Anyone using a P4?

[Dr Mordrid]

Not buying RDRAM on principle because of RAMBUS's behavior.

They attended standards committee meetings on the future course of RAM standards then applied for patents based on what they heard, after which they sued the other attendees for patent infringement when the new RAM's went into production.

Kinda nervy.

As a result a lot of principled folks are just not willing to reward them them for their tech theft.

Dr. Mordrid

[wkulecz]

I thought they lost the patent infrigement lawsuits over this?
I don't buy based on who's suing whom, but on price/performance.

IMHO the whole patent/IP system is corrupt and being used as a weapon against consumers and fair competition.

OTOH you vote with your $$$ a lot more often than in elections, so perhaps paying more attension to what some of these companies are doing and paying a bit more for good behaviour if necessary would be a wise investment in the long run.

--wally.

[Dr Mordrid]

Price/performance is also a very valid method for choosing. In this case which ever method you use the bad guys lose

My other peeve would be to have as many folks as possible avoid upgrading to XP until Microcrap improves *their* behavior. In this case their insistance on XP's MPA and its content ads in the strangest places.

Dr. Mordrid

[wkulecz]

I agree about Microsoft and XP!

I advise folks to get W2K while they still can, except for notebooks where Product Activation shouldn't be an issue and there are real improvements when having to use the notebook on multiple networks over the behaviour of W2K in this situation.
But lose the juvinile Fisher-Price user interface and set the desktop to "classic" mode.

--wally.

[Jerry Jones]

Jerrold Jones

I believe Intel has made some serious mistakes during the past 2 years, but I will admit they seem to be trying to put things back together and their future looks pretty good.

This RDRAM debacle has been an example of their stupidity.

I have a new 1.7Ghz Pentium 4 here at work with 512mb of ram.

I'm running SSE2 enabled version 6.5 of MediaStudio Pro.

When I finish building my home system around the AMD Athlon 1.4Ghz (Thunderbird core), I'll be running MediaStudio Pro 6.5 on that machine, too.

I am going to EYEBALL the DIRECT impact of SSE2-enabled software where TRANSITIONS AND FILTERS are concerned.

First hand experience is where the pedal hits the metal.

I also have access to a 733Mhz G4 Macintosh and I plan to lump that into my comparison, too.

When I finish the tests... in November probably... I'll build a table and show the results.

I suspect the P4 with SSE2 is probably going to do pretty well, but I'm not as convinced as Terry is that it will be up to twice as fast as the Athlon.

I've seen too many tables published by other testers - some of whom have used SSE2-enabled software. And the results really weren't that convincing.

But if I'm wrong, I'll be the first to admit it after my tests are done.

[wkulecz]

How about a quick P4 encode of a 10 minute DV clip to NTSC-DVD preset (5500 Kpbs, VBR) vs the same clip encoded with the same MSP6.5 settings on the Athelon. A factor of two or more would decide the issue for me right now.

My lowly PIII is doing this at about 10X the clip run time, Doc has what should be a similar setup and says he gets about 5-6X, I futzed around a lot with the MSP6.0 MPEG2 settings and its not clear what the upgrade"inherited" so my DVD MPEG2 encoding settings may be sub-optimal, but I am happy with the quality when played back full-screen with WinDVD.

My DVD-R burner should be here Monday, playback in a stand-alone DVD player connected to a 32" TV is the benchmark I really care about :-)

--wally.

[droopy]

Woooow, I had to go away for a couple of days and when I got back I see this. Anyway, I hope you don't mind, but I still would like to reply to your earlier message Dr Mordrid.

Apples and oranges.
Your apples and oranges are the only way that most people are able to compare the different processors and their performances. Luckily these apples and oranges are consistently used by the people who test these things to at least give us an idea on what to do / buy. Apart from that, I find it to be really apples and oranges, as you implied, when you start comparing performance of a program which is optimized for the P4 using an Athlon. That same Athlon would shine if it's program was optimized for it.

