the P4 is dead!

[Rigoletto]

Look over at:
http://www6.tomshardware.com/column/00q3/000903/bb-03.html

the P4 is going to struggle from the start against DDR Athlons. Especially since Intel is going to be greedy and start forcing RAMBUS on users again (groan).
All this strong AMD competition is gonna be great for prices! Celerons increased by 177 MHz in the last 7 months too, and it's still beaten by Durons.

 

[DaveJ]

Linikfied...

 

[KarsinTheHutt]

Rigoletto -

This has been around for awhile. Anway, its hard to say that its "dead" with no benchmarks from reliable 3rd parties (ie AnandTech, Sharky, Tomshardware), no user reports, and no systems available.

I've never seen so much FUD in my entire life - not just on this site, but practically all message boards and sites from Tom's to ZdNet. (Excepting the launch of Athlon 15 months ago). Please stop FUDging until we get some hard evidence - not just rumors or rigged IDF demonstrations.

 

[Rigoletto]

The bad news about IPC is extrapolated from Intel's own figures. Intel are not known for their great modesty or integrity either. What Tom's Hardware writes seems logical enough. What do you think is going to change things, exactly? Any specific ideas?
P4 is going to cost more than Athlon.
DDRAMBUS is going to cost more than DDSDRAM.
Athlon could still be faster than P4!
So who ya gonna choose when P4 arrives?

 

[KarsinTheHutt]

"Intel are not known for their great modesty or integrity either"

As I recall, Intel dumbed down P6 back in '95, and really socked it to PPC and other RISC chips (at least in integer performance) when it was finally released.

"So who ya gonna choose when P4 arrives?"

I'll choose whichever system has the best combination of stability and price/performance regardless of who makes the CPU. Which system will that be? WHo knows? Anyone who makes the choice now can either travel in time or is an arrogant fool.

Edit: Right now I think I'd either go Duron 600@900 or P!!! 700e @ 1 GHz.

 

[Rigoletto]

What do you mean about Intel dumbing down P6 in 95? Do you mean "underplaying"?
I'm just gleeful Intel is going to struggle against AMD still! The competition is great!

 

[KarsinTheHutt]

Yes. I remember reading this stuff about IPC and pipelines and how P6 was gonna get screwed over with a 12 stage pipe (PC Magazine can't remember which issue). Seems ok to me.

BTW - I'm gleeful too about competition I love watching those prices fall!!!

 

[Sir Fredrick]

I remember reading a long long time ago that the PII performed worse when clocked the same as a Pentium MMX

 

[Adul]

It was the Pentium Pro that performed worse the the pentium mmx

 

[Rand]

It's hard to say something id dead when we don't even have any reliable benchmarks on it's performance.

 

[Rand]

It's hard to say something is dead when we don't even have any reliable benchmarks on it's performance.

 

[Chupa420]

Yep! It's dead.

 

[optoman]

I just hope that I can overclock that mother. Seriously, we can't judge the performance until it actually hits the shelves. Maybe the software that comes out when the P4 comes out will kill anything we have ever seen. Intel has a way with Micro$oft. :disgust:

 

[Pariah]

The Pentium Pro was slower than the P5 only in 16bit software, which unfortunately for intel was the majority of software at the time. The software industry migrated to 32bit software slower than Intel thought. In 32bit software, the Pentium Pro was clearly head and shoulders above the Pentium.

 

[BFG10K]

I remember reading a long long time ago that the PII performed worse when clocked the same as a Pentium MMX

PII = Pentium Pro core + MMX.

It was the Pentium Pro that performed worse the the pentium mmx

For 16 bit apps perhaps. But the graphics performance especially in games was much better due to write combining. Plus the 256K L2 cache running at half the CPU speed also helped a lot as well.

The Pentium Pro was more powerful than the Pentium MMX.

 

[jmcoreymv]

Of course the P4 is going to suck, the benchmark programs werent made with compilers that were written to run on the p4.

