Lost all Faith

[Guest]

Since when were special compialations ever used for new processor designs in the past? I understand this is a very new and different design but so was Athlon and no special compialations were used for it?

What is the deal? I thought you were objective? You are no longer showing much objectivity these days.

You should really be slamming the coffin on P4. The fact of the matter is Intel used to be king of the FPU until Athlon came along and stomped that title into the ground. Now instead of a good fight back they just change the rules. Instead of calling them on this new BS tactic you do all that you can to cater to them.

Bah

 

[Grizely1]

HA you think Tom is going for P4? Check out the www.planethardware.com review of the P4!!!! I think they got paid by Intel or something! "As these benchmarks show, the Athlon is way up ahead of the P4. We <i>again</i> believe the benchmark software is faulty and does not reflect the true performance of the P4...."

i mean give me a break!

 

[mpjesse]

I think that everyone forgets that the P4 is BRAND new and has not matured. Look at the VERY first benchmarks of the Geforce 2 compared to the GeForce. There was an average of a 5 to 10 percent performance increase between the 2. Not much? Now look at it. The GeForce 2 MX spanks the old GeForce. It's all about maturity. Most of you forget that most benchmarks use SSE which give the P3 a good rating. The benchmarks out today don't use SSE2 and give the P4 a disadvantage. AMD isn't innocent of that fact either. Take their 3D Now! instructions away and I bet it would do horribly on any benchmark.

True, Intel is relying heavily on software to be SSE2 optimized- but rest assured soon every game will be SSE2 optimized.

Hey, at least Intel is trying to move on into a new type of processor while AMD sits on it's ass hoping sledgehammer will do as well as the Athlon. I don't see any major changes in AMD's roadmap in the near future. All I see is faster Athlons and DDR. I think it's time for AMD to step up and throw something NEW out.

-MP Jesse

 

[Grizely1]

Jesse! They are! In 2 months their new Palomino will be out! Where have you been?

 

[mpjesse]

Palomino? Oooo... 266mhz FSB. Like I said, faster Athlons. NOT anything new. I wish AMD would release the Sledgehammer already!!!!!!!! That would kill intel.

-MP Jesse

 

[Grizely1]

MAYBE IF THE SLEDGEHAMMER WAS FINISHED!!!!!!!! THEY DONT WANT TO SCREW UP LIKE INTEL DID WITH THE 1.13 AND RELEASE IT TOO EARLY! JEEZ YOUR PARANOID

 

[Grizely1]

And so what if it's a modified Athlon? It still kicks the crap out of the P4. And it is new. You obviously are undereducated so please don't post stuff you dont know about. Thnx

 

[Tom]

The idea behind the posting of results with the re-compiled software was to show the potential of P4. I was not trying to condone the fact that Intel designed P4 so that it sucks with todays software and depends on re-compilations. However, as time goes by you will see that the P4 will look a lot better, regardless if this is fair or not. I was simply trying to show what is possible.

I also pointed out that AMD has not supplied me with a special version of FlasK yet, which is why those scores are by no means useful to compare P4 and Athlon.

I wish you guys would read my articles a bit more thoroughly before you jump to conclusions.

Thanks,

Tom

 

[Guest]

3D Now! is FAR from being widely supported and the athlon relies more on it's very good FPU than anything else...

Seoman. Newbie at last!

 

[mpjesse]

3D Now! is supported in just about every benchmark, game, and multimedia program I can think of. Including OpenGL and DirectX Hell, look at the back of any game box and it'll say "3D Now!" or "Athlon" optimized/supported. But more importantly, since version 5.0 directX takes advantage of 3D Now!. If you don't believe me then go into the properties of your video card and find directX properties and OpenGL properties and you actually have the option of actually turning off "Enhanced CPU instruction sets." That's straight from OpenGL's properties. Same thing with 3D Mark 2000 and 3D Bench. WinAMP also takes advantage of 3D Now! I would not call that "far from being widely supported". Who doesn't have WinAMP, DirectX, or OpenGL? I know I have them all and so do most people.

