What happened to the P4

[Divexo]

P4 is dead - ag shame hahahaha

http://www.tomshardware.com/business/20050531/computex_day1-11.html

Thank God i got myself a 939

[Thallium]

i have just moved from a prescott to a 939 2 days ago.. :)

[[B.Inc]=_sNaKe_=]

Ja, saw the article the other day lol! Shame man, over-priced AND outdated heap of junk? Woah... :D

[LeG@cY]

ROFL !!! GO look again The Pentium M CPU Fuxors up the FX55 ag shame & it is only running @ 2.5Ghz

Close your doors AMD

http://www.tomshardware.com/cpu/20050525/index.html

[Ph@ntom]

Frankly, I couldn't care whos more dead that who. During the past 3 years we've seen the Prescott jump from just 3.2 to 3.8..... AMD followed a similar path and the main focus was rather only core adjustments and extra features. And dual-core hasn't been the answer for all type of apps...
Swak... Time they pull their fingers outta their asses, stop making excuses, and for once give us faster chips!! :P

[Ph@ntom]

Yeah Leg, I wouldn't mind getting a Pentium M myself. That thing is just killer! I still remember the days they compared the P4 against a P3 clock to clock. Only apps optimised for the P4 ran faster, yet older apps, suffered big time. Glad I never touched a P4 in my life:

- its tiny tiny 486-style L1 cache. 8K was the smallest in 10 years... It forced software developers to write code that only worked with tiny amounts of data at a time.

- Crippled instruction decoder; which could only decode one instruction at a time, not 3 as with the P3 and Athlons. Again not since the days of a 486 has a chip been crippled like this

-Useless trace cache replacing the code cache. Too small in size. It limited the amount of code that the P4 could cache and puts additional strain on the crippled decoder

-Not enough floating point units, making the Athlon beat it in math intensive tasks and games.

-No L3 cache as promised in its design plans. Pretty much explains why even a G4+ trounce it in Photoshop.

-Distribution of execution units not optimal meaning that too many simple integer units sit idle alot, but not enough floating point units

-The shift, a very basic operation of the CPU has been crippled by means of Intel eliminating the barrel shifter.

-The partial register stall was fixed, yet it was even worse than the original problem.

-Not enough work performed per clock cycle, meaning that the raw efficiency of the chip is worse than the P3 and Athlon, and now the newer Pentium M.

Shame on Intel for not continueing the P3 like they should have!

[GunJitsu]

Who cares about the CPU, most new games run fine on a celeron 2Ghz. Your gfx card is the most powerfull device in the pc anyways.

[hai2u]

Frankly, I couldn't care whos more dead that who. During the past 3 years we've seen the Prescott jump from just 3.2 to 3.8..... AMD followed a similar path and the main focus was rather only core adjustments and extra features. And dual-core hasn't been the answer for all type of apps...

Wtf are you actualy saying here?

[Ph@ntom]

Wtf are you actualy saying here?
Intel still hasn't reached 4Ghz and AMD hasn't even gone past 3Ghz yet. Ony extra features such as SSE3, larger L2 caches etc. gave extra performance. Mhz increments has gone up at a snail's pace...

Due to thermal dissipation issues, processors can no longer be given substantial clock speed improvements.

[hai2u]

Increasing the clock cycle rate is by far the least effective way to achieve increased performance.

Go and look at the CRAY CPU's, R1200, SPARC and and the like.

[Divexo]

yes the PIII is da man. BUT Intel moved away from this technology and gave the public some inferior crap. Sais a lot for them.
BTW I agree with Karney (for once) :)

[Gamos]

With AMD firmly taking the gaming crown and the Nfoce4 giving SLI to AMD long before Intel was forced to approach nVidia, you have to wonder about the state of Intel's P4 roadmap. That article is very interesting speculation. After years of pushing the Mhz, Intel finally met the proverbial brick wall. Their 64bit tech are critiqued to be an afterthought based on AMD's algorythm. Their duel core processors are described as a less elegant and with low clockspeeds, offer nothing to current single core buyers. Add an incredibly rediculously expensive EE chip, that still gets clobbered by AMD, and you have to wonder - what now?

I am not one for brand loyalty, but it does look increasingly bad for Intel.

Still, they rule the marketshare...

[hai2u]

Still, they rule the marketshare...

They make much more than CPU's - like leading gigabit range network interface cards, fibre GBICs etc, routers, PHY chipsets and other network microprocessors and lots of stuff used in VOIP.

Also, their Pentium M range has alot to offer - namely a stable chipset.

[Terror]

Intel is all i know - whats the average dissapointment rate of manufacturing errors made by each party. Wich is the most reliable is the real question! More ghz might not always mean better CPU- its got alot more to do it that that- if u dont believe me - ask google!

