Does Conroe emulate 64-bit or not?

Dmens , I have a question.

There is a rumor which says the following : Intel did apparently made a 64bit extension to x86 of their own ( Yamhill project ).AMD got involved with MS sooner and had their backing.When Intel came with their different 64bit extension MS refused to support 2 different implementations and asked Intel to make their AMD64 compatible.As a result they had to drastically modify Yamhill to turn it into Prescott when the design was finished already.
Is there any truth in this ?
 
xFlankerx said:
I have a relatively simple question - Does the Conroe processor still emulate 64-bit capability like the Pentium 4s, or is it now a truly 64-bit processor like AMD's Athlon 64s?
Let me couter this with a question for you: does it matter? My processor is a black box, it consumes instructions and data and produces a result. I do not care how it is done internally, provided that it provides the performance that I want.

In the end, AMD could have a monkey sitting inside of my CPU pushing random buttons, if it outperforms my current processor, I'd buy it. Honestly, I cannot come up with a reason why it would matter whether the P4 is working in 64 bits internally or not. In the end, what should matter is that you get the correct result as fast as possible. Even if you were using some instruction for timing purposes, you can always consult the manual, which will tell you the execution time. Once again you don't need to know how it works internally.
 
drizzt81 said:
Let me couter this with a question for you: does it matter? My processor is a black box, it consumes instructions and data and produces a result. I do not care how it is done internally, provided that it provides the performance that I want.

In the end, AMD could have a monkey sitting inside of my CPU pushing random buttons, if it outperforms my current processor, I'd buy it. Honestly, I cannot come up with a reason why it would matter whether the P4 is working in 64 bits internally or not. In the end, what should matter is that you get the correct result as fast as possible. Even if you were using some instruction for timing purposes, you can always consult the manual, which will tell you the execution time. Once again you don't need to know how it works internally.

THIS is the problem;

In other words, they aren't true 64 bit processors and thus as a result they bottleneck themselves as they have to keep emulating code that they can't really process in its native mode.

Assuming Vista is the resource hog it is, we'll have to shift to 64 bit processes in order to obtain a larger addressing size (2^64 bytes). This will basically result in an internal memory bandwidth bottleneck in any EM64T processor as it has to keep switching between two instruction sets.

Conroe outperforms Athlon 64 in 32-bit applications, thats today. When Vista comes out, if thats true -- Conroe is screwed. Which is why I still wait for this bickering to end and to get a freakin answer.
 
xFlankerx said:
THIS is the problem:
In other words, they aren't true 64 bit processors and thus as a result they bottleneck themselves as they have to keep emulating code that they can't really process in its native mode.

Assuming Vista is the resource hog it is, we'll have to shift to 64 bit processes in order to obtain a larger addressing size (2^64 bytes). This will basically result in an internal memory bandwidth bottleneck in any EM64T processor as it has to keep switching between two instruction sets.
Conroe outperforms Athlon 64 in 32-bit applications, thats today. When Vista comes out, if thats true -- Conroe is screwed. Which is why I still wait for this bickering to end and to get a freakin answer.
Once agian, it does not matter whether Conroe is a "true" 64-bit CPU, it matters whether it perfoms adequately when running 64-bit software.

On page 1 of this thread, someone gave you Woodcrest (Conroe Server) benchmarks:
http://techreport.com/etc/2006q2/woodcrest/index.x?pg=5


I'm not sure one could say either CPU responds better to 64-bit binaries; the Opteron gains more in the rendering test, but the Woodcrest gains more in the shading test.

Looking at the benches, Conroe improves when running a 64-bit executable. Given that one assumes that these benchmarks are real, it seems that Conroe is able to run 64-bit code faster than 32 bit code. It also runs both the 32 and the 64-bit code faster than the Opteron system that they compared it to.

I do not know what software you will be running in 3 years, and few other people know that either. By that time, there will have been changes to the compiler, a different OS and certainly at least 1-2 revisions of CPUs. It really does not make sense to worry about the performance of your CPU in the far future, when you are buying now.

So let me summarize this for you:
It does not matter how a CPU works internally. As a user of a CPU you should be interested in perfomance, not the technical implementation. Benchmarks have shown that the Conroe architecture is able to execture 64-bit code faster than the equivalent 32-bit code. In both instances the Conroe outperforms the Opteron system that it was compared with. So the performance is available in 64-bit apps.
 
xFlankerx said:
In other words, they aren't true 64 bit processors and thus as a result they bottleneck themselves as they have to keep emulating code that they can't really process in its native mode.

Assuming Vista is the resource hog it is, we'll have to shift to 64 bit processes in order to obtain a larger addressing size (2^64 bytes). This will basically result in an internal memory bandwidth bottleneck in any EM64T processor as it has to keep switching between two instruction sets.

- Wrong on "true 64 bit processors" (what is that?)
- Wrong on bottleneck (what bus?)
- Wrong on emulating (how do you emulate a definition?)
- Wrong on "addressing size" having an effect on performance (cutting out the latency of large memories, the pipestage delay from decode to memory access in the CPU would still be the same, regardless of addressable space)
- Wrong on "internal memory bandwidth" (where's the bus?)
- Wrong on "switching between two instruction sets" (the backend is not aware of x86, the frontend doesn't do anything fancy between the old and new).

