64 bit Cpus, what are the fine details?

[hiddendog]

whilst in theory i agree that he should search, in practice these forums are about the sharing of info. if someone already understands what he is asking then he is more likely to get an answer here that he understands than some technical paper in the subject he finds by googling.

I totally agree what your are said. And personally, I think he is very smart. No need to use the browser & but can get the answer.

 

[spud]

I think that what jDOG is afraid of is this.
Currently, AMD processors have a clear execution advatage over Intel, when using 64 bit kernal, under linux. Since Core2 duo is not expected to have better 64 bit execution than P4, this advantage will be taken to Vista.
In other words, K8 will outperform conroe, once M$ brings out thier new OS.
Personnally, I dont believe, or disbelieve. I want to see it before I decide.

Who expects this, You? Talk about being a hypocrite; you constantly claim that the engineering samples of the Conroe aren't to be taken seriously because it isn't the official release. Yet here you are saying you know what the performance delta will be on Vista in 64bit mode, and no one has seen Conroe's 64bit performance so now you sound like 9-inch guessing at what you hope will be the ideal situation.

 

[sailer]

I think I'll try an answer that should bring some perspective to the 32 bit-64 bit question. 64 bit is not just twice as large as 32 bit. It also relates to how much memory the chip can access. A 32 bit system can access about 4.3 billion bytes of information in memory (2 to the 32 power= 4.3e9). A 64 bit system can access about 18 million terabytes of information in memory (2 to the 64 power =1.8e19). No programs are available that I know of which take advantage of the true power of a 64 bit system. I find the idea of 18 million terabytes mindboggling. At the same time, it might allow some real holographics in video imaging (think of the holodeck from Star Trek).

Check http://artechnica.com/cpu/03q1/x86*64/x86-64-2.html for further technical description.

Another way of comparing it might be to compare one of the original video games like Pong to a new one like Oblivion. Such might be the comparison of Oblivion to a game possible using the full power of a 64 bit processor/operating system. But all that's in the future and depends on the gaming industry to write programs to use the 64 bit system. Oh yes, if you think the games have bugs now on 32 bit systems, think of how buggy they could get when they fully use a 64 bit system.

 

[turpit]

I think that what jDOG is afraid of is this.
Currently, AMD processors have a clear execution advatage over Intel, when using 64 bit kernal, under linux. Since Core2 duo is not expected to have better 64 bit execution than P4, this advantage will be taken to Vista.
In other words, K8 will outperform conroe, once M$ brings out thier new OS.
Personnally, I dont believe, or disbelieve. I want to see it before I decide.

Who expects this, You? Talk about being a hypocrite; you constantly claim that the engineering samples of the Conroe aren't to be taken seriously because it isn't the official release. Yet here you are saying you know what the performance delta will be on Vista in 64bit mode, and no one has seen Conroe's 64bit performance so now you sound like 9-inch guessing at what you hope will be the ideal situation.

Speaking of 9inch, I havent had the mis pleasure of seeing any posts from him or or Mad Mod Mike lately. Have they been kicked, or did their mommies just ground them?

 

[endyen]

Talk about being a hypocrite;

Oh, so you have to believe before you can comment on conroe. Let me tell you what I believe.
1 Benchmarks. Had you put the latest XE chip up against that FX60, it would have done almost as well. Does that mean P4s are better than K8s?
2 The "FX62" Well, it seems that one of the Intel PR team has already linked to a sight that shows thier own non-OCed FX60 beats Intel's OCed version.
3 The AMD mobo. Okay, so the FX only looses a couple of persent there.
4 Algorithms It seems that Victor Wang's sciencemark2 score may have been aided by some non-standard algorithms

like I said, that lately SM result is with some unofficial binaries (I haven't them, and I don't know if they are available for download, at least, I've not found them).

link Wonder where he got them? and if maybe an overzelous Intel tech installed similar on the conroe benchmark machine.
5 No, they did not compensate for AM2 at all
6 64 bit. Anyone who does not accept that the A64s are better than the P4s at 64 bit, has thier head in the ground. I have seen nothing to suggest that Intel has enhanced thier 64 bit perf on conroe. I am reasonably sure that had Intel done anything along those lines, one of the Intel PR team here would have picked up on it. You would be a prime source, since you did so much investigative reporting on prescott.

None of that proves a gd thing. I still expect that conroe will be a ggodly chip. I fully expect to get one in the fall. However, I wont know for sure, until I know for sure how well it truly performs.