-MP Jesse

 

[Kodiak]

You go Tom !
Rest assured, some of us read your articles thoroughly, even if we don't understand all the technical details, we get the gist
(whew, all that e-mailed talk about codecs went a wee bit over my head... but at least I tried

 

[Guest]

To whoever said that the Sledgehammer was nothing new, you obviously haven't researched it much. It is BASED on Athlon designs, but it IS a new processor. I see a very promising future in AMDs future. Roadmaps are pretty stupid anyway. Technology changes so quickly that any sort of long term plan designs can and usually do result in bad failures. Look what happened to Intels 5 years worth of planning for RAMBUS. DDR SDRAM kills it in most benchmarks, but 5 years ago it was untouchable. Both Intel and AMD have learned that it is a good thing to adjust plans on a daily basis now. The sledgehammer is not necessarily incomplete, either. There are working prototypes and have been for some time now. However, AMD has decided to wait and tinker with it and find any possible problems and fix the bugs and wait until the market is ready for it first. Intel should take that advice. The P4 benchmarks are simply pathetic so far. First impressions are important to everyone except us few who really read into this stuff. Intel should've known better than to realease something that is obviously not ready yet. What I want to know is didn't Intel Benchmark the stupid thing first? How could they not have foreseen these results?

 

[Rop]

3DNOW! is far from being a standard in most games/apps. In fact some of the benchmark you site that do support 3Dnow also support SSE as default and not allowing the user to turn them off like 3Dmark2000.
The Athlon does need 3Dnow to spank current procesors out their in a raw FPU shootout.
I do think that in the future more things will be optimized for SSE2 and by that time the P4 will be a viable option, but in our present day and time it is not at least for me...
And by the time most software gets recompiled and SSE2 optimized AMD might have SSE2 running in the hammer...





Why do I use LINUX ? Cause its the BEST OS
Why do I use Windows? Cause its the BEST Nintendo..

 

[RavenPrime]

Getting a little heated?
As Tom mentioned in his article, Intel isn't catering to gamers with its P4. Its catering to the average Joe who walks into his local Best Buy looking for the coolest looking case with the highest GHz rating.
"Benchmark? Is that what you call what happens to the top of your desk when you use a mouse without a mousepad?"

Intel is trying to retake its market share using raw GHz power. When SSE2 becomes commonplace, and it will, then it will start to CATCH UP to AMD MHz per MHz. That is until AMD adds SSE2 to its processors and it has already said it will.

My biggest gripe with Intel is lack of upgradability. The first P4s are a dead end--no upgrade path without replacing your motherboard! Of course,AMD is not innocent in this either--remember SlotA, that was out for what, a week!

James

 

[Tom]

While it may be true that quite a few software titles are carrying the 3D Now!-logo, most of them don't make a lot or good usage of it. Quake 3 Arena e.g. is to my knowledge not 'enhanced' with any meaningful 3DNow!-acceleration, especially not the 'enhanced 3D Now!' of Athlon. The reason is actually rather simple. AMD has still not provided decent tools to implement 3D Now!, while Intel is providing excellent compilers and tuning software (e.g. VTune). Obviously Intel was able to implement SSE2 into FlasK within one night. That's mainly due to an excellent and up-to-date compiler. AMD can only break through this Intel-optimization wall once it supplies compilers and other development tools that are equivalent to Intel's. The alternative for AMD is to use the same extensions as Intel, as they do with SSE2 in the Hammer-line.

On the software front AMD has got a very hard standing. There's millions of 3D Now! processors out there but hardly any software that makes good use of it. It seems as AMD has given up on this issue once they decided to implement SSE2 into Hammer.

The good thing for AMD and Hammer is the fact that by the time the Hammer-products get released, there will be lots of SSE2-enabled software available.

