Intel x86-64. **Updated 2/17** It IS AMD64 compatible...out next quarter

[MadRat]

The Xeon derivative of Prescott has not yet been demonstrated, basically the Prescott with 4MB of L3 and some extra internal functionality they never disclose until release. There is no reason to assume a working Tejas sample is ready to display.

 

[nortexoid]

i suppose the saddest thing to extract from all of this is the fact that AMD is officially extending the life of x86...i hope intel doesn't.

itanium baby.

 

[MadRat]

Why?

 

[buleyb]

Originally posted by: nortexoid
i suppose the saddest thing to extract from all of this is the fact that AMD is officially extending the life of x86...i hope intel doesn't.

itanium baby.

agreed. Shoot the horse and find a newer one for the race ahead.

 

[mikecel79]

Originally posted by: MadRat
Why?

Because the x86 ISA is over 20 years old now and it's time to start a new architecture. X86 is inefficent enough. It's only the vast amount of x86 machines out there and backwards compatibility that have kept it alive this long.

 

[AIWGuru]

Originally posted by: mikecel79

Originally posted by: MadRat
Why?

Because the x86 ISA is over 20 years old now and it's time to start a new architecture. X86 is inefficent enough. It's only the vast amount of x86 machines out there and backwards compatibility that have kept it alive this long.

If we can keep making it faster and faster, it's a fine ISA. The ISA only needs to be replaces when we run out of room to work with it.
Also, madrat, did you read all of what I posted above regarding what the president of Intel said?

 

[mikecel79]

If we can keep making it faster and faster, it's a fine ISA. The ISA only needs to be replaces when we run out of room to work with it.

Ask anyone doing assembly program what they think of the x86 ISA. Most will have a few choice words about it....

 

[sandorski]

Originally posted by: mikecel79

If we can keep making it faster and faster, it's a fine ISA. The ISA only needs to be replaces when we run out of room to work with it.

Ask anyone doing assembly program what they think of the x86 ISA. Most will have a few choice words about it....

Screw those guys! hehe, j/k

...but. Keeping x86 alive has it's advantages. It's a sure thing and gauranteed Market for all those Assembly coders. Throw out x86 without a cross compatibility transition to a new standard and you basically start from scratch.

 

[grant2]

Originally posted by: mikecel79
Because the x86 ISA is over 20 years old now

And this is directly relevant how..? Because you like shiny new baubles?

and it's time to start a new architecture.

As Madrat asked, "Why?" Simply saying so doesn't make it true.

X86 is inefficent enough.

Actually, x86 is efficient enough... otherwise it wouldn't be used.

It's only the vast amount of x86 machines out there and backwards compatibility that have kept it alive this long.

.. which is obviously important to millions of purchasers who have passed over competing ISAs to buy x86.

What intel & AMD have recognized (unlike many techno-nerds) ... most users don't care about how "elegant" the ISA is, in fact, most programmers don't care either, thanks to compilers. They just want a cheap processor that lets them run lots of cheap software reasonably fast.

 

[MadRat]

Originally posted by: AIWGuru

Originally posted by: mikecel79

Originally posted by: MadRat
Why?

Because the x86 ISA is over 20 years old now and it's time to start a new architecture. X86 is inefficent enough. It's only the vast amount of x86 machines out there and backwards compatibility that have kept it alive this long.

If we can keep making it faster and faster, it's a fine ISA. The ISA only needs to be replaces when we run out of room to work with it.
Also, madrat, did you read all of what I posted above regarding what the president of Intel said?

I saw your verbatim of the article but couldn't find his original comments. I assume he didn't fingerpoint to Tejas or which exact ISA (AMD64/IA64) will be included with his upcoming demonstartion. They have previous created an emulator to translate x86-32 to IA64, so I don't find it out of the question that they release a demonstration of an emulator to translate IA64 to their 64-bit tecnology in the new Prescott-generation. Intel engineers and executives have said that Prescott has 64-bit functionality disabled. Perhaps they only meant to turn it on in their Xeon line? That would mean the Prescott-generation of Xeon has the distinctions of more cache (4MB of full-speed, on-chip L3 cache) and 64-bit functionality.

 

[NFS4]

Here's an interesting story on the matter:


http://news.com.com/2100-1006-5159067.html

 

[Duvie]

Seems about right then.....I would like to see A64 and opteron use the HT technology as well as that should work both ways....

You definitely can see where the development of INtel is occurring...Oregon!!!! I live in Clackamas County.....

IN the past they have used the names and words.

Rogue
La Grande
Willamette
Deschutes
I believe Nehalem
now clackamas

and a couple more.....So far most all of them have been named after rivers here in the state....

 

[Duvie]

Seems about right then.....I would like to see A64 and opteron use the HT technology as well as that should work both ways....

You definitely can see where the development of INtel is occurring...Oregon!!!! I live in Clackamas County.....

IN the past they have used the names and words.

