• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

64bit any better?

- 64-bit numbers can be dealt with without using multiple registers and multiple clock cycles
- memory addressing is extended far beyond the 32-bit limit of 4gb
- even for memory under the 4gb limit the notions of highmem and lowmem are redone so there's no stupid confusing mixups regarding high and low memory
- x86-64 has twice as many GPR's (general purpose registers) as old x86. there are now 16 instead of 8.
 
though 64 bit programs are still basically non-existant in the desktop world, and i haven't actually done testing with xp-64 on my 631, everything i see indicates that intel + 64bit os is a somewhat bad idea..

if someone could give the specifics on why, i would be appreciative :D
 
Prescott is a dog in general, that might be why. Conroe and Woodcrest should be much better in either 32 or 64 bit mode.
 
(cf)Eclipse said:
though 64 bit programs are still basically non-existant in the desktop world, and i haven't actually done testing with xp-64 on my 631, everything i see indicates that intel + 64bit os is a somewhat bad idea..

if someone could give the specifics on why, i would be appreciative :D
Click to expand...

Prescott's EMT64 was dismal.

Conroe's is significantly better. Better than AMD's 64-bit, in fact.
 
Pabster said:
Prescott's EMT64 was dismal.
Click to expand...
i know that, but why? :p

and yeah, conroe's sheer power helps it continue to best the K8 in 64bit mode, though whatever article it was, i think xbit's, shows K8's gain to 64bit to be 16% and conroe's is 10%, most likely cause i think the op-fusion doesn't work in 64bit mode
 
Pabster said:
Prescott's EMT64 was dismal.

Conroe's is significantly better. Better than AMD's 64-bit, in fact.
Click to expand...

I believe I've seen benchmarks that indicate otherwise...let me dig up the article.

EDIT: Found it...

Here's the article:
http://www.xbitlabs.com/articles/cpu/display/core2duo-64bit_5.html

Here's is the first page of benchmarks:
http://http://www.xbitlabs.com/articles/cpu/display/core2duo-64bit_5.html

Nice graph on the last page of the article that summarizes the benches:
http://www.xbitlabs.com/articles/cpu/display/core2duo-64bit_7.html

It's hard to separate the general performance of the chip due to clock speeds, and its 64 bit implementation, but the conclusion of the article basically says that Conroe does
a fine job with its 64 bit implementation. It also says that Athlon users will experience a greater increase in performance when switching to 64 bit, but that's a relative measure.
Conroe basically offsets that by delivering better performance at the same clock rate.


me (From Article) said:
Conclusion

As we have expected, nothing serious has happened. CPUs with Intel Core microarchitecture and EM64T technology work normally in 64-bit modes. No dramatic performance drop has been detected in most benchmarks.

Of course, there are a few applications, when Core 2 Duo work slower in their 64-bit versions than it would in their 32-bit ones. Among them are Windows Media Encoder 9 or 7-zip archiving tool, for instance. However, since the other testing participants have also lost some of their performance in these tasks, the problem is most likely to be not in the microarchitecture. EM64T technology of Core 2 Duo processors has a positive effect on the performance in the majority of applications.

At the same time I would like to point out that it looks like Athlon 64 processors ensure higher performance increase when switching to 64-bit work mode. The average performance improvement we have seen from Athlon 64 FX-62 equaled 16%, while Core 2 Extreme X6800 demonstrated only 10% average performance boost. This way, there is a certain difference: AMD K8 turns out 6% mode efficient in 64-bit mode than Intel Core. However, this difference cannot compensate for the 20% performance advantage of the Intel Core 2 Duo over the Athlon 64 X2 working at the same clock speed, which we have pointed out in our previous articles.
Click to expand...
 
you guys saying 64bit is more complex? why would someone use more than 4gb of ram, graphics/lan parties/games? in the time of 16bit, what did 32bit gave us? what's the future?
 
64 bit is more complex in terms of the CPU design due to an increased number of instructions, expansion of registers to 64 bit, and support for 64 bit addressing modes. x86-64 also increased the number of general purpose registers. Though the additional registers are a feature of that specific 64 bit implementation rather than being anything related to innate 64 bit-ness.

