blue_heart

Distinguished
Oct 29, 2001
298
0
18,780
<A HREF="http://www.vanshardware.com/articles/2002/06/020620_Intel_To_Dump_Yamhill/020620_Intel_To_Dump_Yamhill.htm" target="_new">Intel To Dump Yamhill?</A>

and

<A HREF="http://www.theinquirer.net/19060235.htm" target="_new">Intel won't produce AMD clone</A>

wish if there was UnDo in the life
 

AmdMELTDOWN

Distinguished
Dec 31, 2007
2,000
0
19,780
that's funny I don't recall Intel ever saying that they were making a AMd clone, talk about putting words in Intel's mouth.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
 

spud

Distinguished
Feb 17, 2001
3,406
0
20,780
Good old Vans and the Inquirer too boot...

-Jeremy

<font color=blue>Just some advice from your friendly neighborhood blue man </font color=blue> :smile:
 

juin

Distinguished
May 19, 2001
3,323
0
20,780
There only 1 hick

Why intel dont have made comment faster it like 6 month after.I think there were not able to include X86-64 on prescott at time before the chip start to be in testing.

Still that a theory that we will never know the true.

cheap, cheap. Think cheap, and you'll always be cheap.AMD version of semi conducteur industrie
 

eden

Champion
Then how can they respond in the event x86-64 is widely accepted by Microsoft, and the entire hardware industry that follows the 32-bit users?

--
:smile: Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G! :smile:
 
when i hear Intel say it themselves then i'll believe it. Funny how they say it, but intel hasn't said anything.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
 

Dark_Archonis

Distinguished
Apr 20, 2002
286
0
18,780
Hehe, that won't be for a while (if it ever does occur). I believe that Intel will instead improve the 32-bit speed on the Itanium platform.

------------------------------------------------
Montecito & Chivano; Intel's Big Guns.
 

LED

Distinguished
Apr 28, 2002
511
0
18,980
"I believe that Intel will instead improve the 32-bit speed on the Itanium platform."
What?

I sold my sig for $50.
 

eden

Champion
Yes but nevertheless how can Intel possibly let go of the next WinXP if it turns out x86-64? We've so far heard talks in the background featuring AMD and MS, I just cannot imagine Intel letting AMD get their own OS just like that. I doubt improving 32-bit Itanium is of any importance here, it would never reach desktop pricing anyway (McKinley or Madison).
Instead, focus on Yamhill, or just call it the x86-64 chip, no big codename really...

--
:smile: Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G! :smile:
 

LED

Distinguished
Apr 28, 2002
511
0
18,980
MS isnt going to make an OS just for AMDs processors.....AMD has a lil over 10% of the pie...That would just be silly.

I sold my sig for $50.
 

eden

Champion
Then explain the conversations on MS' website during a court session week, where Sanders was with Gates, discussing it. It's almost sure there will be one, otherwise how will we move to 64-bit in the future, preserving the current program codes intact, not needing an emulation like the IA-64 does?

--
:smile: Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G! :smile: <P ID="edit"><FONT SIZE=-1><EM>Edited by Eden on 06/24/02 09:35 PM.</EM></FONT></P>
 

FatBurger

Illustrious
Then explain the conversations on MS' website during a court session

Perhaps Sanders was talking out of his ass and has no real evidence? Oh wait, that never happens...

otherwise how will we move to 64-bit in the future

We could do it the same way we did the 16-bit to 32-bit move.

not needing an emulation like the IA-64 does?

And why is that such a bad thing, I ask?

<font color=blue>Hi mom!</font color=blue>
 

eden

Champion
I thought the emulation takes longer and would not be as efficient?

Also how was the move to 32-bit done?

--
:smile: Intel and AMD sitting under a tree, P-R-O-C-E-S-S-I-N-G! :smile:
 

FatBurger

Illustrious
Currently yes, emulation takes longer. But that doesn't mean that they couldn't improve it and make it fast enough to keep up when the time comes.

And I'm sure there are others that remember the 16-32 move more clearly and can explain it better than I.

<font color=blue>Hi mom!</font color=blue>
 

bront

Distinguished
Oct 16, 2001
2,122
0
19,780
Van's hardware is probably the most AMD slanted hardware review site I've ever seen, so I wouldn't take their word for much.

English is phun.
 

Kelledin

Distinguished
Mar 1, 2001
2,183
0
19,780
Also how was the move to 32-bit done?
General-purpose register size was increased to 32 bits, as well as stack-pointer register size, index register size, and instruction-pointer register size. This allowed for intrinsic operations on 32-bit integers and access to a 32-bit address space.

Segment descriptor format was expanded to allow access to up to 4GB of linear address space.

Two segment registers ("FS" and "GS") were added to the original four segment registers. Segment registers were left as 16-bit, as they have been since the 8086, but the values stored in segment registers came to have different meaning.

A new "protected mode" was introduced, where the default operand size was 32 bits. (80286 had protected mode, btw, but was limited to 16-bit operands and 16MB of address space)

About the same time, the FPU was updated to allow better parallelism with the integer unit, as well as intrinsic instructions for trigonometry and logarithms. By default, the system still booted up in ye olde 16-bit, non-protected mode for the sake of compatibility; it was up to the operating system to switch the CPU to 32-bit protected mode.

