• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

P4

Nemesis3

Member
There seems to be a lot of conjecture when it comes to the Pentium 4 and the concurrence seems to be this:it will probably be out in late October and on a clock for clock basis it will not be as fast as a P3. Which leads me to the reason besides raw clock speed I personally was waiting on the p4(sold my 450pII and puttering around on a 233pII since March) as I was under the impression it would be attractively packaged with the following:
1. Native ATA100 Any performance boost?
2. 400Mhz FSB Surely we should see some benifit here?
3. USB 2.0 Will it be present?
4. i850= dual channel RDRAM memory setup... might offer bandwidth and speed but at what price?
If I am erroneous on these let me know🙂

What I would like to see but won't:1394b,serial ATA,Copper interconnects, and PCI-X

Tossing aside "it will give Intel headroom" for the reality that if you buy a p4 its not because you care what a later generation of the core at a higher speed and a new motherboard/chipset will do, you care what it will do it the form you bought it. This leaves us with with the cost of RDAM, and whether or not it may quickly become extinct en leui of DDR or higher speed RDAM versions. Thoughts on whether the p4 and it accompanying mobo/chipset will perform well enough to make it worth waiting for? I guess if nothing else it might lower p3 prices some more eh?
 
You need to remember that Intel is toying with their first totally new core in five years (yes, the P6 has been around since the Pentium Pro, which debuted in '95). Raw power isn't necessarily everything; what they are trying to do is make a core that is capable of doing more natively. I would imagine that some of the new features that will be present on i850 boards will make up for the P4's "less-efficient" 20-stage pipeline.

1. So far, ATA/100 offers no relative performance boost for IDE devices. It's largely marketing ploy. Considering that today's fastest IDE hard drives barely surpass 40MB/s (well under ATA/66's maximum), it will be quite a while before perfections in design can truly take advantage of this, if at all. I'm betting that Serial ATA will render ATA/100 useless technology.

2. Theoretically. 4 times the effective data throughput sounds wonderful and will help the P4 compete with the Athlon. But as always, bottlenecks in system buses (PCI, AGP, etc.) will prevent totally efficient use of this additional bandwidth.

3. Most likely.

4. RDRAM is getting more affordable, but it's still not at the level most consumers want it to be. It's no surprise that most Coppermine-based OEM systems run on VIA 133A chipsets. Using the i820 with RDRAM puts these things WAY out of reach. We'll have to wait and see if DDR SDRAM really creates the impact that some are expecting it to.

 
this chip is just pure crap. They designed it with a 20 stage pipeline just to get clock speeds up real high. Marketing will be happy with that.
 
Hmmm how is it totally new, when they have just put 20 stage pipeline. I don't see the new design in that.

AoD

P4...Seems like a sucker...AMD here I come
 
Maybe there is some benefits I am missing then?
That translates to:
1. Ata100 =useless
2. Slight improvement at best
3. Might not be there.
4. Overpriced, unknown performance value, and maybe quickly obsolete...which is something when speaking of Ram types, because unlike processor speeds they are a little slower to go obsolete🙂
 
Native ATA100 Any performance boost?

Should be, but not too much to begin with.

400Mhz FSB Surely we should see some benifit here?

Yes but again probably not as much as you'd think.

USB 2.0 Will it be present?

Yes.

i850= dual channel RDRAM memory setup... might offer bandwidth and speed but at what price?

The RDRAM is the biggest hoax ever made in computing. Despite its theoretical 3.6 GB/sec bandwidth, it never reaches that because it is much narrower than normal SDRAM. It costs an arm and a leg but it is either slightly faster than PC133 SDRAM or much slower, depending on how it is setup. Add the fact that you must have it in pairs and you can see why RIMMS are such a rip-off.

DDR RAM is definitely the way to go. All video cards use DDR RAM because they need high bandwidth. Any video card using RAMBUS instead of DDR would perform very poorly and would cost much more. Avoid RAMBUS at all costs.
 
Being that you guys shouldn't bash either amd or intel for their designs.

This CPU has been design and is on the marketing board.

1. Being in Intel's positions and using a old outdated code, Intel WOULD NOT bring this to the market unless they have a speed & processing advantage.

2. I believe 400FSB is only for the memory - I could be wrong ie. AMD 2x100 FSB only for CPU to memory.

3. Hopefully USB 2.0 & Firewire will be available.

4. By then, Rambus would be affordable and they do have that new rambus thats 4x or something.

You can't run SDRAM forever.........
 
"The RDRAM is the biggest hoax ever made in computing."
Perhaps; certainly your opinion.
"Despite its theoretical 3.6 GB/sec bandwidth, it never reaches that because it is much narrower than normal SDRAM."
False. That bandwidth calculation already takes into consideration the 16bit path.
"It costs an arm and a leg..."
False. $250 bucks for 128 megs of PC-800 is NOT an arm and a leg.
"... but it is either slightly faster than PC133 SDRAM or much slower, depending on how it is setup."
True. Trick is to set it up right!
"Add the fact that you must have it in pairs..."
...only on an OR 840, not i820
" and you can see why RIMMS are such a rip-off." -BFG10K

(just a little proof reading 😛 )

 
Had to know fkloster couldn't resist. Fkloster, will you upgrade to a P4 when it is out?
 
