Apple will probably be satisfied this month, seen Motorola completing the G5, the next generation of PPC CPUs, which it will be 64-bit, but it will also run 32-bit code in almost full speed. Apple is preparing a 64-bit version of MacOSX and they insist that porting the OSX apps over to the new CPU, it will be as easy as recompiling the app for the new CPU. The CPU will clock from 1 GHz to 1.6 GHz for its first generation. For now, Apple has released a long awaited dual 800 G4 machine, while it upgrades the iBook and PowerBook series of laptops with more RAM, speed and features. In the meantime, Transmeta announced yesterday their new Crusoe CPU which it will clock 1 Ghz and it will be available sometime next year. Intel is getting ready to release the first mobile Pentium4, which it will start clocking at 1.5 Ghz, while AMD strugles to produce new products that can compete with Intel directly and finds refuge in marketing tricks, renaming their line of CPUs as Athlon XP or Athlon 1800+. Same tricks Cyrix was doing 3-4 years ago when they could not produce CPUs with faster clock speed than the competition.
Motorola Completes its new 64-bit CPU, G5
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
20 Comments
2001-10-17 10:39 am
Eugenia,
you’re talking utter BS!.. I suggest you get your head cleaned up from Intel propaganda and inform yourself about the facts!
a) The Visual C++ Compiler by M$, which they’re also using to compile Windows itself has barely reached the P6 (or even P5?) stage (p5 was the Pentium 1, P6 was Pentium Pro, P-II and P-III!), Native P7-optimized Compiling in VC++ is A LONG WAY OFF! Some very small parts like the Windows Media Encoder might be optimized for the P4s SSE2, but *that’s it*!
b) TANSSASRWTP4 – There ain’t no such thing as “a simple recompile” with the P4! Stuff has to be MANUALLY optimized in Intel’s Vtune to take advantage of the P4, a long and hard process which is due to the fact that Intel just plain FUCKED UP the design! There are wellknown developers out there optimizing for the P4 and they say stuff like “a programmer would rather have a brain surgery than to optimize for the P4”! >:-) Plus, to take advantage of SSE2, the only *real* improvement the P4 has, one has to *vectorize* the code *manually*, auto-vectorization programs that translate normal C++ into SSE2-Vectorcode are also a LONG way off!
c) Intel just pain and simple blew it with the P4!.. For proof, read here: http://www.emulators.com/pentium4.htm – It is ANYTHING but a “brilliant CPU”, i wonder where you got that crazy idea from if not from Intel Marketing brochures! Fact is that the Athlon trounces the P4 in almost ANY Test with alot less MHz, except some (manually optimized!) stuff well suited for the P4’s ISSE2 like Video-Encoding!
d) P4 doesn’t even “behave like a big P3” with non-optimized code, which can be easily verified by the hilarious fact, that the 1.5GHz was surpassed in ALOT of tests by the 1GHz P3 when it was released!
e) I personally hate the x86-ISA, but i respect AMD for making a kickass CPU on an ancient totally outdated zombie-ISA that was supposed to be dead a decade ago! If it *really* has to be x86 and you want performance – go get an AMD, it’s without a doubt the best value-for-money out there and basically *all* comparisons show the Athlon XP really really ahead of any P4 (not intel’s “benchmarks” ofcourse!)
f) Just because Intel reached 2GHz now does NOT mean that the competition “cannot keep up”, has it ever occured to you that PERFORMANCE is what matters, and not MHz?? And that some CPU-manufacturers put more focus on the actual PERFORMANCE than on empty marketing-figures that can be used to fool the general public and naive people like yourself? Please, do tell me: Why is Intel making the Itanic with only 733MHz? 8) Have they lost it now? Or *could* it be that MHz is really just ONE factor that determines a CPUs overall performance?
Look at other CPUs – there are Tests where a 867MHz G4 _can_ be faster than a 1.7GHz P4! How do you explain that? Or look at the Alpha – It kicks the ass of ANY CPU out there in Floating Point, even the Athlon, yet it just reached 1GHz!
(and no, Intel-optimized SpecFP does NOT cut it, sorry, look at stuff like Cinebench or POVray instead! <:-)
g) Where on earth do you have the idea from, that the Linux-Kernel is optimized for the P4?? Given the state of GCC and its compiles i consider this complete nonsense!..
2001-10-17 10:55 am
“while AMD strugles to produce new products that can compete with Intel directly and finds refuge in marketing tricks”
not just a load.. but a shitload of crap.. strugling?.. i don’t think so.. no wonder they invented that annoying rating system.. it’s for dumbasses who thinks more Mhz is faster.. and AMD has been fair about it on top of that.. their XP 1800+ is superiour to the 2ghz P4.. (i guess that explains the “plus”)
2001-10-17 11:08 am
2001-10-17 12:18 pm
Just that this new fancy processor has a 64-bit word size, doesn’t mean it’s actually <em>better</em>. It’s not faster with compiled C++ code, since you still need the same number of variables and operations to complete a task. Keep that in mind. In fact, the Single Instruction Multiple Data idea is a whole lot better. Also note that Intel’s own C++ compiler does much better in optimization than MS-C++.