On the software side, we started seeing 32-bit DOS Protected Mode Interface (DPMI) apps for MS-DOS. Most were using the DOS/4GW stub, a rudimentary DPMI layer provided by Watcom with their C/C++ compiler. Windows 3.x began supporting some 32-bit features, and Windows NT 3.51 shipped with 32-bit protected mode support. Finally Windows 95 was shipped with full (if somewhat buggy) 32-bit protected mode support. The 32-bit Windows API ("Win32") got back-ported to Windows 3.1 as the Win32s add-on. The rest is history.

Basically AMD is doing the same thing with the Hammer.

So far, it's confirmed that Microsoft <i>is</i> developing a 64-bit Windows for Hammer. It's been on some Hammer demo systems, so it exists. The question is, how available will this be to consumers? Will it be like Windows XP IA-64, available only to specific customers with specific requirements? Remember, the 386 was released some time in '91 or '92, and it took three years for a fully 32-bit consumer O/S to ship (NT 3.51 was not readily available to consumers).

<i>I can love my fellow man...but I'm damned if I'll love yours.</i><P ID="edit"><FONT SIZE=-1><EM>Edited by kelledin on 06/25/02 04:53 PM.</EM></FONT></P>
 

ritesh_laud

Distinguished
Nov 16, 2001
456
1
18,780
Also how was the move to 32-bit done?
On the desktop? It was done when developers really started feeling the limitations of 16-bit memory addressing. That occurred around 1994ish. Although Windows 3.1 running in 386 Enhanced Mode could access 256 MB of real memory and also take advantage of the 386's Memory Mapping Unit for 256 MB more of virtual memory, it was still a 16-bit OS and hence could only allocate a maximum of 16 MB per application.

Microsoft timed the release of Windows 95 well, because developers by that time really needed the ability to address more than 16 MB.

The same logic should apply to the transition to 64-bit. Developers will start coding 64-bit apps when they feel the need to address more than a couple gigabytes (ain't happening anytime soon for games).

Ritesh
 

Dark_Archonis

Distinguished
Apr 20, 2002
286
0
18,780
Actually, you're right. Come to think of it, Intel's forgetting about 32-bit altogether, and is really pushing 64-bit aggressively.

Eden, it seems that Microsoft is developing not only an x86-64 version of Windows, but it's also working on IA-64 version of Windows. In fact, Intel demoed Mckinley running a 64-bit beta of Windows at IDF this year.

------------------------------------------------
Montecito & Chivano; Intel's Big Guns.
 

MisterMaya

Distinguished
Jun 17, 2002
58
0
18,630
My take on x86 emulation is, the current *app* apps you'd want to run already run crazy fast on x86, so if you bump up proc speed a lot and introduce emulation, then your old apps would run bout the same as they always did. Seemed to work with N64, nintendo, all that... emulated, but on a faster proc so stuff runs fine.

Granted you gotta have hundreds of times the performance to emu those systems, but that's 'cuz you have to emulate custom dsp chips, the whole completely different architecture of the machine, and all kinds of analog weirdness... x86 should be easier, right?

-- Monkeys? What does this .sig have to do with monkeys? --
 

FatBurger

Illustrious
Absolutely, native x86 is far better than emulation. I don't think anybody is going to argue that.
The point where I'm torn between x86-64 and IA64 is which is better in 64-bit. I'd hate to go with one and then 5 years down the road we all look back and wish we'd gone with the other.

<font color=blue>Hi mom!</font color=blue>
 

MisterMaya

Distinguished
Jun 17, 2002
58
0
18,630
I'd bet on the new technology being better for us... having programmed sizeable things in x86 assembly, I don't think it's the most efficient way to go, it's <i>extremely</i> nonparallel by nature without extensions, has very very very few registers! ARGH I just revived memories of trying to index values in four-dimensional arrays using only two memory-base registers and two iterator registers! Oh, the pushing-popping-pain! (sorry... painful experience... must get therapy)

but anyway, x86 to me looked bad at working with lots of types of data that were very uncommon when it was made, but are used every day now. The extensions help with that, but there are also some other crazy things... like the way multiplies work is very inflexible, the input regs and output reg are hardwired at least for some versions of the multiply op. Plus, the break between 16-bit operands and 32-bit operands has just got to go, period... the 32 bit ones are faster than the 16 bit ones now, at least according to my profs. The extensions have also taken over so much stuff that the x86 is practically there as backup anyway. Even the floating point unit is an extension, for crying out loud.

Time for a new instruction set, I think.

-- Monkeys? What does this .sig have to do with monkeys? --
 

Kelledin

Distinguished
Mar 1, 2001
2,183
0
19,780
Time for a new instruction set, I think.
Ja, pretty much. Trouble is, there's too much precompiled stuff tied to x86. There was already too much of that before the 80286 came out; that's why the 80186 failed. :tongue:

I for one would be happy to run an Alpha as my main workstation. I could even afford a pretty decent one off of eBay and put Linux on it. But I'd have no hope of running Q3A on it with decent speed, or running the nVidia Linux driver either. OpenDivX would be out of the question as well. :frown:

<i>I can love my fellow man...but I'm damned if I'll love yours.</i>
 

TRENDING THREADS