

[juin]

Consider that intel will move FSB to 166 after 200 and after ??? maybe we will see BBUL packing appear at the same time.

cheap, cheap. Think cheap, and you'll always be cheap.AMD version of semi conducteur industrie

 

[slvr_phoenix]

and I could have had a 2.0GHz that performs like a 2.0GHz for a couple hundred less from AMD.

It sounds like you're not very up to date on AMD then. If you were say, to buy an Athlon XP 2000+, it is <i>actually</i> a 1.67GHz chip.

Which, in a way is better, if you think that a 1.67GHz AMD chip is competing against a 2.0GHz Intel chip. Yet at the same time, you can't help but feel slightly miffed that an Athlon XP 2000+ is only 1.67GHz. Imagine how much it'd whoop up on a 2.0GHz P4 if it were actually a 2.0GHz Athlon.

Anywho ... back to sanity. <font color=blue>Blue Wizard</font color=blue> now has <font color=red>reflective shot</font color=red>. Ooooooh!


Tech support said take a screen shot.
Putting it down with my .22 was the humane thing to do.

 

[FatBurger]

You're in a strange mood this afternoon. Gauntlet is a good game though.

<font color=blue>Hi mom!</font color=blue>

 

[nanyangview]

How do you spell BULLSHIT Fat burger?
You see that chess match rite NOW? @ http://www.heise.de/ct/schachduell/ The P4 XeonCRAP is having a 1.6Ghz advantage compared to the Athlon MPs but the CRAPP4Xeon is losing.
Once again how do you spell BULLSHIT?

 

[FatBurger]

I can spell "troll" very well, thank you.

<font color=blue>Hi mom!</font color=blue>

 

[MisterMaya]

yeah, I've been kinda talking in past terms because that's the topic I started the thread on, talking about my 850 being practically the same performance as the 1400. I only bother getting up to the minute information when I'm shopping for a new system. That stuff changes too fast, and besides I was making a hypothetical analogy; sorry if it wasn't accurate to modern performance expectations.

still, if an actual 2GHz Athlon performs at 84% the speed of an actual 2GHz P4, I'd say that's at least ballpark... this P4 of mine was performing at 60% of what its clock said it should (assuming the Athlon was at 100%), and that's just not right. Even if clock speed isn't everything, it should at least be something.

but anyway, I'm going back to the graphics and sound forums where I belong. I just wanted to post a little rant about my p4 1400 sucking so bad and pissing me off.


-- Monkeys? What does this .sig have to do with monkeys? --

 

[FatBurger]

if an actual 2GHz Athlon performs at 84% the speed of an actual 2GHz P4, I'd say that's at least ballpark...

It doesn't really matter, since there is no 2GHz Athlon.

this P4 of mine was performing at 60% of what its clock said it should

If the clock speed was 1400MHz, then that is correct. Intel has never made any promises about per clock performance in regard to other processors.

Even if clock speed isn't everything, it should at least be something.

And it is. A 2GHz P4 still outperforms a 1.4GHz P4.

<font color=blue>Hi mom!</font color=blue>

 

[MisterMaya]

Found some papers on the system - no sdram, it was pc800 rdram, 256MB of it, and no HD access during the test render btw. Oh, and found out it's actually a 1.5 not a 1.4, not that that makes much difference.

-- Monkeys? What does this .sig have to do with monkeys? --

 

[nanyangview]

Basically Intel's architecture is INFERIOR and crap no matter what their clock speed it. FULL STOP.
AMD is really THRASHING Intel 7-10 rite now.
HAHAHAHA!!

 

[FatBurger]

Not really, the 1.4 and 1.5 both sucked almost equally hard :frown:

<font color=blue>Hi mom!</font color=blue>

 

[FatBurger]

By the way, how does it feel to have the lowest possible user rating over at Anandtech's forums?

<font color=blue>Hi mom!</font color=blue>

 

[lhgpoobaa]

nods. well its probably better than what i expected then.

the P4 really doesnt shine unless its got 500 extra Mhz under the hood compared to anything other processor design.

also, the p4 design seems optimised better towards certain applications. video editing and the like is good with the P4, lots of memory bandwidth, and anything with SSE2 enabled will run well too.
The P4 is NOT very good at floating point intensive applications though, especially in the adsence of SSE2 acceleration. maybe that is the 'problem' you are experiencing. whoknows? would help if we knew what kind of thing your program uses most... FPU? ALU? SSE2? lots of membandwidth etc



<font color=green>Proud member of THG's</font color=green> <font color=blue>Den Of Thieves</font color=blue> :lol:

 

[AmdMELTDOWN]

AMD is really THRASHING Intel 7-10 rite now.
HAHAHAHA!!

LOL! yeah but AMD is sweating at the table while fritz is cool, calm and collected, many more games left, you damn fool.


"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"