It's all BS... mmmkay?
 
drizzt81 said:
Once agian, it does not matter whether Conroe is a "true" 64-bit CPU, it matters whether it perfoms adequately when running 64-bit software.

So let me summarize this for you:
It does not matter how a CPU works internally. As a user of a CPU you should be interested in perfomance, not the technical implementation. Benchmarks have shown that the Conroe architecture is able to execture 64-bit code faster than the equivalent 32-bit code. In both instances the Conroe outperforms the Opteron system that it was compared with. So the performance is available in 64-bit apps.

The difference is instead of being 30% faster in 32bit it will be 25% faster in the 64bit version of that same app, kind of moot isn't it?

I found it kind of funny that the same folks who absolute hate Netburst and Love 64bit can't see where Netburst seems better at dealing with 64bit than X2 or Core 2 Duo, that doesn't make it faster LOL! Gee I'd have loved to have seen Yamhill and 90nm on Northwood C? 20 is still long.
 
xFlankerx said:
THIS is the problem;



Conroe outperforms Athlon 64 in 32-bit applications, thats today. When Vista comes out, if thats true -- Conroe is screwed. Which is why I still wait for this bickering to end and to get a freakin answer.

http://www.hardforum.com/showpost.php?p=1029629164&postcount=109

introduction of a macro-op fusion, where two instructions are executed simultaneously, will provide enhancements of floating point operations. The design of such implementation in Core 2 is based on 32-bit. But 64-bit environment contains less operation, Fusion is less useful. The case is similar to K8. Netburst however is an outstanding in the performance gain in 64-bit environment. The Trace Cache design in Pentium 4 benefits a longer instruction, in which will offer another 10-15% boost in 64-bit environment. In conclusion, Intel Core 2 family would provide a very good performance in 64-bit as in 32-bit, yet it’s no more gain between twos.

If Conroe is screwed, the X2 is as well please note* "But 64-bit environment contains less operation, Fusion is less useful. The case is similar to K8.". I posted what I did above so that it wouldn't be taken out of context. dmens is more than likely right about the reason, but the end result still is very clear.

Vista has 32 and 64 bit modes BTW. I sticking with WinXP until MS kills it off. Then I'll get off my ass and put more time into Linux.
 
xFlankerx said:
THIS is the problem;

Conroe outperforms Athlon 64 in 32-bit applications, thats today. When Vista comes out, if thats true -- Conroe is screwed. Which is why I still wait for this bickering to end and to get a freakin answer.

If the point was made well enough.

http://www.digit-life.com/articles2/cpu/intel-conroe-2-13-ghz.html
Software
Windows XP Professional x64 Edition SP1.
3ds max 7.0
Maya 6.5
Lightwave 8.5 x64 Edition WinRAR 3.51
7-Zip 4.32 x64 Edition
LAME 3.98
Monkey Audio 4.01
OGG Encoder 2.8 (Lancer)
Windows Media Encoder 9 x64 Edition
MATLAB 7.1
Pro/ENGINEER Wildfire 2.0
SolidWorks 2005
Microsoft Visual C++ Professional 6.0
CPU RightMark 2005 Lite x64 Edition
F.E.A.R. 1.3
Half-Life 2
Unreal Tournament 2004 build 3339
Quake 4 Point Release 1.1
FineReader Professional 8.0
Adobe Photoshop CS2 (9.0)
Canopus ProCoder 2.01.30
DivX 6.1.1
Windows Media Video 9 VCM
x264 v.438
XviD 1.1.0 Release
Apache 2.0.55 for Windows
 
Donnie27 said:


Oy, which is exactly what I was aying to begin with. You started the bashing on altivec. I just stated htat it has its place, and in many ways is far better then SSE.
 
dmens said:
But since you seem to think that there are "elements" which are common on both P4 and K8, care to extrapolate which ones matter for 64-bit operation?

Transistors. :D
 
duby229 said:
Oy, which is exactly what I was aying to begin with. You started the bashing on altivec. I just stated htat it has its place, and in many ways is far better then SSE.

No I didn't. Dewd, give it up already. I said anyting can be coded for and they prove it. That means it was SIMD dependent, not that SIMD (Even Altivec) sucks. ;) You're making another moot statement. I didn't start any freakin' arguement with you over if everything has its place or not. Or if one SIMD was better than the other or not.
 
Very nice to see dmens has joined in finally(a fine addition.He beat the tar out of duvie as did I.) Most you guys know how I feel . So here's the straight. Intel conroe isn't as good at 64. tech as amd. Intel basicly says so. But they also say Penryn will be better than AMD 64. Want links look for them.

Really guys on the desktop 64 bit has shown me zero. But it did wonders for amd e-penis morons. 64 bit is cool and all that . but most these coders out there can't do 32 bit well.