 

[endyen]

Talk about being a hypocrite;

Oh, so you have to believe before you can comment on conroe. Let me tell you what I believe.
1 Benchmarks. Had you put the latest XE chip up against that FX60, it would have done almost as well. Does that mean P4s are better than K8s?
2 The "FX62" Well, it seems that one of the Intel PR team has already linked to a sight that shows thier own non-OCed FX60 beats Intel's OCed version.
3 The AMD mobo. Okay, so the FX only looses a couple of persent there.
4 Algorithms It seems that Victor Wang's sciencemark2 score may have been aided by some non-standard algorithms

like I said, that lately SM result is with some unofficial binaries (I haven't them, and I don't know if they are available for download, at least, I've not found them).

link Wonder where he got them? and if maybe an overzelous Intel tech installed similar on the conroe benchmark machine.
5 No, they did not compensate for AM2 at all
6 64 bit. Anyone who does not accept that the A64s are better than the P4s at 64 bit, has thier head in the ground. I have seen nothing to suggest that Intel has enhanced thier 64 bit perf on conroe. I am reasonably sure that had Intel done anything along those lines, one of the Intel PR team here would have picked up on it. You would be a prime source, since you did so much investigative reporting on prescott.

None of that proves a gd thing. I still expect that conroe will be a goodly chip. I fully expect to get one in the fall. However, I wont know for sure, until I know for sure how well it truly performs.

 

[Vampire_Lestat]

Oh boy another stuck up Moron who trying to act so cool. Grow up SidVicious. This dont make you look cool. If you want to help People then help people. If you dont. Then shut the Hell up and live with your mommy.

See MatTheMurdera asking a honest question. I think we need more Mods on the forums. For this is a place where people should get help.

 

[spud]

Talk about being a hypocrite;

Oh, so you have to believe before you can comment on conroe. Let me tell you what I believe.
1 Benchmarks. Had you put the latest XE chip up against that FX60, it would have done almost as well. Does that mean P4s are better than K8s?
2 The "FX62" Well, it seems that one of the Intel PR team has already linked to a sight that shows thier own non-OCed FX60 beats Intel's OCed version.
3 The AMD mobo. Okay, so the FX only looses a couple of persent there.
4 Algorithms It seems that Victor Wang's sciencemark2 score may have been aided by some non-standard algorithms

like I said, that lately SM result is with some unofficial binaries (I haven't them, and I don't know if they are available for download, at least, I've not found them).

link Wonder where he got them? and if maybe an overzelous Intel tech installed similar on the conroe benchmark machine.
5 No, they did not compensate for AM2 at all
6 64 bit. Anyone who does not accept that the A64s are better than the P4s at 64 bit, has thier head in the ground. I have seen nothing to suggest that Intel has enhanced thier 64 bit perf on conroe. I am reasonably sure that had Intel done anything along those lines, one of the Intel PR team here would have picked up on it. You would be a prime source, since you did so much investigative reporting on prescott.

None of that proves a gd thing. I still expect that conroe will be a ggodly chip. I fully expect to get one in the fall. However, I wont know for sure, until I know for sure how well it truly performs.

1. What are you talking about I never mentioned anything about the Pentium 4?
2. Again what are you talking about?
3. ?
4. Conspiracy theories again?
5. AM2 is a platform that’s AMD's failure not the users.
6. We all know K8's are better at 32bit and 64bit why do you have to rehash it every single day? As per 64 bit mode, you have also seen nothing that suggests anything in regards to either of our arguments, because no one has seen anything on the subject. I'm also reasonably sure you have no reasons to base your unreasonable assumptions based on that fact you have no reasonable amount of information to say yes or no on the 64bit performance.

Again with the Prescott good god man I had no clue they were adding additional stages, I had no clue the software used to redesign the Prescott failed miserably, I had no clue they increased latencies on the L1 and L2 caches, all I knew was what was provided by Intel and online sources which appeared to for all purposes feasible and reasonable assessments based on that information.

At this point I don't have a clue what the 64bit performance will be like on the Core processors and frankly neither do you for that matter. One can be certain there will be a performance increase due to the change in architecture, but no one can say for sure.

Also what if it sucks then what? What changes in this dominate 32bit machine market? What possible outcome is possible if it sucks so bad at 64bit it makes baby Jesus cry? Nothing because we are still in a very dominate 32bit market and well be that way for years to come regardless of Vista being 64bit or not.

 

[endyen]

In a nutshell, I just showed you that the conroe will be on par or slightly better than the real FX62, until vista comes into play.
Ever hear the expression "fool me once, shames on you. Fool me twice shame on me?" Intel is about to fool you again. Pity.
As for vista, there should be a major delta going to long mode. It will be nice that windows will be basicly held to the high registers, on one core. Well at least for AMD. Remember, Intel hasn't set up thier EM64T that way. Access to the extra SSE unit will add some as well.
AMD should gain about 20% from a workable 64 bit OS, just like it does with linux.
So, yup Intel will have a paper crown till Vista. Three cheers.