Rogue
La Grande
Willamette
Deschutes
I believe Nehalem
now clackamas

and a couple more.....So far most all of them have been named after rivers here in the state....

 

[digitalsm]

Originally posted by: MadRat
The Xeon derivative of Prescott has not yet been demonstrated, basically the Prescott with 4MB of L3 and some extra internal functionality they never disclose until release. There is no reason to assume a working Tejas sample is ready to display.

Nocona(the Prescott based Xeon) will debut in May/June.

 

[digitalsm]

Originally posted by: sandorski

Originally posted by: mikecel79

If we can keep making it faster and faster, it's a fine ISA. The ISA only needs to be replaces when we run out of room to work with it.

Ask anyone doing assembly program what they think of the x86 ISA. Most will have a few choice words about it....

Screw those guys! hehe, j/k

...but. Keeping x86 alive has it's advantages. It's a sure thing and gauranteed Market for all those Assembly coders. Throw out x86 without a cross compatibility transition to a new standard and you basically start from scratch.

Thats why you do what Intel is(or atleast was) planning, and slowly phase x86 out

 

[digitalsm]

Originally posted by: NFS4
Here's an interesting story on the matter:


http://news.com.com/2100-1006-5159067.html

And people said Intel was stupid.

Let AMD suffer the teething pain, then swoop in.

 

[Snoop]

Originally posted by: digitalsm

Originally posted by: NFS4
Here's an interesting story on the matter:


http://news.com.com/2100-1006-5159067.html

And people said Intel was stupid.

Let AMD suffer the teething pain, then swoop in.

Wow, that is devastating for AMD. It looks to me like intel has absolute control over the cpu market.

 

[RaynorWolfcastle]

Originally posted by: Duvie
Seems about right then.....I would like to see A64 and opteron use the HT technology as well as that should work both ways....

You definitely can see where the development of INtel is occurring...Oregon!!!! I live in Clackamas County.....

IN the past they have used the names and words.

Rogue
La Grande
Willamette
Deschutes
I believe Nehalem
now clackamas

and a couple more.....So far most all of them have been named after rivers here in the state....

Hyperthreading isn't patented since it is actually SMT (Simultaneous Multi-Threading). The issue with implementing SMT is that it makes debugging and validation much more lengthy (and difficult?). AMD could implement SMT without Intel's permission even without the cross-licensing agreement. BTW, my understanding is that the cross-licensing agreement applies to instructions; so AMD could technically make a CPU that uses SSE3 instructions under this agreement but that doesn't mean that Intel has to show them how they're implemented in hardware, obviously.

 

[JeremiahTheGreat]

Some of the instructions in SSE3 actually originated in 3DNOW! would you believe...

For all those crying for the Itanium ISA to crush AMD64, please think about this - there will be no competition! Neither AMD nor Cyrix (cough) have a license, so this is one route no one should be asking for unless you're rich..

 

[NFS4]

Originally posted by: Snoop

Originally posted by: digitalsm

Originally posted by: NFS4
Here's an interesting story on the matter:


http://news.com.com/2100-1006-5159067.html

And people said Intel was stupid.

Let AMD suffer the teething pain, then swoop in.

Wow, that is devastating for AMD. It looks to me like intel has absolute control over the cpu market.

We shall see...

 

[MadRat]

If Intel endorses AMD64 then its a major blow to Intel IA64 partners, partners who's power is nothing to underestimate. After all Intel is built on the backs of its largest partners, not your typical mom and pop stores, so a move away from IA64 is going to piss off some pretty big partners like HP.

 

[MadRat]

Originally posted by: JeremiahTheGreat
Some of the instructions in SSE3 actually originated in 3DNOW! would you believe...

For all those crying for the Itanium ISA to crush AMD64, please think about this - there will be no competition! Neither AMD nor Cyrix (cough) have a license, so this is one route no one should be asking for unless you're rich..

You don't need a license for IA64-compatibility. You need a license for the process to connect to Intel's bus, something that AMD already might have access to through Intel directly or indirectly through someone like IBM. AMD won't falter any faster than its larger partners allow it to falter. Intel would need to persuade IBM to go 100% behind IA64 to get AMD's lifeblood let once and for all. But to get IBM 100% behind IA64, it would take IBM basically killing off its low- and mid-range processor markets.

 

[buleyb]

Originally posted by: MadRat
If Intel endorses AMD64 then its a major blow to Intel IA64 partners, partners who's power is nothing to underestimate. After all Intel is built on the backs of its largest partners, not your typical mom and pop stores, so a move away from IA64 is going to piss off some pretty big partners like HP.

Well I doubt it will be a 'blow' to the IA64 backers, because I'm sure they already know Intel's plans. But this isn't Intel ditching IA64 regardless, but it could stay a high end niche chip.

 

[NFS4]

http://news.com.com/2100-1006_3-5159754.html?tag=nefd_top

 

[ViRGE]

Yeesh, could they be any more sketchy on the details?