Sledgehammer vs. Itanic

[matrox]

http://www.theregister.co.uk/content/3/16307.html
Benchmarks - Itanic 32bit emulation is 'unusable'

Unbelievable
itanium 32bit emulation really sucks

Sledgehammer can´t sink the Itanic (because intel is a monster compared to tiny AMD)
but Sledgehammer is at least a big iceberg for the Itanic, because SH is a 32-Bit monster

 

[slvr_phoenix]

I won't argue for a second that Itanium is going to be better than Sledgehammer. Obviously for pure processing power, we know which is better.

I would however like to point out something: Itanium isn't about JUST being 64 bits. It's about exploring a much different architecture based on x86. In a way it falls into the category of ingenuity like Transmedia and Radeon. It's a unique product that may or may not really have a purpose, but at least it's something different with new and unheard of ideas.

You don't see AMD working on anything intellectually exotic, do you? They're just working on upgrading things the exact same way they've been upgraded for the past few decades.

Despite all else, you have to at least give Intel credit for trying new things. And for still standing tall even when those new things don't always work as desired.

I'll look to AMD for the best of today's chips.

I'll look to Intel for providing the chip solutions of the future when today's chip advancement ideas simply run out.

- Sanity is purely based on point-of-view.

 

[AmdMELTDOWN]

yeah, try and put the newest formula 1 racecar on a dirt road and bring all your test equipment then see how well it performs.

Unfortunately, the internet is no substitute for human logic.

- Amd Helpdesk -

 

[slvr_phoenix]

I think you've completely lost me.

1) What does a F1 car running off-road have to do with anything?

2) What does the internet have to do with the price of tea in China?

If it weren't for the bad FPU, the P4 would be a great chip.
And I bet we'll see AMD taking a lot of the design concepts from it and working them into their own chip. But more importantly, I'll bet we'll see Intel taking a lot of the fundamental architecture and improving the rest.

And still, what does ANY of that have to do with the Itanium? Not a thing. But then neither does car racing or the internet.

- Sanity is purely based on point-of-view.

 

[Guest]

In all honsety, I appreciate that Intel is being innovative, it drives technology. However, no idiot would try and put an F1 car on a dirt road. It's not made for that. If you made an F1 car that could off road, you would have wasted a lot of time. I don't dislike Intel, I think they are great, I feel the same about AMD. However, you can't give AMD -ish because they keep making the F1 faster and handle better. That is what an F1 car is supposed to do.I really appreciate innovation, but lets keep comparisons somewhat realisitc. And maybe just maybe, give AMD a little credit.
my $.02,
Loki

 

[Grizely1]

You don't see AMD working on anything intellectually exotic, do you?

No, I don't see AMD making any chips that don't have purposes right now. The SledgeHammer will be the first "new" processor since the K7, but hey. The Pentium series (up until P4) all used the same architecture so who cares.

 

[TheAntipop]

good points across the board. i havent heard a whole lot about the itanium in a few months so im not sure how intellectually exotic it really is. Care to fill us in?

If at first you don't succeed, skydiving isn't for you.
<font color=blue>Intel Inside</font color=blue> = Idiot Outside

 

[mpjesse]

Actually- you should be looking at IBM. Remember that IBM were the ones who patented copper interconnects and a whole of other things that if not invented- Intel could have never invented. Intel is no different that AMD really. They look to other companies to actually provide the means to invent their stuff.

No one can sink IBM's R&D department and no one has ever tried. They apply for more patents than ANYONE in the world.

-MP Jesse

 

[Guest]

I have to agree, Itanium will be interesting to look at. It's interesting that it will be using a VLIW (very long instruction word) ISA along with EPIC (Explicitly Parallel Instruction Computing). This effectively allows the computer to do away with things like complex branch prediction, because you can execute both sides of an if-then-else at the same time while waiting for the answer to be calculated and at the end just use the side that was right. It's quite exciting from a computer architecture point of view. I honestly think EPIC will become quite successful if it gets implemented properly. Itanium may not be the one to do it, but it looks cool anyway. There already is one VLIW architecture out there already in Transmeta's Crusoe processor, which, for all its failings, is a <i>very</i> clever piece of kit. Go to <A HREF="http://friday.editthispage.com" target="_new">http://friday.editthispage.com</A> if you are interested about Crusoe.