 

[Action_Man]

Proof?

 

[davidflet9]

could you put a link or some infomation up on where you are getting your data on 64bit performance for both AMD and Intel processors please, it should make for some interresting reading
cheers
PS i know this is in reply to action man but but a question for all!

 

[Da_Banig]

A short question like this actually has two pages 8O

 

[TechnologyCoordinator]

I might get flamed for being helpful and on-topic, but here it goes:

http://en.wikipedia.org/wiki/64-bit
http://en.wikipedia.org/wiki/32-bit
http://en.wikipedia.org/wiki/16-bit
http://en.wikipedia.org/wiki/8-bit

That ought to keep you busy for a while!

From Wikipedia:

A change from a 32-bit to a 64-bit architecture is a fundamental alteration, as most operating systems must be extensively modified to take advantage of the new architecture. Other software must also be ported to use the new capabilities; older software is usually supported through either a hardware compatibility mode (in which the new processors support an older 32-bit instruction set as well as the new modes), through software emulation, or by the actual implementation of a 32-bit processor core within the 64-bit processor die (as with the Itanium processors from Intel, which include an x86 processor core to run 32-bit x86 applications). One significant exception to this is the AS/400, whose software runs on a virtual ISA, called TIMI (Technology Independent Machine Interface) which is translated to native machine code by low-level software before being executed. The low-level software is all that has to be rewritten to move the entire OS and all software to a new platform, such as when IBM transitioned their line from the older 32/48-bit "IMPI" instruction set to 64-bit PowerPC (IMPI wasn't anything like 32-bit PowerPC, so this was an even bigger transition than from a 32-bit version of an instruction set to a 64-bit version of the same instruction set). Another significant exception is IBM z/Architecture, which readily handles applications concurrently with different addressing expectations (24, 31, and 64 bit).

While 64-bit architectures indisputably make working with huge data sets in applications such as digital video, scientific computing, and large databases easier, there has been considerable debate as to whether they or their 32-bit compatibility modes will be faster than comparably-priced 32-bit systems for other tasks.

Theoretically, some programs could well be slower in 64-bit mode. Under some architectures, instructions for 64-bit computing take up more storage space than the earlier 32-bit ones, so it is possible that some 32-bit programs will fit into the CPU's high-speed cache while equivalent 64-bit programs will not. In basic terms moving 64 bits at a time to perform otherwise 32 bit work simply requires more processing effort to/from memory. A common argument is that, in applications like scientific computing, the data being processed often fits naturally in 64-bit chunks corresponding to double-precision floating-point types, and will be faster on a 64-bit architecture because the CPU will be designed to process such information directly rather than requiring the program to perform multiple steps — this is erroneous, however, because most 32-bit CPUs already have a 64-bit wide data bus and 64-bit registers for floating-point quantities. The only speed advantages come for manipulating 64-bit integer quantities, but this is rarely a performance-limiting task even for applications (such as large-file I/O) that require such manipulations.

All performance assessments are complicated, however, by the fact that in the process of designing the new 64-bit architectures, the instruction set designers have also taken the opportunity to make other changes that address some of the deficiencies in older instruction sets by adding new performance-enhancing facilities (such as the extra registers in the AMD64 design).

 

[TechnologyCoordinator]

stealing my thunder eh i already posted a wikipedia link about 64-bit.

Guess I stopped reading after the flame-fest started, eh?

I, TechnologyCoordinator, officially give all the thunder to strangest stranger I (don't) know: StrangeStranger

But you've got to admit that my posting of links to the 32-bit, 16-bit, and 8-bit articles as well as quoting some text from the 64-bit article was truly ingenious.

 

[windego]

Easy 64 bit cpus can process 64 bit instructions which are twice as long as 32 bit.

Incorrect. The word size is double that of 32 bit. This means that chips add additional 64 bit registers. The instruction size is not 64bit, it's simply the addressable width of pointers, and the maximum width at which basic operations can be performed.

 

[spud]

In a nutshell, I just showed you that the conroe will be on par or slightly better than the real FX62, until vista comes into play.
Ever hear the expression "fool me once, shames on you. Fool me twice shame on me?" Intel is about to fool you again. Pity.
As for vista, there should be a major delta going to long mode. It will be nice that windows will be basicly held to the high registers, on one core. Well at least for AMD. Remember, Intel hasn't set up thier EM64T that way. Access to the extra SSE unit will add some as well.
AMD should gain about 20% from a workable 64 bit OS, just like it does with linux.
So, yup Intel will have a paper crown till Vista. Three cheers.