2001-10-17 1:44 pm
Get your head on straight. Cyrix was barely as fast in a limited number of areas as competeing intel products. AMD beats intel on all fronts with its Performance rating versus real P4 clock speed.
The simple fact is that the P4 is optimized for clock speed and NOT performance. The P4 is the cpu intel’s marketing department has been asking for for years! At this rate I would guess a 2.1 Ghz AMD XP would outperform a 3Ghz P4, and we should see this sometime next year.
Stop knocking AMD. It doesn’t have the luxury of paying R&D to design an underperforming processor that can ramp clock speed up enough to be competetive. They are forced to design the best processors they can from a technical perspective and hope the marketing department can convince the retards of the world to see that no matter what the clock speed, this processor is faster than the intel competition.
Here is a good analogy. You can make a car that can handle its engine running at 9000 rpms and 250 horsepower. You would think it would be a fast car. But if you can’t put a transmission and gears in it that get the horsepower to the ground it will only look good on paper. The marketing guys will love it. But wil the users
2001-10-17 5:07 pm
I find it strange, DeeKay, that I read a review a while ago (on ArsTechnica if I’m not wrong) where they said that the latest intel c++ compiler could do auto vectorization (if not all that extremely well). Sure, it was beta then, but still. You say that “auto-vectorization programs that translate normal C++ into SSE2-Vectorcode are also a LONG way off!”. Well, I say you’re wrong.
That said, I don’t think Motorolas PowerPC processors will ever be able to compete head on with AMD or Intel offerings. The market is just too small for Motorola to invest much money in its desktop offerings. That can clearly be seen in the yield problems they’ve been having with the G4 during its lifetime. Nothing says this will change with the G5, sadly. I’ve heard some rumors of Apple bying Motorola’s PowerPC assets, but really. That’s as much B.S. as Apple not releasing 10.2 for 32-bit processors.
Still, with higher prices and lower raw power, I find myself using an Apple Computer. It’s all in the software.
2001-10-17 7:42 pm
> As for Pentium4, I personally believe that it is a brilliant CPU
What exactly is this belief based upon? Any undergrad ce/ee/cs major can tell you that the P4 is the example they use for “not” how to design a CPU. Talk to the engineers at Intel and they will tell the P4 is a product of the marketing dept. The P4s entire design is tuned for one thing, to raise the clock speed, not to make a better CPU, not to make a truely faster CPU, but to make the unknowledgeable public pay more money based upon a number they dont understand. There is numerous literature describing the P4, I suggest you research it rather than believing Intels marketig dept. and making bold public statements about the quality of the P4.
> AMD strugles to produce new products that can compete with Intel directly and finds refuge in marketing tricks
While AMD is showing questionable judgement here, its nothing compared to Intel. AMD makes as best an CPU they can given the ancient x86 ISA, then their marketing dept slaps a misleading number on it. Intel, on the other hand, Intel designed an entire CPU based upon misleading marketing. Which is worse?
2001-10-18 1:34 am
How true about BeOS. BeOS has some problems in it, as all OSs do, but it has so much going for it. A lot of things in BeOS could be fixed rather quickly, if only we had access to source code. Linux’s VM, for example, was rewritten recently in the space of a few weeks by one person. Thing of what a small but dedicated group of coders could do to BeOS! But I digress. As for the G5 thing, its seriously cool! For the longest time, the PPC chips were architecturally powerful, but hampered by clock speeds so low that the less efficient x86 chips still beat them. A 1.6 GHz G5 will absolutely OWNEZ any x86 chip out there. Hopefully, IBM open PPC spec takes off (yea right) and I get a shiny new G5 for my next computer.
2001-10-18 12:47 pm
GHz 1.2 1.4 1.6 2.0
G5 SpecInt2000 987 1151 1340 —
G5 SpecFP2000 1005 1173 1359 —
P4 SpecInt2000 — — — 656
P4 SpecFP2000 — — — 714
Hopefully those IBM POP motherboards and AmigaOne motherboards start to appear, that PPC POWER needs to be RELEASED !!!!
Even though I am not a PC user (except at work), I have friends that are quite satisfied with AMD’s CPUs compared to Intel’s. Most say that they get their moneys worth and get better performance on top of saving some bucks. And even user reviews on CNET have got the same notion saying the the Pentium 4 is just a bad product plain and simple… but I will let you PC folks duke it out… I am not affected by this battle!!