 

[Midnight Rambler]

Exactly.

Plus calling Uncle Tom is reliable is a huge stretch. Most people don't even bother with his drivel anymore.

 

[jpprod]

PII = Pentium Pro core + MMX

This is not quite right, PII also had much-improved 16bit performance. Ppro lacked something called stack register cache (correct term?) which all other modern x86 CPUs had, because at that time Intel didn't feel 16bit performance was important. How wrong were they proven

 

[Remnant2]

jmcoreymv: To this day, VERY VERY few benchmark programs have any optimization at ALL for the Athlon. About the only benchmark that tries to optimize for it is Sandra 2000 and 3dmark 2000 (and only a bit at that).

So, does that mean the Athlon sucks in performance? A new CPU has to run existing software well. bottom line.


In regards to the topic of this thread : The P4 is definitely NOT dead. We don't know what Intel is sandbagging, but shoot guys, they aint stupid -- stumbling around like a drunk yes, but they didn't become the leader in the CPU market by acting stupid. I'm an AMD fan, but I fully expect that the P4 is going to give AMD a serious run for its money for the high end crown, finally. With clockspeed, if nothing else.

 

[OneEng]

The P4 is not dead. This processor will sell because the clock speed on the box is high and it says "Intel Inside".

What is most likely true is that the AMD Mustang and Palamino will outperform the P4 by a signifigant margin and at a reduced cost. This does not mean that AMD will be more successful than Intel. Marketing has much to do with the success of a processor.

Just look at marketing Intel has already put in place for the P4 ... "Netburst technology". I mean, be serious. The same old "Intel will make your internet experience come alive" line of BS ..... and people eat it up.

All Intel needs to do is advertise with a good line like this:

Don't just make her moan, MAKE HER SCREAM ... with the Pentium4 from Intel

People will flock in droves to get the P4 hoping to improve their sex life!

 

[AndyHui]

jpprod: it was a segment register cache.

 

[jpprod]

Andyhui: Yeah that sounds about right, thanks

 

[BurntKooshie]

For all you RAMBUS bashers: Why is API (the alpha processor guys) using 8 rambus channels for their 21364? BECAUSE ITS DAMNED FAST. They just don't implement it the way that Intel does. They allow it to have more pages open per bank. They also massively reduced the latency issue by integrating the memory controller onto the die of the processor. It will have 12.8(!!!)gbytes/second of bandwidth. With nice latency. So RAMBUS isn't terrible technology. The way that Intel has been implementing it is!

The Penium IV will get a greater benefit from RAMBUS than the Penium III did - remember, the FSB was limiting the bandwidth of the memory because the memory operates faster than the FSB! With the Pentium 4 bus, it will take a new speed grade of rambus (533mhz, 1066 mhz effective - and yes, it is a planned speed grade), and two channels to soak up the nice 4.2gbytes/sec of bandwidth on the P4's FSB. Don't discount the Pentium4

and no, i'm not an intel zealot. I do, infact, have a Cyrix Processor.

I recommend that people http://www.realworldtech.com/page.cfm?ArticleID=RWT091000000000. And http://www.aceshardware.com/cgi-bin/ace/tech.pl?read=7281

P4 might suck. But get some technical reasoning before making the statement. Or, it might rock. Again, read logical arguments about either opinion. EVEN BETTER! wait for reviews before digging a products grave!

 

[Sunner]

BurnKooshie, true, but OTOH the Alpha isnt targeted at the same market as the P4, the Alpha can afford to cost a $hitload of cash, which it does not cause its an expensive CPU(well it is, but not THAT expensive), but rather cause stuff like 8 way RAMBUS interleaving, high end SCSI dis systems etc etc is expensive.

And of course the fact that its a low volume high end product makes it expensive by itself.

 

[BurntKooshie]

sunner - my point was that the bus of the PIII is over saturated - while the P4's will not be.