Hey Adul! I will upgrade to a P4 board & chip cause all my other stuff i.e. memory will slide right on over w/no probs.
 
flokster: You should wait for the second-generation P4 chipset, due out this spring...rumor has it that the i850 will only last 6 months. When the P4 moves to a .13u process in Q2 2000, Intel is changing the packaging from 423 pins to 453 pins (I think those are the pin numbers). The i850 will then become useless for further upgrades...
 
Sohcan is correct; the first generation P4 motherboards will be incompatible with second generation P4 motherboards, which are due out in October and May respectively. I'll put up a link as soon as I find one.
 
>>Native ATA100 Any performance boost?
>Should be, but not too much to begin with.
My point here is not to discredit the need to the technology ever being beneficial.. but to bring into focus that the p4 is not going to get a big performance boost by its presence with current hds.

>>400Mhz FSB Surely we should see some benifit here?
>Yes but again probably not as much as you'd think.
Once again no complaints for anything faster but focusing on real life present benefits of the p4 when it is released.

>>USB 2.0 Will it be present?
Yes.
Hmm anyone know for solid its performance specs vs. USb 1? I believe I read 450MB for USB2 vs the hat 12mb for USB1?

>>i850= dual channel RDRAM memory setup... might offer bandwidth and speed but at what price?
>DDR RAM is definitely the way to go.
No one is debating DDR vs RDAM as currently DDR is not an option.

>1. Being in Intel's positions and using a old outdated code, Intel WOULD NOT bring this to the >market unless they have a speed & processing advantage.
Definite logic here, but one could have said the same for the 1133 but after it I agree there will be even more focus not to #### this one up yet when it was being designed Intel was in a total diffferent place.

>2. I believe 400FSB is only for the memory - I could be wrong ie. AMD 2x100 FSB only for CPU to >memory.
Anyone know?

>3. Hopefully USB 2.0 & Firewire will be available.
We have heard one definite on USB 2.0 and not even a maybe on 1394b,serial ATA,Copper interconnects, and PCI-X etc..

>4. By then, Rambus would be affordable and they do have that new rambus thats 4x or something.
By then is October are you thinking prices will dramatically fall to where price and performance are closer matched?

>>"Despite its theoretical 3.6 GB/sec bandwidth, it never reaches that because it is much >>narrower than normal SDRAM."
>False. That bandwidth calculation already takes into consideration the 16bit path.
So it does reach 3.6 GB/sec?

>>"... but it is either slightly faster than PC133 SDRAM or much slower, depending on how it is >>setup."
>True. Trick is to set it up right!
Hmm even set up right it is only slightly faster.... all that great given the price?

>>"Add the fact that you must have it in pairs..."
>...only on an OR 840, not i820
I was under the impression the p4 would be paired with the i850 chipset?


>flokster: You should wait for the second-generation P4 chipset, due out this spring...rumor has >it that the i850 will only last 6 months. When the P4 moves to a .13u process in Q2 2000, Intel >is changing the packaging from 423 pins to 453 pins (I think those are the pin numbers). The >i850 will then become useless for further upgrades...
>Sohcan is correct; the first generation P4 motherboards will be incompatible with second >generation P4 motherboards, which are due out in October and May respectively. I'll put up a >link as soon as I find one.
Very good point indeed is that we can expect the "life" of the mobo to be 6-8 months. Obviously what quickly comes to mind is the expected life of this 2nd generation mobo? One might also focus on the RDAM issue of several flavors are going to be released soon from what I understand. If anyone knows this roadmap as well....?

So our concurrence is:
1. ATA100 useless for sometime to come
2. 400Mhz FSB less benficial than one might imagine.
3. USB 2.0 will be present
4. RDAM IF CONFIGURED CORRECTLY offers only a slight performance increase at a substanially higher cost and if incorrectly then it perorm worse.

We won't be seeing the following:1394b(firewire), serial ATA, Copper interconnects, and PCI-X

Sounds like I probably need to sit on my non-gaming 233 for at least another 6-8 months as sadly the fastest thing going is a p3 on an old BX board🙂 Ouch 14 months on a 233 notebook, at least I have got 6 of them out of the way...woo...woo
 
this chip is just pure crap. They designed it with a 20 stage pipeline just to get clock speeds up real high

- What a dumb comment. We haven't even seen the real benchmarks yet.

P4...Seems like a sucker...AMD here I come

- Another dumb comment. Seems? I'm waiting for Anand's take on this.

End result, more expensive, no more performance, nuff said.