Now something else, if you take windows media encoder, which is optimized for the P4, you will see that the P4 2.0 indeed outperforms a normally Thunderbird 1.4 easy. When you take the same program with the Palomino 1.2 overclocked to 1,5 it will still be outperformed, but with a very low margin though. This was done on Tomshardware site or AhtlonOC, I'm not sure anymore, I have to look it up again. The Palomino clocked to 1.5 / 1.6 and 1.7 is what AMD wil release in the coming 3 months. So..............

Additionally, the P4 is on the market for about a year now I guess. It takes a bit long don't you think, for these P4 optimized programs to materialize!!!!!!! They will come though that's for sure. I heard that TMPGEnc will soon come with their encoder using it.

DDR vs RDRAM
As for this one, I'm not going to say anything on it as this was exactly one of the points that I tried to make in my previous message.

Now to try and put some fire fighting foam on your gas fire ;-)

Northwood core at 3.5.
Nice, very nice, but sure not something which is going to materialize for us consumers in the near future. Say 1 to 1.5 years??? Intel will for sure first want to milk us consumers with their current P4 2.0, 2.2 and other processors first. Apart from it, if this is the way to go, AMD will definitely follow. Wouldn't you think so? This is something though which you shouldn't consider now. That P4 3.5 processor will likely have a different socket again, so not usable in your currently available mobo's. I heard a rumor though that AMD is going to stick to their socket A for the next 1 to 2 chip generations!!!!!!!!!

AMD cutting jobs.
He Dr Mordrid you disappoint me. From the previous messages I read from you I didn't think you were the mud throwing type ;-)))). (Just joking) I just want to remind you here of Intel's past track record. You will find that they have not been behaving very nicely on this subject either. Apart from that, their current i845 DDR policy reminds me of something. What was it again. It was a company on whose policy we probably all agree on, or was it against. Ooooooh yeah. Intel has been watching Microsoft policy and saw what they can get away with. They probably thought, "What Microsoft can do we can to". As such they forbid manufacturers to introduce DDR support for this chip and file charges against the one which ignores them. Is this good for us consumers. You give the answer.............


As a last I just want to say that I neither prefer AMD nor Intel. I also do not have stock in either of them. I just look what's best for me at a certain period in time. At the moment this is still AMD, but who knows what it will be 4 months from now, especially if the Palomino promise doesn't materialize.

All the best, Leon

[Dr Mordrid]

The P4 optimized TMPGEnc was just released.

I guess with my gas-fired analogies is proper to note the my father in law is a retired fire marshall & brother in law is still an active firefighter

As for the length of time it takes major companies to launch CUP optimized code, I read recently that the average timetable is 2 now years. It's called "corporate inertia".

I dunno about the Palamino. A while back I read it'll have Advanced 3DNow! + all 144 SSE2 commands + double precision FPU but now all the published info says it'll only have the few SSE2 instructions already in Advanced 3DNow! (19 math, 5 DSP). Looks like full SSE2 waits for the Hammers.

Dr. Mordrid

[droopy]

@ALL

I'm a bit frustrated on this as I've been looking for that comparison test I mentioned in my first message where windows media encoder was compared on a P4 2.0 and a to 1.5 overclocked Ahtlon MP 1.2. Can't find it anymore. I hate it when that happens to me.

Anyway, it seems like the Palomino promise is coming true. Now if only the prices would follow that same path. Ah well, we have to wait and see. Decision time for the next PC is still about 2.5 to 3 months away. Who knows what Intel will come up with in the mean time.

Here's a URL to a page containing some Palomino and P4 comparisons. Definitely some interesting reading there.

http://www.theregister.co.uk/content/3/22126.html

Hope you like it.

Regards, Leon

[Dr Mordrid]

If you look deep into many of the published tests you'll find the AthlonXP (Palamino) ahead in general computing tasks (business apps etc.) where its slightly faster FPU give it an edge, but the P4 is ahead where streaming large chunks of data are involved (ie: video). Those big, fat pipes of the P4 have a lot to do with this.

Add software with SSE2 compatability & P4 optimizations (for those fat pipes) and things are not as clear as you presume.