From a developers perspective the compiler can automatically utilize the extra registers so coding for a pure 64 bit environment is mostly a question of using the right data types (mostly for pointers and memory references). It's not really an increase in coding complexity (though porting poorly written 32 bit code to native 64 can be a hassle).

Having more than 4GB of RAM is useful for any scenario dealing with large data sets, and is common in detailed simulations, data mining, and for large databases or models. The larger the set of data you can hold in RAM the less disc IO is required during processing.

For games, right now, the extra address space isn't terribly useful (though it may become so eventually), although the extra registers DO give a performance benefit.

The transtion from 16 bit to 32 bit also had several facets. One was the increase in address space. A native 16 bit address space allows for 64 KB of memory. In most cases 16 bit processors in the PC space had either a 20 bit (1MB) or 24 bit (16MB) addressing capability. 32 bit systems used a 32 bit addressing model allowing for 4GB of RAM.

32 bit x86 CPUs also support hardware addressing virtualization (physical support for paged memory) and memory protection (the association of a specific block of memory pages to a process).

So with 32 bit operating systems it became much easier to set things up so that when one process failed it was physically incapable of affecting other processes on the system.

Future wise, I'd expect additional instructions to be added under the current 64 bit model (e.g. SSE4), and possible expansion of the register model, and eventually a switch to 128 bit CPUs.

Again, the native benefit of higher bit widths is just address space and the ability to store larger values in a single register. Other additions are possible without increasing the width of a CPU register or addressing capabilitiy, but typically an increase in register width and address range is going to add other things as well.
 
When 64-bit Registers have RAM expandable upto 128GB then why today's motherboards support maximum 4GB of RAM?
 
True 64 bit addressing permits a 16 EB (exa-byte, the progression being giga, tera, peta, exa etc.) flat address space, not 128GB. x86-64, whether AMD64 or EMT64, actually uses a 40 bit physical addressing model and a 48 bit virtual address space. That allows for 1 TB of physical RAM and 256 TB of virtual address space. The x86 model is provisioned to for expansion to either a 52 or a 56 bit addressing mode for later support of up to 16 PB.

There are plenty of motherboards that support more than 4GB of memory, and have been for a good while now. Intel’s D975XBX is certified for 8GB for example, various nForce 4 boards are good for 8-16 GB, and if you want to consider workstation boards there are lots that are good for 32+ GB.

The limit there tends to be a design decision or limit on module size, rather than any specific failing in the addressing models of the hosted CPU architecture. Also you won’t often see these boards offered in systems with more than 2GB of RAM, but that’s the vendor electing to avoid ill-educated purchasers complaining that their 32 bit operating system is only seeing <3GB of RAM when they purchased 4 or more.

128 bit CPUs are a long way off, though that’s the eventual progression. Each successive increase in bit width has taken longer to come, since it takes long to reach the limits of those increases. 4 bit to 8 bit transitioned in one year (1971 – 1972), for example. Sticking to PC CPUs (to keep things simple) fully 16 bit processors showed up in 1978 (the 8080 had limited 16 bit capability via register combining).

There were various refinements to the pure 16 bit model (protected memory, 24 bit addressing) over the next few years (think 80286), and it was 1985 when the first fully 32 bit PC CPU launched … that being the 80386.

Ten years after the 386 launched we got the first mainstream 32 bit PC operating system (Windows 95, 9 years if you include NT Workstation), and it was another 8 years before the first 64 bit x86 compatible CPU showed up (AMD’s Athlon and Opteron 64s in 2003).

So 18 years of 32 bit CPU availability before 64 bit PC processors showed up. During that time lots of changes occurred, both in terms of instruction set additions (on boarding of the FPU, addition of MMX, SSE, 3DNow! etc) and architecture. So that likely puts 128 bit CPUs a couple of decades out, and definitely not something to worry about.
 
OKS..............So I guess it will take another 3-4 yrs for 64-bit to fully develop? Long way road. My motherboard (intel 945GNT) says that it supports maximum 4GB or RAM and out of which only 3GB will be utilized, rest will be utilized for operating system.

