Were P3 cpu's better than P4 cpu's?

[lindochico]

Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

 

[RichUK]

With the lower clocked Willamette cores, yes. Or so i had read.

 

[aidanjm]

apparently the Pentium M is based on the same "P6" architecture that underlied the Pentium IIIs. i.e., the Pentium M has more in common with Pentium III than Pentium IV. And I guess now with core duo/ yonah and upcoming conroe etc. Intel has gone back to that P6/ pentium III lineage. So I guess eventually the pentium IV will seem like an aberration.

 

[Ika]

Originally posted by: aidanjm
apparently the Pentium M is based on the same "P6" architecture that underlied the Pentium IIIs. i.e., the Pentium M has more in common with Pentium III than Pentium IV. And I guess now with core duo/ yonah and upcoming conroe etc. Intel has gone back to that P6/ pentium III lineage. So I guess eventually the pentium IV will seem like an aberration.

QFT :thumbsup:

 

[Kakumba]

Clock for clock, the later PIII would beat the early PIV. what was the name of the latest PIII core? I cant remember.

 

[RichUK]

Originally posted by: aidanjm
apparently the Pentium M is based on the same "P6" architecture that underlied the Pentium IIIs. i.e., the Pentium M has more in common with Pentium III than Pentium IV. And I guess now with core duo/ yonah and upcoming conroe etc. Intel has gone back to that P6/ pentium III lineage. So I guess eventually the pentium IV will seem like an aberration.

The Pentium M architecture was derived from the P3, yes, but that doesn?t answer the OP?s question.

 

[coldpower27]

Pentium M has the same philosophies as Pentium 3, high performance per clock, lower power output, but it had many changes to it's core, though the Pentium M does use soem technologies borrowed from Pentium 4 as well. Intel to me has gone back to this philosophy high performance per clock and low energy output.

Core Architecture borrows technolgoy from both Pentium 4 and Pentium M, so I would say it's a hyrbird but it follow the high performance per clock philosphy started by I dunno what Pentium Pro?

 

[coldpower27]

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

Yes a Pentium 3 and at the same clock as a Pentium 4 would outperform it. The Pentium 4's didn't really start to take off until Northwood at 2.0GHZ and beyond...

 

[thecoolnessrune]

IMO OVERALL:

Opteron > Athlon X2 > Athlon 64 > Pentium 4 Northwood > Athlon XP > Pentium 4 Prescott > Pentium 3 > Pentium 4 Williamette.

 

[dmens]

"Better" is a very loaded word since there are so many axes to compare against... I've had multiple day discussions on this very question. But from a work per clock cycle point of view, P3 is better.

 

[F1shF4t]

Originally posted by: thecoolnessrune
IMO OVERALL:

Opteron > Athlon X2 > Athlon 64 > Pentium 4 Northwood > Pentium 4 Prescott > Athlon XP > Pentium 3 > Pentium 4 Williamette.

There u go fixed, even presHOTTS are better than athlon xps, athlon xps were one weird and suckfull cpus, lol i hate the darn things for some reason, i guess i said that numerous times before, and thats why i have a polomino to try it out .

Also since the p4s have such a long pipeline, only when they went beyond 1.7 ghz was it possible to see their advantage over best p3's. Otherwise they were more efficient than p3's, or it would take a 2.0ghz p4 to be equal to a 1.0ghz p3.

 

[Avalon]

Originally posted by: Kakumba
Clock for clock, the later PIII would beat the early PIV. what was the name of the latest PIII core? I cant remember.

Tualatins

 

[BlueWeasel]

I had a P3 1Ghz that hit 1.33Ghz (10x133) purchased from Zap about 4 years ago. Damn, that was a nice chip.

 

[Zap]

Yeah, that was a monster of a Coppermine core, wasn't it? Tualatins had nuttin' on that... except for models with more cache... and heatspreaders... and lower temperatures... nevermind.

Still, that was a rarity to have a Coppermine P3 running at 1.33GHz.

Some extreme overclockers were getting 512k cache Tualatin P3 chips to run upwards of 1.5GHz. With really low latency RAM those would put the smack down on even 2GHz P4 chips.

 

[Fox5]

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

Yes they were, but not by much. The early P4 processors were rushed out (produced on a lower fabrication process than was originally intended, essentially design samples), didn't have adequate memory, and had half the cache the design was supposed to have. That said, the P3s had basically been elipsed by the Athlons in that day, so P4s weren't that bad compared to P3s. It only took like a 1.4-1.6ghz P4 to match a 1.2ghz p3, while it took a 1.8ghz to 2ghz p4 to match a 1.2ghz athlon.

