Is intel EMT 64 bit pure or amd 64 bit pure

[GlacierFreeze]

Yeah well over half of that stuff listed, the AMD64 has and the EM64T doesn't.

 

[evilr00t]

gj, contradicted yourself then.

All I see is EM64T lacks this and that, AMD Supports that.

 

[levicki]

The K8 has expansion of another 8 GPRs(general purpose registers) when it is running in 64bit mode, which is not the case with Netburst.

False. Both have additional 8 GPRs and additional 8 SIMD registers.

The X86-64 is performing better on K8.

That was before Conroe hit the streets.

AMD traded x86-64 with Intel for iSSE3.

I don't know where you are getting your info but what you are saying here is purely speculative crap.

Intel had EM64T way back in the first Prescott core only it wasn't enabled until the 5x0F series showed up. Subtract approx. two years from the release of Prescott to account for the length of the CPU design cycle and you will know when Intel already had EM64T in the works. It is not an aftermath hack as many incorrectly think.

The other is the EMT64 which was pretty much a straight up copy that I believe intel actually pays AMD for the rights to use in their more recent chips (cant remember exactly).

What's the matter doc, too much crack inhaled lately? Just keep on dreaming. 😀

 

[49ers540]

Yeah well over half of that stuff listed, the AMD64 has and the EM64T doesn't.

Shutup noob. Xeon has always ran strictly 64bit until AMD wanted to push 32/64 together. EMT64 is just another way of saying 64bit. Intel Cross license with AMD. They just push it and made people think they are better than Intel. Intel didn't push it because desktop isn't ready for 64. Even today, programs are still written in 32. 64 or 32, it will all run the same speed. Get a life noob and stop posting crap!

 

[cxl]

The main reason they call it 64 bit is because of the ability to do 64 bit floating point operations without any 'work' arounds.

Nope, even 486 CPU can do 64-bit FP ops without "workarounds".

In fact, specifying "processing width" of CPU is prety fuzzy bussines. (E.g. since SSE2, x86 is capable of processing even 128-bit integer, but still is considered 32-bit CPU).

Anyway, in this case, going from 32-bit to 64-bits meant that the most common integer registers and integer operations were widened to work with 64-bit operands.

The 40/48 bit limitation is likely a hardware design shortcut to save silicon. The instruction set is likely capable of 64 bits.

Yes. In fact, AFAIK, there is 40 bits of physical address, but AMD64 is capable of addressing of 48 bits of virtual memory. In reality this means you can have 2^40 bytes of physical memory and 256x more in swap file.

That means if you use 64-bit address in your code, upper 16 bits are simply ignored.

 

[cxl]

But there is little or no 64-bit driver support let alone software that runs 64-bit. So it's a pointless question.

Its been what? 4 years since AMD tried selling AMD64 as the reason to buy AMD. Guess we are still transitioning....

That is no surprise. Actually, things are going much faster that the last time when 16->32 bits was happening - first 32-bit CPU, 80386, was introduced in 1986, but first widely used 32-bit OS (well, mostly 32-bit , Win95, was introduced in 1995. By the time, 80386 was already phased out... With AMD64, there is actually chance you will use origininal version of the chip with Vista

 

[levicki]

With AMD64, there is actually chance you will use origininal version of the chip with Vista

True if:

- You haven't had to change your mainboard to support more than 3GB of RAM -- many S754 mainboards had this stupid limitation despite being build around 64-bit CPU and rarely you can see desktop board supporting more than 4GB anyway.

- You haven't had to change CPU three times from S754 to S939 for dual-core and from S939 to AM2 to get DDR2 memory. Not counting having to get Venice cores for SSE3.

So you might use AMD64 architecture all the way but not the same chip 😀

 

[ivan_lee05]

First. Intel didn't trade its SSE3 for AMD64.. AMD bought SSE3 and had a cross-license agreement to intel.

Second. Intel didn't use AMD's AMD64 Architecture. instead, they convert their own x64-IA64 technology to x86-64 technology or what they called EM64T.

Third. Intel doesn't need to buy or trade their SSE3 for AMD64. because they have their own facility (R&D) to research and create EM64T.

there is also a site who benchmarked those EM64T enabled, and A64 processor.

xbit-labs

you can clearly see the result...



so stop saying AMD64 is good in 64bit and EM64T is not...