The only reason I upgrade my PC from intel 865GBF with P4 2.4GHz to 945GNT with Pentium D 915* 2.8Ghz was to experience the power of 64-bit processing. But now I've realized that I am bit disappointed
 
It's hard to say when 64 bit will become maintstream. I expect Vista will accelerate the process, much as Windows 95 did for 32 bit adoption.

When AMD was first touting the wonders of 64 bit computing, Intel said "Bleh, not relevant for most users until 2008" and they were probably close to correct (I'm not claiming they were saying it for any reason other than to deflate AMD's marketing, but that doesn't mean they were wrong). So that falls within your 3-4 year guestimate, and even a little sooner.

If you want the benefit of 4GB of RAM with Windows, regardless of your motherboard, you have to run a 64 bit Windows build. Right now that means either Windows XP Pro 64, or 64 bit Vista (which isn't released yet, though RC1 is due soon).

Your Pentium D 915 is dual core of course, and that should be showing useful performance improvements over your previous P4. While the REAL benefits to that come in properly multi-threaded applications, it should have made normal use of the machine vastly smoother and more responsive.

For example, you should be able to run a long video transcode operation and still use the machine to do other demanding tasks without issue.

...

Also, even without using the additional address space available in 64 bit machines, there are instruction set/register additions present only in the 64 bit CPUs that do make a postive difference in performance regardless. It's just not a night-and-day kind of difference.
 
Thanks for the advice torque, I really appreciate that. I do have 64-bit Windows XP pro. Its pretty obvious that you need a 64-bit Operating system, otherwise its just like any other 32-bit Processing.

I have one important question. 64-bit CPUs come up with this extra security feature "Enhanced Virus Protection (EVP)". Can you tell me how can I enable protection of my applications and associated files from buffer overrun attacks?

Does EVP also means that we can discontinue the use of 3rd party anti-virus softwares as a part of their security strategy? Because in EVP the virus protection resides within the firmware only, if I am not wrong, pls suggest me
 
So, EVP is the AMD marketing name for their NX (No eXecute) implementation (Intel call theirs XD, for eXecute Disable). In Windows this technology, regardless of whether you're running an AMD or Intel CPU, is called "Data Execution Prevention" (DEP) .. and that's what I'll call it for the rest of this post.

You're absolutely correct about it being a system for preventing buffer overruns from executing instructions that exist in a data segment in memory. And that's literally all it does. In practice, while it cuts of one specific attack vector for getting malicious code running on a target machine, it is just that one attack vector that is protected.

You sound pretty clued up, so some of what follows is general info for others that are unaware of what this really means, and then I'll get to how you turn it on etc. :)

DEP is worth having for sure, but it doesn't prevent all types of buffer overruns (just the easy-to-pull-off ones), and it doesn't protect against executing malicious code that exists in a code segment rather than a data segment.

So, it's good, in general, for stopping a malformed URL or a dodgy web component from randomly executing code that might install a virus or trojan or whatever. What it won't do is stop such malicious code executing if it gets on to your box as part of another program that you manually execute.

A simple example would be if you downloaded a crack or a P2P client from a questionable source, and the installer for that also installed a virus. The virus code itself at that point could run like any executable and would reside in a code segment and thus DEP isn't involved.

DEP should stop an attack that relied on a specially formed MP3 file causing a buffer overrun in a specific media player application, to install the virus code however (depending on the precise nature of the buffer overrun of course).

DEP etc. are absolutely NOT alternatives to other safe computing practices. There are many vectors for getting malicious code to run on your PC, and DEP only addresses one of these, so you still need to run AV and Spyware protection/scanning and a decent firewall (or a NAT router).

For 64 bit Windows viruses are currently much less of a worry. Unless they're written to run as 64 bit code they're somewhat limited in the level of system access they have, and in most cases simply fail to install or run (Kernel mode code and drivers must be 64 bit).

Now, how to turn on EVP/DEP/NX/XD ... :)

This feature is available for Windows XP Pro SP2 32 bit as well as the 64 bit versions, and the setup is the same.