[Thallium]

- its tiny tiny 486-style L1 cache. 8K was the smallest in 10 years... It forced software developers to write code that only worked with tiny amounts of data at a time.


Where the hell did you get that from ?? even if there is NO L1 cache. it does not effect execution.. What you smoking??

[Ph@ntom]

I aint smokin shit. Why do you think the P4 sucks clock per clock against a P3 or Athlon? Go research its flaws.
Most of that technical data comes from Anandtech and emulators.com, just to name a few... The list goes on.

[Thallium]

so in other words, all written code is slow, just for the P4.. mmm, i never knew that..

smaller L1 cache will cause a slower execution, not less.. it does not work that way.. that statement is incorrect

[[B.Inc]=_sNaKe_=]

ROFL !!! GO look again The Pentium M CPU Fuxors up the FX55 ag shame & it is only running @ 2.5Ghz

Close your doors AMD

http://www.tomshardware.com/cpu/20050525/index.html

My dear dear friend... read more carefully ...:

Meanwhile, AMD should begin preparing a suitable answer to Intel's upcoming accelerated 65 nm dual-core processor, code named "Conroe."

http://www.tomshardware.com/cpu/20050525/pentium4-21.html

[Divexo]

Meanwhile, AMD should begin preparing a suitable answer to Intel's upcoming accelerated 65 nm dual-core processor, code named "Conroe."

hmmm - dude if Intel screw up the 65 nano meter process like they did with the 90 Nano meter then you are in for a big shock. The Intel 90nm cpu's use more watts then the 130nm even @ idle.

[[B.Inc]=_sNaKe_=]

hehe...ok, I never liked Intel anywayz...:D

[DawEEd]

Quite interesting that Intel had to go back 4 years in their technology to find sumthing that works. So what's excuse gonna be to the millions of people they suckered into buying their defective, in-effiecient junk. "We just wanted to make some money for our next range which hopefully might work.", most probably followed by a statement like "We didn't lie to our customers, we just didn't know betta."

I think that was the purpose of why Div started the thread, was just to show Intel showing their telkom colours. As Gamos said u really have to wonder about the Intel's road map. I remeber when AMD started testing their 64bit technology before they shipped it, Intel was saying their is no point to it, 64 bit technology was not required for the next 2 decades, four years later they tried to incorporate it into their flawed design. AMD seems to have a very good road map for now.

Quite funny to think that intel will now be asking you to upgrade from pentium 4 to pentium 3. And if they count like that (1,2,3,4,3), how do their processor do it?

Hopefully Intel do sumthing good with their pentium 3 technology, like start bringing some competition back on the market. Or VIA should pick up the pace on the developemnt of their CPU, if they still working on it.

At least the GPU segment is getting liberated a bit now. No longer gonna be a 2 horse race.

[<A>FluX.v2.0]

No longer gonna be a 2 horse race.

so who are the other horses? XGI? S3? intel (rofl)?

I would love to see these "contenders" catch up to nvidia and ati.

[DawEEd]

In my opinion XGI and Matrox, are "the best of the rest" for now, and XGI have sumthing cheaper and competitive in the nvidia6600 range. Matrox might be joinging the mass market soon too. I'd say give them about 2 years more and things should be getting very competitive. And Who in his right mind would buy an Intel VGA card?

[Shrimp]

People in the same mind as would buy an AMD or ATI?

[joke tag]

[<A>BioGizzard]

rofl, baai p4, send me a postcard from the junk yard....:P

[DawEEd]

ROFL! Shrimpy, Funny! I'd love to see the person who buys a AMD CPU and puts a Intel VGA card on it. He's most probably asking to be stoned to death. But then again the machine wouldn't even work. THE CPU would feel so insulted about having such rubbish in it, it wouldn't start up.

[sneaky]

WTF is a P4???

[<A>FluX.v2.0]

http://graphics.tomshardware.com/graphic/20020514/

interesting

but the arcticle seems like its from 2002

[Ph@ntom]

http://www.tomshardware.com/cpu/20050525/index.html --> keep in mind that these are OC'ed PentiumM results, and none of the P4's or AMD's were OC'ed as well.

The power consumption is the real plus here, and I wonder by these kinda advantages, what Intel will scale the chip up to. I'd get one if there are sufficient supply of a number of mobo's to choose from and the chip itself is more or less on par with comparitive Athlon pricing.

[Shrimp]

My Pentium M notebook plays Chopper Challenge more smoothly than my P4 at home... And the P4 is 1.1GHz faster :P

[Macera]

lo0l i stay wiht amd