Regards,

Tom

 

[Rop]

Exactly the point I was trying to make.
I feel that we should all wait and see what the future brings.
Though the P4 might be a great product in the future who knows if by that time it is hammer time.

if AMD or Intel show me a great procesor for my current games and appz i will be the FIRST to STAND IN LINE to buy it. Since i just upgraded a month ago from my good old Celery 300a (which hit 464 flawless for over a year) to my New AMD Tbird 800 and Abit KT7MB i have all the time in the world to wait and see what comes. Plus you cant beat the Tbird when it comes to everything Linux

One last thing Tom more Linux based benchmarks would be great!

thanks




Why do I use LINUX ? Cause its the BEST OS
Why do I use Windows? Cause its the BEST Nintendo..

 

[mpjesse]

Yes you CAN turn off 3D Now! in 3D Mark 2000!!! Run a custom benchmark and change "CPU Optimization" to software T&L... that simple man! Anyways, I never said 3D Now! is a standard, all I said was it is utilitized in most games and apps... 3D Now! is old- i know this. But, if it was turned off, Athlons scores WILL hurt. Try it. I've done it using 3D Mark 2000. Same with SSE. That's my whole point, SSE, 3D Now! and SSE2 are important and do affect performance. So give the P4 some time.

-MP Jesse

 

[Rop]

I would invite you to reread my post as i did not say that you cannot turn off 3DNOW!, I said you could not turn off SSE.
A Standard is something that is widely adopted and by saying "it is utilitized in most games and apps" you are refering to it as a "standard".
As i said before Athlon does not need any help from 3DNOW! and it spanks any current x86 raw FPU out there.
The problem with 3Dmark2000 is that yes you can turn off Athlon optimizations but you cannot turn off SSE and thus 3Dmark is intel biased.

Try Sandra 2001 as an example. Turn off SSE in the CPU benchmark and see for yourself.

Your point being that SSE, SSE2, 3DNOW in principal are important, i agree. Though they are certainly not important if it does not become a "standard".





Why do I use LINUX ? Cause its the BEST OS
Why do I use Windows? Cause its the BEST Nintendo..

 

[Guest]

the athlon did the same with the slot A platform, it is not good but I guess they just want to get it out there, I would definatly wait for the platform to stablilise before buying any new chip (although I do have a slot Athlon as part of my collection)...

M

one of the first UK T-Bird users....

 

[Guest]

ok...... WRONG!
DirectX 6
and Also not ENCHANCED 3Dnow!

 

[Guest]

Hey Tom
Great Article, Great work. On AMDZONE they have information on a 3DNOW! optimised FlaskMPEG. CHeck it out. By the looks of it, it evens things up ALOT
Great Site!

 

[agriffin]

I understand the inclusion of instruction sets and what they mean to new chips. I also agree that as these mature, obviously new sets will be released. My main question is, if these instructions were developed and copyrighted by Intel. How is it that AMD is able to include them on their line of processors?

That's kind of like Nintendo making a game system that has the codecs to play Playstation games and Nintendo games.

I've looked, though not too hard, but haven't really found any detail on the bylaws. I was hoping someone here might be able to enlighten me.

Aaron

 

[Guest]

Intel was never the king of the FPU. That title has belonged to (at various times) IBM, HP and DEC for at least the last 5 years. I don't believe any X86 chip has ever had the best FP performance...

 

[Guest]

Yea, I agree that some people are shooting off their mouths before reading the articles. My only problem was the inclusion of Flask, period, and the weight given to it. If you do MPEG for a living, why burden your PC with a task vastly better suited to a DSP?

I'd venture to guess that a TMS 320C80 DSP would beat any general purpose CPU (even RISC chips with real FPU's) by a large margin, if not orders of magnitude. The figure of merit quoted is usually the time to execute a 1K FFT, and the figures are usually in microseconds. I bet that a 1K FFT probably takes milliseconds on a general-purpose CPU.

 

[Guest]

oh my god! all we're talking about the x87 FPU, not the IBM, HP and DEC stuff.
why don't you compare these fpu with the one inside NASA?!

<font color=orange>What do you think? </font color=orange>