 

[cxl]

- You haven't had to change your mainboard to support more than 3GB of RAM -- many S754 mainboards had this stupid limitation despite being build around 64-bit CPU and rarely you can see desktop board supporting more than 4GB anyway.

So you might use AMD64 architecture all the way but not the same chip

You can use Vista with 1GB of ram in 64-bit mode. Sure, you will loose the main advantage (>2GB RAM), but still get improved architecture bonus.

I was using 64-bit Fedora on my s754 two years ago... (and it WAS faster).

 

[Heyyou27]

The K8 has expansion of another 8 GPRs(general purpose registers) when it is running in 64bit mode, which is not the case with Netburst.

False. Both have additional 8 GPRs and additional 8 SIMD registers.

The X86-64 is performing better on K8.

That was before Conroe hit the streets.

AMD traded x86-64 with Intel for iSSE3.

I don't know where you are getting your info but what you are saying here is purely speculative crap.

Intel had EM64T way back in the first Prescott core only it wasn't enabled until the 5x0F series showed up. Subtract approx. two years from the release of Prescott to account for the length of the CPU design cycle and you will know when Intel already had EM64T in the works. It is not an aftermath hack as many incorrectly think.

The other is the EMT64 which was pretty much a straight up copy that I believe intel actually pays AMD for the rights to use in their more recent chips (cant remember exactly).

What's the matter doc, too much crack inhaled lately? Just keep on dreaming. 😀Actually, Conroe's EM64T performance isn't any better than Netburst's was, relative to each core's 32-bit performance.

Yeah well over half of that stuff listed, the AMD64 has and the EM64T doesn't.

Shutup noob. Xeon has always ran strictly 64bit until AMD wanted to push 32/64 together. EMT64 is just another way of saying 64bit. Intel Cross license with AMD. They just push it and made people think they are better than Intel. Intel didn't push it because desktop isn't ready for 64. Even today, programs are still written in 32. 64 or 32, it will all run the same speed. Get a life noob and stop posting crap!You act like backwards compatibility is a bad thing.

First. Intel didn't trade its SSE3 for AMD64.. AMD bought SSE3 and had a cross-license agreement to intel.

Second. Intel didn't use AMD's AMD64 Architecture. instead, they convert their own x64-IA64 technology to x86-64 technology or what they called EM64T.

Third. Intel doesn't need to buy or trade their SSE3 for AMD64. because they have their own facility (R&D) to research and create EM64T.

Oh right, they can research and create something already available on the market! Too bad patent law still exists in the U.S. :roll: x86-64 was an AMD technology, but Intel has made plenty of important technologies of their own.

 

[quantumsheep]

The K8 has expansion of another 8 GPRs(general purpose registers) when it is running in 64bit mode, which is not the case with Netburst.

False. Both have additional 8 GPRs and additional 8 SIMD registers.

The X86-64 is performing better on K8.

That was before Conroe hit the streets.

AMD traded x86-64 with Intel for iSSE3.

I don't know where you are getting your info but what you are saying here is purely speculative crap.

Intel had EM64T way back in the first Prescott core only it wasn't enabled until the 5x0F series showed up. Subtract approx. two years from the release of Prescott to account for the length of the CPU design cycle and you will know when Intel already had EM64T in the works. It is not an aftermath hack as many incorrectly think.

The other is the EMT64 which was pretty much a straight up copy that I believe intel actually pays AMD for the rights to use in their more recent chips (cant remember exactly).

What's the matter doc, too much crack inhaled lately? Just keep on dreaming. 😀Actually, Conroe's EM64T performance isn't any better than Netburst's was, relative to each core's 32-bit performance.

Yeah well over half of that stuff listed, the AMD64 has and the EM64T doesn't.

Shutup noob. Xeon has always ran strictly 64bit until AMD wanted to push 32/64 together. EMT64 is just another way of saying 64bit. Intel Cross license with AMD. They just push it and made people think they are better than Intel. Intel didn't push it because desktop isn't ready for 64. Even today, programs are still written in 32. 64 or 32, it will all run the same speed. Get a life noob and stop posting crap!You act like backwards compatibility is a bad thing.