You also have to take some of the benches with a grain of salt, especially where you see very few SSE2 optimizations in the software being used for the test. The rendering tests using POVRay are a perfect example and this is noted in the tests text.

Dr. Mordrid

[mgu2]

Dr. Morbid,

you seem to be a big advocate of Intel and P4, I suggest you read http://www.emulators.com/pentium4.htm

AMD's faster FPU helps mainly in 3D Studio benchmarks, 90% of all other apps do not use FPU at all. The real reason behind the benchmark results is Athlon's much better instruction/clock ratio.

Pentium 4 architecture is flawed:

- Ridiculously small (i486 era) L1 data cache: 8KB (Athlon has 128KB), this simply KILLS data processing performance

- Only one instruction can be decoded per clock (Pentium III decodes 3 + 6 micro), trace cache not even on PIII level

- Bit shift operations require 4-6 clock cycles (the ages old i486 requires 1 cycle)

It seems Intel relies on "consumer buys the MHz" instead of making quality products.

[wkulecz]

We can go around this forever.

The definitive answer for me (or it would have been if DVD-R had adaquate compatability with existing players) would be from the the times for encoding 1hr of DV into 5500VBR mpeg2 using MSP6.5 NTSC-DVD setting. My PIII-850 usually does this in a bit under 6 hours. If and upgrade can't cut this in at least half, I'll save my money, and for the time being stay with what I have until I can get this level of improvement.

Time to encode 1hr DV to 5500 VBR ==

Athelon 1.4GHz (or whatever) = ???
P4 1.6GHhz (or whatever) = ???

Replacing the ??? with actual measurements would likely be the decider for me on my next upgrade. Of course one needs the same DV clip on each machine for a truely fair test.

It gets more complicated when you add the possibilities of Rambus vs. DDR vs. SDRAM, but best vs. best would be Rambus P4 vs DDR Athelon.

--wally.

[DJ]

Ditto Wally! Thats the kind of benchmark I would like to see.

DJ

[droopy]

I tend to agree as well. The comparisons though for video are hard to find and far between. Every now and again you find one and you have to nurish it else it's gone. That's why I was so frustrated of not being able to find that overclocked Palomino MP 1.2 and P4 2.0 comparison anymore. Additionally with a few you also see MPEG4 encoding as comparison. Here too though, few and far between. So simple as it may sound, the comparisons that I get are used by me to compare and try to find the best CPU / value for money for the next buy.

Apart from above, this was not the reason why I posted the earlier URL. I was just abusing this post to provide some more information on both the Palomino, which for the moment is my favorite, and the P4. I just thought this was a good page with some more info available. If I find more interesting info, I'll post it again. If this info is in favor of the P4, It'll be posted.

To be honest, I'd encourage anybody if he or she finds a good video comparison for these CPU's to post it. I'm sure that apart from me, there's more out there whom whould be greatly interested.

I tend to agree with Dr Mordrid though on his comment of the P4 being better, for the architecture it uses, with streaming data. The Palomino is getting close quickly though with its raw power. I heard the rumor that within a few weeks there'll be the 1900+ and before the end of the year the 2000+. But also the rumor is going round that the P4 2.2 will be here soon.

And yes, it's a never ending story. Let's keep it that way, as this means that there is competition between AMD and Intel where each is trying to outperform the other in some way. Good for us consumers I'd guess. The advantage at the moment though is with Intel as they still rely very much on their past reputation and their Mhz numbers for the less informed consumers.

I read a good comment on that, I believe it was also on this forum, not sure there. IT managers are still looked upon strangely / fired when they suggest / buy AMD operated computers instead of the established Intel ones. It's a comment which for sure is over the top, but gives a good idea on the idea with consumers for AMD and Intel.

Read you next time around

Regards, Leon

[droopy]

@ALL

I said it, I'LL BE BACK. My god, I sound like a cheap Arnold Schwarzenegger imitation. To think I could fall so low.

Check this one. Especially the MPEG2 encoding part!!!!!!

http://www.aceshardware.com/read.jsp?id=45000224

Hope you like it, or not.

Regards, Leon

PS: This is just as information sharing, nothing more .