P4 netburst Intel had a vision Ahead of the tech that would actually make good on that vision. So intel falls back on the old Pro series and plays around with it. Low and behold its better than AMD 's current offerings. Look at K8L the only improvements made are copies of Intel merom . This RH/T softwear surprize . Is laughable . Mitosis on the otherhand is very serious shit.


Any way nice to see you here dmens.
 
$BangforThe$ said:
Really guys on the desktop 64 bit has shown me zero. But it did wonders for amd e-penis morons. 64 bit is cool and all that . but most these coders out there can't do 32 bit well.
It's shown you zero because you haven't bothered to look into it.
http://anandtech.com/linux/showdoc.aspx?i=2127&p=2

Some 64bit linux programs that deal in encoding, rendering, gaming, and SQL (sql isnt exactly a desktop ap, of course). 64bit was consistantly as fast or faster. Look at the MP3 encoding and POV-Ray! Pretty impressive. Gaming benchmarks are kind of surprising, actually.
 
dmens said:
Really? Where?

To find . Go back to the first time Intel publicly displayed conroe . than reread the arch. topics . Intel states that 64bit for conroe is the same as it is for the p4D cores. But than its stated it will be updated with penryn. I really didn't pay much attention to it . Its just recall. I and millions upon million use ms only so until 64 bit has shown something on the desktop ms os than 64 bit tech is like a bad taste in mouth after a nites sleep. Brush well and it goes away.
 
$BangforThe$ said:
Really guys on the desktop 64 bit has shown me zero. But it did wonders for amd e-penis morons. 64 bit is cool and all that . but most these coders out there can't do 32 bit well.
I'm seeing 15-22% speed improvements in audio and video encoding in Vista Beta 2.
 
$BangforThe$ said:
Really guys on the desktop 64 bit has shown me zero. But it did wonders for amd e-penis morons. 64 bit is cool and all that . but most these coders out there can't do 32 bit well.
From my understanding, it's more a compiler than a programmer issue. In the end, I do not think that the majority of programs is written in assembly, but rather some higher-level language. The compiler converts them into executables. Depending on the flags, they will either be 32 or 64-bit. Any "well written" program should be able to be compiled to create either executable.

How does this relate to the topic though? Let me remind you of the question:
"Does the Conroe processor still emulate 64-bit capability like the Pentium 4s, or is it now a truly 64-bit processor like AMD's Athlon 64s?"

which has been answered like 4000 times already: "It does not matter. [some links] show benchmarks indicating that Conroe can outperform an AMD solution, when running 64-bit applications. Who care about the implementation, if the performance is excellent?"
 
drizzt81 said:
.....which has been answered like 4000 times already: "It does not matter. [some links] show benchmarks indicating that Conroe can outperform an AMD solution, when running 64-bit applications. Who care about the implementation, if the performance is excellent?"

QFT! As well as quoted for common f#$%$@& sense! :D Yes, VIsta isn't going to change the outcome.
 
Sorry, I am trying to clear doubts in my mind. The present applications are mostly 32 bit for Windows XP. How good these applications are running under the 64 bit OS (i.e. Vista 64 Bit and XP 64 bit) as compared with 32 bit OS (say Windows XP)? Say you have Conroe (with EM64T) and AMD64 and you have Windows XP64 and the forthcoming Vista 64, will these 64-bit setups run 32 bit applications better (by how much?) or worse (by how much?)?

Are we here buying 64 bit processors running 64 bit OS now and waiting for the 64 bit applications yet to be developed?
 
Sam Ontario said:
Sorry, I am trying to clear doubts in my mind. The present applications are mostly 32 bit for Windows XP. How good these applications are running under the 64 bit OS (i.e. Vista 64 Bit and XP 64 bit) as compared with 32 bit OS (say Windows XP)?
Depends a lot of the application I'd think. Probably somewhere from "the same" to "slightly faster" to "much slower/ not at all".

Sam Ontario said:
Say you have Conroe (with EM64T) and AMD64 and you have Windows XP64 and the forthcoming Vista 64, will these 64-bit setups run 32 bit applications better (by how much?) or worse (by how much?)?
depending on how good MS's WoW (windows-on-windows) layer is, they'll run just about the same.

Sam Ontario said:
Are we here buying 64 bit processors running 64 bit OS now and waiting for the 64 bit applications yet to be developed?
Yes, but without a 64-bit OS, there couldn't be any 64-bit apps. I am sure that once 64-bit becomes more mainstream, more and more apps will offer a recompiled, 64-bit update.
 
xFlankerx said:
Which is why I still wait for this bickering to end and to get a freakin answer.

How about this:

The discussion whether or not EM64T and AMD64 really are 64-bit processors can be confusing to some because Intel calls its architecture “Extended Memory 64 Technology”. We know that it extends the IA32 instruction set. Therefore, are EM64T and AMD64 processors “real” 64-bit chips? The answer is absolutely yes. When these processors operate in 64-bit mode, the addresses are 64-bit, the GPRs are 64 bits wide, and the ALUs are able to process data in 64-bit chunks. Therefore, these processors are full-fledged 64-bit processors in this mode.

From ibm.com
 
Back
Top