First. Intel didn't trade its SSE3 for AMD64.. AMD bought SSE3 and had a cross-license agreement to intel.

Second. Intel didn't use AMD's AMD64 Architecture. instead, they convert their own x64-IA64 technology to x86-64 technology or what they called EM64T.

Third. Intel doesn't need to buy or trade their SSE3 for AMD64. because they have their own facility (R&D) to research and create EM64T.

Oh right, they can research and create something already available on the market! Too bad patent law still exists in the U.S. :roll: x86-64 was an AMD technology, but Intel has made plenty of important technologies of their own.

Just one word comes to mind here.....OWNED!

 

[dmantech]

They are really the same in 64bit instructions, the only difference is in the cache. Intel Core Duo cores share cache and AMD Cores have their own per core.

 

[quantumsheep]

They are really the same in 64bit instructions, the only difference is in the cache. Intel Core Duo cores share cache and AMD Cores have their own per core.

How does Cache have anything to do with 64 bit instructions?

 

[gman01]

Even today, programs are still written in 32.

Programs are not written in 32 or 64 or any other bit... You just write the programs, and then use a 32 or 64 bit compiler....

My operating system, and all my programs are compiled in 64 bit(except 1).... People think it is a big deal because microsoft is 3 years late to the game.... Their problems with vista has nothing to do with it being 64 bits....

 

[spud]

Pointer values essentially dictate whether the processor is 64bit or 32bit.

 

[spud]

Second. Intel didn't use AMD's AMD64 Architecture. instead, they convert their own x64-IA64 technology to x86-64 technology or what they called EM64T.

Actually Intel did use AMD64, the only difference is Intel followed AMD's specifications to a tee while AMD omitted a few instructions. This was the reason Windows XP 64 initially did not work with Intel's implementation of AMD64. But AMD64 and EM64T are buzz words for x86-64; there isn't a real fundamental difference between the two.

 

[PorkyTheClown]

Pentium MMX 166@233
32MB PC66 SDRAM CAS2
ATI Rage II+ 2M
200W Generic AT PSU
4X Sony CD-ROM
1275MB Seagate 3600rpm



Hell yeah evilr00t!

 

[quantumsheep]

Even today, programs are still written in 32.

Programs are not written in 32 or 64 or any other bit... You just write the programs, and then use a 32 or 64 bit compiler....

My operating system, and all my programs are compiled in 64 bit(except 1).... People think it is a big deal because microsoft is 3 years late to the game.... Their problems with vista has nothing to do with it being 64 bits....

Good points there. Is it feasible to convert a program that uses a 32 Bit compiler to use a 64bit compiler or vice versa?

If that is so it seems the majority of software developers are being quite lazy.

 

[gradius]

Well there's a simple answer here. They're both exactly the same. The reason for this is basically 'cos microsoft is lazy and couldn't be bothered to develop for 2 different 64bit instruction sets

Are you serious? Do you even know how redundant and timewasting it would be to develop for two instruction sets?

 

[quantumsheep]

Well there's a simple answer here. They're both exactly the same. The reason for this is basically 'cos microsoft is lazy and couldn't be bothered to develop for 2 different 64bit instruction sets

Are you serious? Do you even know how redundant and timewasting it would be to develop for two instruction sets?

Yes i admit to it, it would be a pointless exercise, but it COULD have happened if Intel chose to use a different instruction set.

 

[gOJDO]

The K8 has expansion of another 8 GPRs(general purpose registers) when it is running in 64bit mode, which is not the case with Netburst.

False. Both have additional 8 GPRs and additional 8 SIMD registers.

I am not sure how many GPRs are missing in the EM64T implementation: SYSCFG, TOP_MEM, TOP_MEM2. Thats why the additional 8 GPRs is not the case for EM64T.

The X86-64 is performing better on K8.

That was before Conroe hit the streets.
Ture, I forgot about Conroe

AMD traded x86-64 with Intel for iSSE3.

I don't know where you are getting your info but what you are saying here is purely speculative crap.
They have agreement for reverse engineering or they have traded the x86-64 for iSSE3.