The Pentium M performs much better than the P3 in most things primarily to a much larger and improved cache along with jacking the P4's memory subsystem. In situations which aren't dependent on memory or cache that much, the PM performs almost exactly like a P3. Which is also why the PM wasn't suitable for desktop, workstation, or server usage overall, because there are just too many situations where it would get rocked by a Pentium 4 or Athlon 64, or even an Athlon XP. (it was great for games though, but the actual core lacked any punch, something core duo makes up for by being dual core and conroe makes up for by having more punch than any other x86 processor produced yet)

BTW, in everyone of these threads I must always say how Conroe should not be called a suped up P3 anymore than the P4. The p4 was a P3 made to run at high clock speeds and with the memory subsystem and cache upgraded to support that decision, just because Conroe doesn't have a focus on high clock speeds doesn't mean it's anymore a P3 than a P4 is, otherwise the Athlon is just a K6.

BTW, was a Tualatin a desktop chip? I remember it being primarily marketted as a mobile chip (it was the direct precursor to the PM), and I think, like the K6-3, it was relegated mostly to the mobile market since it came out late and the P4 had already begun its clock speed climb by then.

 

[stevty2889]

Originally posted by: Fox5

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

Yes they were, but not by much. The early P4 processors were rushed out (produced on a lower fabrication process than was originally intended, essentially design samples), didn't have adequate memory, and had half the cache the design was supposed to have. That said, the P3s had basically been elipsed by the Athlons in that day, so P4s weren't that bad compared to P3s. It only took like a 1.4-1.6ghz P4 to match a 1.2ghz p3, while it took a 1.8ghz to 2ghz p4 to match a 1.2ghz athlon.

The Pentium M performs much better than the P3 in most things primarily to a much larger and improved cache along with jacking the P4's memory subsystem. In situations which aren't dependent on memory or cache that much, the PM performs almost exactly like a P3. Which is also why the PM wasn't suitable for desktop, workstation, or server usage overall, because there are just too many situations where it would get rocked by a Pentium 4 or Athlon 64, or even an Athlon XP. (it was great for games though, but the actual core lacked any punch, something core duo makes up for by being dual core and conroe makes up for by having more punch than any other x86 processor produced yet)

BTW, in everyone of these threads I must always say how Conroe should not be called a suped up P3 anymore than the P4. The p4 was a P3 made to run at high clock speeds and with the memory subsystem and cache upgraded to support that decision, just because Conroe doesn't have a focus on high clock speeds doesn't mean it's anymore a P3 than a P4 is, otherwise the Athlon is just a K6.

BTW, was a Tualatin a desktop chip? I remember it being primarily marketted as a mobile chip (it was the direct precursor to the PM), and I think, like the K6-3, it was relegated mostly to the mobile market since it came out late and the P4 had already begun its clock speed climb by then.

Tualatin was a desktop chip, it was the final P3 version. And there were a lot more changes to the pentium-m than just added cache. The netburst architecture is not similar to the P6 architecture even remotly, it's a totaly differant architecuture. Thats like saying a Porche is a Dodge Neon designed to run at higher speeds. You might as well call all of them souped up 286's since they are all x86 based.

 

[WT]

I still have a Coppermine P3, but a rather rare one at that. Its a P3 1ghz 100 FSB chip. Its still in use even, and I often get the notion to Ebay it. I have yet to find any Ebay auctions for that chip, as they are always the common 133fsb chips available, but for the old BX rigs still running, this was the best chip available.

 

[Maximilian]

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

Yep they pretty much mauled the early pentium 4's. The coppermine varient on the PIII reached its limits around 1ghz, and so could only really compete with the 1.3ghz pentium 4, but i believe the pentium 4 was still actually better than it. The tualatins on the other hand included an IHS, 512kb cache for the Tualatin-S and were built on the 130nm process, probably to test it out for the later pentium 4's, similar to the situation with cedar mill on 65nm. The tualatin chips ran at 1.4ghz max, and this was roughly the equivilent of a p4 1.8ghz, although i think the p4 would win in the gaming benchmarks though, because of its much higher FSB or SSE2 or somthing. I remember it beating the PIII as i researched this when building a dual PIII system.

So overall tualatin 1.4ghz = wilamette 1.8ghz. Coppermine shouldnt really be part of the equation, the p4 was better than it. Perhaps not the 1.3ghz p4 but certainly every other speed after that.

 

[piasabird]

I still have a running PIII (Celeron) with 256kb L2 cache 1.2gig Processor. This processor competes well with even low end P4 as well as Athlon XP processors. One reason is that the PIII uses less power and runs cooler. However, Intel made many versions of the PIII some of which could use the 133mhz bus and RAM with larger cache sizes.