It's basic behavior should be enabled by default. You can get to the settings by right-clicking on "My Computer", selecting "Properties", choose the "Advanced" tab, hit the "Settings" button for "Performance", and then choose the "Data Execution Prevention" tab.

By default DEP is enabled only for specific Windows applications and services. This is to ensure that some software which "legimately" (though through bad design) executes code from data pages will still run.

You can choose the other option ("Turn on DEP for all programs and services except those I select") and that will enable DEP for everything. Note that some applications may not operate correctly in this mode, and in those cases you can add the offending applications to an exclusion list by using the "Add" button on this same configuration page.

...

It's worth noting that on 64 bit systems DEP has all the same benefits, plus some extra capabilities when dealing with Kernel-mode code (via Kernel Mode DEP). In 32 bit mode DEP applies to drivers and the Kernel's stack. In 64 bit mode it adds protection to the paged pool and session poll as well, and drivers are prevented from executing code on the stack.

Sorry that's a bit on the long side, but hopefully it helps!
 
Torq said:
True 64 bit addressing permits a 16 EB (exa-byte, the progression being giga, tera, peta, exa etc.) flat address space, not 128GB. x86-64, whether AMD64 or EMT64, actually uses a 40 bit physical addressing model and a 48 bit virtual address space. That allows for 1 TB of physical RAM and 256 TB of virtual address space. The x86 model is provisioned to for expansion to either a 52 or a 56 bit addressing mode for later support of up to 16 PB.

There are plenty of motherboards that support more than 4GB of memory, and have been for a good while now. Intel’s D975XBX is certified for 8GB for example, various nForce 4 boards are good for 8-16 GB, and if you want to consider workstation boards there are lots that are good for 32+ GB.

The limit there tends to be a design decision or limit on module size, rather than any specific failing in the addressing models of the hosted CPU architecture. Also you won’t often see these boards offered in systems with more than 2GB of RAM, but that’s the vendor electing to avoid ill-educated purchasers complaining that their 32 bit operating system is only seeing <3GB of RAM when they purchased 4 or more.

128 bit CPUs are a long way off, though that’s the eventual progression. Each successive increase in bit width has taken longer to come, since it takes long to reach the limits of those increases. 4 bit to 8 bit transitioned in one year (1971 – 1972), for example. Sticking to PC CPUs (to keep things simple) fully 16 bit processors showed up in 1978 (the 8080 had limited 16 bit capability via register combining).

There were various refinements to the pure 16 bit model (protected memory, 24 bit addressing) over the next few years (think 80286), and it was 1985 when the first fully 32 bit PC CPU launched … that being the 80386.

Ten years after the 386 launched we got the first mainstream 32 bit PC operating system (Windows 95, 9 years if you include NT Workstation), and it was another 8 years before the first 64 bit x86 compatible CPU showed up (AMD’s Athlon and Opteron 64s in 2003).

So 18 years of 32 bit CPU availability before 64 bit PC processors showed up. During that time lots of changes occurred, both in terms of instruction set additions (on boarding of the FPU, addition of MMX, SSE, 3DNow! etc) and architecture. So that likely puts 128 bit CPUs a couple of decades out, and definitely not something to worry about.
Click to expand...
Ahh, thanks for the first part explanation for memory addressing. I was wondering why the new 64 bit processors could only support up to 1tb of ram instead of the full 2^64 bits.
 
samsung.204b said:
you guys saying 64bit is more complex? why would someone use more than 4gb of ram, graphics/lan parties/games? in the time of 16bit, what did 32bit gave us? what's the future?
Click to expand...
my research thesis would likely benefit from > 4GB of RAM.

torq said:
You can choose the other option ("Turn on DEP for all programs and services except those I select") and that will enable DEP for everything. Note that some applications may not operate correctly in this mode, and in those cases you can add the offending applications to an exclusion list by using the "Add" button on this same configuration page.
Click to expand...
I am running this DEP mode on all my machines and I remember only a single incident, where I needed to exempt an application.
 
You must log in or register to reply here.
Back
Top