Intel had EM64T way back in the first Prescott core only it wasn't enabled until the 5x0F series showed up. Subtract approx. two years from the release of Prescott to account for the length of the CPU design cycle and you will know when Intel already had EM64T in the works. It is not an aftermath hack as many incorrectly think.

Maybe, what stoped Intel to release such feature earlier?

 

[jamesgoddard]

Well there's a simple answer here. They're both exactly the same. The reason for this is basically 'cos microsoft is lazy and couldn't be bothered to develop for 2 different 64bit instruction sets

Are you serious? Do you even know how redundant and timewasting it would be to develop for two instruction sets?

Microsoft already code for 2 separate 64bit instruction sets!!!!! The 2 sets in question are the Intel 64bit set and the AMD 64bit set, and they are COMPLETLY different......

OK to understand this you have to understand that since 1999 Intel have been trying to push the world + dog into using the EPIC 64bit set (aka Itanium)... This whole project has been probably the single biggest mistake Intel ever made, the have spent billions and billions of this project and ended up with just about nothing. Microsoft obliged in creating Windows server and XP for Itanuim, these products are still available today, just no ah heck ever uses them. Intels origanal roadmaps had x86 dead and buried by now.

Then, AMD came along and INVENTED AMD64, again Microsoft has obliged and created 64bit Windows for this also. Where the Intel Itanium was a complete failure, the AMD64 started to take off big time, mainly because of the true 32bit compatibility, and ease of coding with the old x86 like code. Intel were in serious trouble by this point, they were up the creak without a paddle.

Intel then tried to fight back and decided to create a new 64bit set to rival AMD's but this time Microsoft stood firm, they had had there chance as the 64bit game with Itanium and Microsoft was not going to oblige Intel for a second time, Microsoft told Intel to adopt AMD64, which they grudgingly did.

To do this Intel used the cross licensing deal and got AMD to send over the specification documentation for AMD64, Intel then worked away for 6 months or more and added 64bit to the net burst architecture with a new CPU stepping, the implementation of EMT64 was a perfect copy of the documentation AMD had sent over some time back. What AMD had 'forgotten' to tell Intel was that a few extra instructions had been added to the AMD64 set after the version of the document that Intel had worked with was released, and oh dear the Intel’s had a few compatibility issues (nice job AND 😉 ) Intel then went back and added these extra instructions, but still you find some very early Intel EMT64 chips were released without the full compatibility.

PS re comments about selling / swapping technology - IE AMD64 for SSE - this is cobblers as the tech cross license deal covers all this. The reason for the cross license deal goes back to the very early days of x86 where Intel were keen to make x86 the de-facto instruction set, and to do this the best way was to get all the other players making the same x86 chips as they do... In later years this deal also held off antitrust laws as Intel never have achieved total monopoly, funny fact of the matter is that now Intel NEED AMD to survive to ensure the long term stability of Intel as a whole, if Intel were ever to get too bid US laws would step in like the baby bells in the 70's and we would say bye bye to Intel as we know it...

 

[djkrypplephite]

AMD developed the x86-64 instruction set and started the 64-bit thing (not counting the Itanium).

Yeah, forget about Alpha, PowerPC, Power, SPARC64, UltraSPARC, IBM's z, HP PA-RISC, MIPS, etc.

I'm afriad to tell you, it was MIPS who invented 64-bit processors back in 1991.

Plus, 64-bit is for sissies. 128-bit is where it's at. I run the entire IPV6 system, so you know, I have a 128-bit system.

 

[quantumsheep]

AMD developed the x86-64 instruction set and started the 64-bit thing (not counting the Itanium).

Yeah, forget about Alpha, PowerPC, Power, SPARC64, UltraSPARC, IBM's z, HP PA-RISC, MIPS, etc.

I'm afriad to tell you, it was MIPS who invented 64-bit processors back in 1991.

Plus, 64-bit is for sissies. 128-bit is where it's at. I run the entire IPV6 system, so you know, I have a 128-bit system.

128 Bit system?!?!?! Bet that thing has some serious power.

 

[djkrypplephite]

Actually, 128-bit systems technically speaking ARE in use. A lot of newer, if not ALL modern processors use 128-bit FP registers, and OS/400 has used 128-bit pointers for, like, ever.