Ok let’s start by apologizing, I snapped in that post needlessly, I just got frustrated.

As per showing me that a Conroe will be equal or slightly faster than a FX62, I don't see it that way. Because at this point you and me and been bickering over opinions not facts.

I have heard the expression but there is a slight difference this time around. Benchmarks of engineering samples are available, which was not the case for the Prescott.

The long mode comment I don't know to argue the point because I don't know how Intel has implemented EM64T on the Conroe but I personally I care very little about the 64bit performance or potential. With regards to the fact I have never been supportive of 64bit machines at this point in time. The last 3 years have shown how serious the industry is into the transition.

 

[MatTheMurdera]

My gosh. I posted this a few weeks ago, people are still on it?

 

[4745454b]

Hey all, got some more fuel for this fire....

Who here knows that emt64 and AMD64 are (almost) the same? The only difference in EMT64 and AMD64 is that Intels EMT64 includes support for Hyperthreading, while AMDs doesn't. This is so that EMT64 will/would work with "current" p4 CPUs. Both chips aren't "real" 64bit CPUs. They still process things internally as 32bit CPUs. AMD simply added 64bit memory registers to their chips. Some reading for anyone who wants to... http://www.hardwareanalysis.com/content/topic/37587 Don't forget to read the links to MS and Intels websites.

 

[SexBomb]

I think that what jDOG is afraid of is this.
Currently, AMD processors have a clear execution advatage over Intel, when using 64 bit kernal, under linux. Since Core2 duo is not expected to have better 64 bit execution than P4, this advantage will be taken to Vista.
In other words, K8 will outperform conroe, once M$ brings out thier new OS.
Personnally, I dont believe, or disbelieve. I want to see it before I decide.

Who expects this, You? Talk about being a hypocrite; you constantly claim that the engineering samples of the Conroe aren't to be taken seriously because it isn't the official release. Yet here you are saying you know what the performance delta will be on Vista in 64bit mode, and no one has seen Conroe's 64bit performance so now you sound like 9-inch guessing at what you hope will be the ideal situation.

It's not a big mistery to know that Intel's upcoming processors will not perform as expected once running in 64-bit mode.
When you see a company like Intel Shouting-Out-Loud about all the enhancements on their new architecture like 4-issue, macro and micro-ops fusion, memory disambiguation, etc, etc, but still DON'T say anything about 64-bit performace, then you start to think hard about it.
Even Hannibal (the guy from Arstechnica) in his extensive review of the core architecture didn't mention any "enhancements" or "additions" to intel's EMT64.

Conroe might perform good on plain old 32-bit apps, but once you put in 64-bit on the table things will start to look nasty for intel.

 

[Action_Man]

So basically your proof is a bunch of assumptions.

 

[Action_Man]

:lol: You win the internets. Well at least he read Hannibals article this time, its some progress, very little but something.

 

[spud]

I think that what jDOG is afraid of is this.
Currently, AMD processors have a clear execution advatage over Intel, when using 64 bit kernal, under linux. Since Core2 duo is not expected to have better 64 bit execution than P4, this advantage will be taken to Vista.
In other words, K8 will outperform conroe, once M$ brings out thier new OS.
Personnally, I dont believe, or disbelieve. I want to see it before I decide.

Who expects this, You? Talk about being a hypocrite; you constantly claim that the engineering samples of the Conroe aren't to be taken seriously because it isn't the official release. Yet here you are saying you know what the performance delta will be on Vista in 64bit mode, and no one has seen Conroe's 64bit performance so now you sound like 9-inch guessing at what you hope will be the ideal situation.

It's not a big mistery to know that Intel's upcoming processors will not perform as expected once running in 64-bit mode.
When you see a company like Intel Shouting-Out-Loud about all the enhancements on their new architecture like 4-issue, macro and micro-ops fusion, memory disambiguation, etc, etc, but still DON'T say anything about 64-bit performace, then you start to think hard about it.
Even Hannibal (the guy from Arstechnica) in his extensive review of the core architecture didn't mention any "enhancements" or "additions" to intel's EMT64.

Conroe might perform good on plain old 32-bit apps, but once you put in 64-bit on the table things will start to look nasty for intel.

I beg to differ 3 64bit integer units; each unit can do one 64-bit integer operation per cycle. That is a significant improvement over Netbursts 16bit double pumped integer units.

Frankly I could care less about 64bit at this point in time but with the changes to the reorder buffer (96 from 40), an additional integer unit, and additional float point unit. I personally think the performance will be there, how much is yet to be determined but if it’s even close to linear from the performance enhancement over Netburst, I see it at least being something to talk about.