Intel just shipped some Single Core 5.5W Yonah-Based processors about 1 Gig built on the 65nm die process. These are really just reworked PIII Core processors. It is like they took the PIII Core and improved it while still making the P4 Core processors. While designed for notebooks they might even be good for gaming with a 2 Meg Cache. Some companies like A-Open build motherboards for the P4M processors. They dont make great gaming rigs, but they run cool and quiet and use very little power.

I think Shuttle makes some cubes that run on P4M as well.

 

[Henny]

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

The latest of the previous generation is almost always as good or better then the first of the new generation.

examples:
Late 486's were better then early Pentiums. (Pentium 60 was a dog)
Late Northwoods were better then the first Prescotts. (Prescott was a dog)

 

[Zim]

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

At the same clock speed, the P3 was faster... but the P4 later scaled to higher clock speeds leaving the P3's in the dust. I'm talking about desktop cpus here... not mobile ones.

 

[zephyrprime]

The early P4 was a worse performer than the P3 on both a clock to clock basis and on an absolute basis.

 

[Auric]

Originally posted by: WT
I still have a Coppermine P3, but a rather rare one at that. Its a P3 1ghz 100 FSB chip. Its still in use even, and I often get the notion to Ebay it. I have yet to find any Ebay auctions for that chip, as they are always the common 133fsb chips available, but for the old BX rigs still running, this was the best chip available.

I had an uber-rare 1100/100 (OEM only) and o'erclocked it to 1234 (and at .1V < default) on a BH6, cooled only by the PSU exhaust fan (ahh, those were the days). However, I did not keep it long before eBay-ing since although it was the best wi'out a socket adapter or mod on the i440BX, the Tualatin outperformed it by 15% at the same clock and of course was able to clock much higher.

 

[Fox5]

Originally posted by: stevty2889

Originally posted by: Fox5

Originally posted by: lindochico
Hello guys, first time poster here!

I'm collecting info. about the fact the Pentium 3 procs. were better performers than early Pentium 4 procs. If this is true, please let me hear from you.

Any comments will be appreciated...

Yes they were, but not by much. The early P4 processors were rushed out (produced on a lower fabrication process than was originally intended, essentially design samples), didn't have adequate memory, and had half the cache the design was supposed to have. That said, the P3s had basically been elipsed by the Athlons in that day, so P4s weren't that bad compared to P3s. It only took like a 1.4-1.6ghz P4 to match a 1.2ghz p3, while it took a 1.8ghz to 2ghz p4 to match a 1.2ghz athlon.

The Pentium M performs much better than the P3 in most things primarily to a much larger and improved cache along with jacking the P4's memory subsystem. In situations which aren't dependent on memory or cache that much, the PM performs almost exactly like a P3. Which is also why the PM wasn't suitable for desktop, workstation, or server usage overall, because there are just too many situations where it would get rocked by a Pentium 4 or Athlon 64, or even an Athlon XP. (it was great for games though, but the actual core lacked any punch, something core duo makes up for by being dual core and conroe makes up for by having more punch than any other x86 processor produced yet)

BTW, in everyone of these threads I must always say how Conroe should not be called a suped up P3 anymore than the P4. The p4 was a P3 made to run at high clock speeds and with the memory subsystem and cache upgraded to support that decision, just because Conroe doesn't have a focus on high clock speeds doesn't mean it's anymore a P3 than a P4 is, otherwise the Athlon is just a K6.

BTW, was a Tualatin a desktop chip? I remember it being primarily marketted as a mobile chip (it was the direct precursor to the PM), and I think, like the K6-3, it was relegated mostly to the mobile market since it came out late and the P4 had already begun its clock speed climb by then.

Tualatin was a desktop chip, it was the final P3 version. And there were a lot more changes to the pentium-m than just added cache. The netburst architecture is not similar to the P6 architecture even remotly, it's a totaly differant architecuture. Thats like saying a Porche is a Dodge Neon designed to run at higher speeds. You might as well call all of them souped up 286's since they are all x86 based.

I'm just saying that conroe is no more a P3 than the P4 is a P3, the cores are both about equally different from a P3, and for the chip as a whole conroe has even more differences from a p3.

 

[Blain]

Originally posted by: Zap
Some extreme overclockers were getting 512k cache Tualatin P3 chips to run upwards of 1.5GHz. With really low latency RAM those would put the smack down on even 2GHz P4 chips.

I played with the Coppermine's, but never made it over to the Tualatin's.
They do sound like very nice CPU's. Especially since they could pull off SMP.