The article is not a battle. I write this with an AMD CPU from a laptop.
The fact is that AMD was up to the spot in the competition up to 3 months ago. Now Intel rides away with lots of CPU horsepower while AMD does not have the hardware in its factories for better yields and Mhz, therefore they are left a bit out in the Mhz race so they resort in such marketing practices. I like AMD, and I hope they will be able to pull this off with their Hammer CPU. But things are not looking too bright, I must say.
As for Pentium4, I personally believe that it is a brilliant CPU, but as PentiumPro was back in the day (and it was critisized as badly as P4 is today), you have to recompile your apps for P4 support in order to really unleash the CPU’s power. If you don’t, then P4 “behaves” as a big PIII, nothing more.
WindowsXP, new Linux kernel versions and that unreleased upgrade of BeOS(BONE/GL) are already compiled for P4 btw.
Eugenia,
Since you are on the support team for BeUnited… are you telling us BeOS folks that an upgrade may see reality for BeOS with BONE/GL? I have tried to keep up with what is going on in the BeOS camp, but with all that has happened in the last month or so has got me side tracked alittle bit!!
No, all I am saying is that the (still unreleased) version I received from Be, Inc. a year ago for the OpenGL benchmarks I conducted for BeNews.com, was Pentium4 enabled (I had mentioned that in my benchmark article as well).
The OpenGL package I received from Be was replacing the BeOS kernel, along with some other system files as well, and it was turning on the MTRRs for Intel CPUs, but I also know for a fact that there was some nifty Pentium4 code in these system files and in the drivers too.
You should help Intel to get that code for the 1.7 GHz P4. Now it´s not even a breakfust for 7450 (G4), both of them running PhotoShop. Why don´t you make a review of a P4 and a dual G4 both running Maya?
I don’t know what reviews you’ve read, but the new AMD XP “Palomino” core CPUs are beating 2Ghz P4’s. The only app that seems to show the P4 in a superior light is Quake3 (by a margin that is hardly noticable BTW).
When AMD goes over to the .13 process (Thoroughbred) it will be even more appealing. The price/performance ratio of an AMD system remains superior to a comparable Intel system.
No I don’t don’t sides in the AMD/Intel debate. I just call them as I see them.
The last line should read…
No, I don’t take sides in the AMD/Intel debate.
The fact is … AMD chips ARE as good or better than their rating says they are (a 1800+ (1.53 GHz) blows away a P4 in most tests) … BUT people buying HP, Compaq, Dell, etc … are not techies and don’t have this information … so AMD changes the name of the processor to “fool” (or more acurately “inform”) their potential customers into seeing them as direct competition to the higher clocked P4s … The tech savy people are not fooled, nor were they ever, because they already understood enough to purchase the great price / performance king … the Athlon … But just like G4 evangilists found out … it’s hard to educate the masses and remove their biases.
Also … back when they FIRST came out … the cyrix chips were almost as good as their ratings said … EXCEPT at floating point ops … which at the time didn’t matter to business users … so a Cyrix 233 WAS as good as a Pentium MMX 233 (except compatibility issues) … BUT by the time they we’re competing with PII’s floating point had begun to become important … and so in some importnat benchmarks they actually performed 30% worse than other chips at their own CLOCK SPEED … not to mention how terrible they looked compared to chips running at their RATED SPEED …hmm
Its no secret that Bone, OpenGL and the Media kit are available from the net – I wouldn’t be suprised if a large portion of the community (all 300 of us) already run them (guilty 😉 A little polish is still needed, and if BeUnited licenced the code tomorrow, you still have to wait 6 months to get 5.5 – maybe even longer since the BeUnited/OpenBeOS crew need to probe the internals. Since it will take at least 6 months, a few graphic artists might as well give BeOS a facelift and call it a R6 release.
BeOS is still one of the best engineered systems out there – the implementation needs constant revisions just like every OS out there, but the architecture is fantastic. Even with the latest MacOSX.1 update, Sherlock still takes over a minute to search my hard disk for files. Its amusing to see all the new apps coming to MacOSX these days – 2 years ago they were promised for and in development for BeOS (Cinema4D, Lightwave etc).
Its sad that the platform lost HWOpenGL demos of Q3Arena (there is a screenshot somewhere) and the unreleased port of Unreal Tournament (Chris, I know of a great server in Russia willing to host it 😉
Apple charge too much for Mac hardware.
64-bit RISC desktop cpu… [drool]
Now if only a hardware manufacturer would use it to create an open-architecture platform for the non-Apple world.
Same tricks Cyrix was doing 3-4 years ago when they could not produce CPUs with faster clock speed than the competition.
Yeah, only this time AMD has a head on the competition (intel). They can beat ’em clock for clock. Unfortunately the consumer doesn’t understand this…
http://monkeysvsrobots.com/“