- More FUD

You all sound like those morons who said K7 would suck. P4 may or may not be good. If it sucks, more power to you. If its awesome I'll laugh in all your faces! Needless to say, we shouldn't judge the product until it's actually released.

 
Perhaps; certainly your opinion.

Well I have yet to see evidence of RAMBUS actually doing something useful.

False. That bandwidth calculation already takes into consideration the 16bit path.

Really? So normal SDRAM is what 800 MB/sec - 1.2 GB/sec? So if RAMBUS does actually have a 3.6 GB/sec bandwidth, wouldn't that make it 3-4 times the speed of normal RAM? I don't think so. As you admit later on, it is only slightly faster than normal SDRAM.

I think they do the calculation based on raw MHz only. If it was the same width as SDRAM it would certainly reach 3.6 GB/sec bandwidth. The fact they don't factor in the narrow width is exactly what the hoax is. Their marketing will always tell you the speed but never the width.

False. $250 bucks for 128 megs of PC-800 is NOT an arm and a leg.

And how much does 128 MB of SDRAM cost? You can't say that RAMBUS is cheaper.

True. Trick is to set it up right!

So it has 3-4 times the bandwidth of normal RAM, but it is only slightly faster than SDRAM, and only if it is properly setup. Plus it does cost much more than normal RAM.

My point exactly.
 
I read Anand's article on the P4 and I am underwhelmed. But of course if the damn thing because cheap enough as a system and AMD don't respond well I will probably get it.
The sucker seems to need a lot of Watts, that's all I can say.
 
BFG10K
RDRAM has a 16 bit bus. SDRAM has a 64 bit bus. I believe RDRAM runs at 400mhz QDR? well 400 x 4
is 1600 with a 16 bit bus that 2 bytes so 3.2G/sec SDRAM at 133 would be 64/8 is 8 8 x 133 is 1064 is basically 1G/sec but RDRAM doesn't really offer any added benifit because of high latency and other things.
 
-KarsinTheHutt

Look at your post: "You all" "dumb comment X 2", "morons"
First avoid flaming with "dumb comment" and calling people "morons" as it has nothing to do with discussing future/current technology and has little purpose.
Also avoid super generalizations of "you all" as people tend to like to be individuals or if nothing else I do.
As far as solid real life benchmarks are concerned yes a lot of this conversation is sure to be conjecture. This discussion is very valid as to the following benefits of P4:

1. Will USB 2.0 be included or not.
2. ATA100 is already out there. What are the benefits from it?
3. The presence of the 1394b etc.

Since "you all" would tend to indicate you meant all posters on this thread then as for the comment "If it sucks, more power to you." I for one really want for some "non-sucking" technology to be released. I.E. I don't want it to "suck". My point in starting this thread was to inquire as to some more specific information on the p4 and not to "judge" it. The reason I was hoping to find more is to figure out whether to continue waiting on it and maybe waiting after it🙂

Over at Sharky Extreme I was told the following by Arcadian:
1. The Pentium 4 platform includes ICH-2, which is the same chip in the i815e chipset. Therefore the ATA/100 is the same as found in boards like the Asus CUSL2.

2. 400MHz front side bus is a definate improvement in CPU bandwidth! Of course, the real benefit comes to the server world, where Pentium 4 Xeon will have more headroom in this area. But desktops should see a small improvement, since 133MHz of the Pentium III was only a small bottleneck.

3. As I said, the Pentium 4 will use ICH-2. USB 2.0 is a feature of ICH-3, which won't be out until 2001. Sorry.

4. Dual channel Rambus has the ability to really make a comeback. The fact that bandwidth is higher is nearly as interesting as the fact that you can interleave between the Rambus channels, and half the aweful latency that Rambus has. Rambus has always had higher bandwidth than PC133, but PC133 performed better on PCs because of smaller latency. Price is also getting lower for Rambus... the Pentium 4 might encourage memory makers to increase supply, which would lower prices further.


So:
1. ATA100 useless for sometime to come
2. 400Mhz FSB less benficial than one might imagine.
3. USB 2.0 will not be present
4. RDAM IF CONFIGURED CORRECTLY offers only a slight performance increase at a substanially higher cost and if incorrectly then it performs worse.
 
Hmmm... I've always wanted to try my hand at trolling - I guess I'm not very good at it. 🙁
No one would even pay attention to me when I insulted them in the video forum 🙂 (They were too busy killing one another :Q ).

"I for one really want for some "non-sucking" technology to be released" - Doesn't everyone?

Comments like: "this chip is just pure crap" piss me off though. How can any of the naysayers back this up - do they have a p4 system to test?
 
Nemesis3
As far as the P4 goes I think the thunderbird/mustang plus DDR RAM would be better but I would wait for benchmarks. RDRAM right now is kind of pointless but once they get the bus of the CPU to RAM up to 400MHz it will improve performance a lot.



 
Did Intel help design USB 2.0? I was under the impresson that they just shruged it off and let someone else do all the work... Are they supporting it?
 
Back
Top