Also, I consider AMD the only way to go with PCs at the moment as well. The bang-for-your-buck factor puts AMD way in the lead.

I wonder how long it will take for Intel to make a real challenge to AMD in this area.

Fat Chucky

 

[slvr_phoenix]

Thank you for nailing it right on the nose for me.

Itanium and the P4 both use some very interesting concepts.

Maybe their first versions won't work too amazingly. But eventually they'll be good products.

Meanwhile what's the only new concept AMD is giving us? A whole 64bit architecture...which isn't even new. While very useful and nice to have, it's certainly nothing to praise as innovative.

The CPU isn't just about speed anymore. It's not just about MHz and copper. The future of the CPU is going to be about handing data better, not just faster. It'll be about having the processor actually process things smarter. And simply, AMD isn't catching on to this yet. They just keep pushing for faster.

To use the very horrible abused (and now often even misrepresented) analogy: AMD is making faster cars, and that's it. Intel is working on making an airplane.

Right now, the car is running faster and looking like the best product. But in the future it's undeniable that the airplane is going to have more potential. And if AMD doesn't start working on products with that potential and leaves it all up to Intel to develop, then who knows how long AMD will really last.

Frankly, I like AMD's products better. But I like Intel's ideas for the future of processors a lot more than I like AMD's. I just wish AMD were the one working on these innovations. But then, that's what has kept Intel on top.

This competition has been great. We've seen chips jump in speed drastically. But Intel is finally recovering from the blow AMD dealt in a way that AMD doesn't seem to be expecting. Intel is working on changing the entire playing field by making a smarter chip, not a faster chip.

For now we can all laugh at Intel's slower new chips and call AMD the king of speed. But for how long will speed be this important? Or what happens if Intel's chips start running faster?

One thing is for certain: It's certainly an exciting time to be observing what new technology is coming out. Computers have never changed so much in such a short time.

- Sanity is purely based on point-of-view.

 

[Guest]

Yes indeedy!

I think you got it right there. Mind you, in AMD's defence they are just pandering to everyone's wants (I almost said needs there, but how many people really <i>need</i> a 1.2GHz processor?!).

The problem I currently have with AMD at the moment is the insistence to perpetuate the 80x86 ISA, which is probably the most brain-dead, ugly, and basicly inefficient piece of nonsense around (we don't have VAXs around to take that place now!). I learnt computer architecture on a Motorola 68008, which, while still very CISC, was a very elegent ISA. I then graduated onto the MIPSco RISC architecture which is simplicity and elegence incarnate (along with kick ass performance!). I've tried to learn about the more complex parts of the x86 ISA, but everytime I do I see something and say "Why the f#*@ did they do it that way" and go back to something that is well designed from the ground up.

It's interesting to note how intel and AMD have got round the kludge that is x86. They've made pretty serious RISC processors and surrounded them with silicon to translate the x86 instructions into something sensible, do what's needed, and translate it back. Intel doing it almost entirely in microcode and AMD doing it in solid state electronics for the simple stuff and the rest in microcode (one reason why Athlon performs so well for old x87 FP stuff).

Hammer will be pretty cool I've no doubt, but Intel's EPIC that is Itanium (sorry about the pun, I couldn't help it ) looks a way to go if it gets working properly. The computing world will be a better place when x86 is a bad memory.

I think I've been a little controversial here, please keep it civil folks. Fat Chucky dons his asbestos suit and looks for cover...

Fat Chucky

 

[slvr_phoenix]

I don't see how anyone can really flame you for stating the truth.

Just that the Mac fans seem to see that as meaning that a Mac runs faster ... which it doesn't. In a clock-for-clock basis in some instances it could. But since when has a Mac's clock speed ever competed with a PC? Heh heh. And the same could be said for a Mac's price tag.

But they do have some better structure. And it looks as though (FINALLY!) Intel is working on making the PC use just as good of a structure. (Wouldn't THAT really kick a Mac where it hurts. It's all they have left.)

And I'm not flaming AMD for not working on things like this. I'm just saying it's sad. And that if they don't fix that, their future will be limited.

To me it seems Intel has their focus on the future, while AMD has locked their focus on to the present.

- Sanity is purely based on point-of-view.