• Welcome! The TrekBBS is the number one place to chat about Star Trek with like-minded fans.
    If you are not already a member then please register an account and join in the discussion!

PC Memory question

ThunderAeroI

ThunderAeroI

Rear Admiral
Rear Admiral
I have a gigibite G-M55SLI-S4 v2, driven by an AMD Athlon 64XP 4200+. I currently have 2g of memory using DDR2 800 - 6300. I went down to wally-world and picked up a stick of 2g DDR2 667 5300 and put it into my machine.

The mobo has four slots, two different colors based upon if your using Duel channel or not. I am currently using all four slots.

When I start the machine up it says there is 4g of memory but windows is only saying there is 3.5g avilabe int the system control panel application.

My question is this; is windows stealing my memory from me for some purpose, or is the lack of 500mb because of the speed difference, or whats causing the discrepancy between the BIOS read and the windows read.

Please dont tell me anything abot going to wall-mart, or mixing memory, tats not the question and not constructive.
 
Are you using a 32bit version of Windows? - If memory serves,it probably won't recognize 4gig, especially if you have a decent amount of memory on your video card too:(
 
4GB is the maximum amount of RAM that you can address in 32 bits, and the reason you see only about 3.5GB in XP is because part of those addresses are reserved for system functions like protected mode execution. This is what keeps Application A from clobbering the memory address of Application B, resulting in the all too common General Protection Fault (GPF) errors from the 16-bit Windows days.

This has been a limitation of ALL 32-bit processors and operating systems.

Vista x64 would be a solution - but so would a hammer, and I consider that about as valid. ;)
 
TEH BABA said:
Whatever you do don't insert your soda on the cddrive.
Click to expand...
Darn you, Coca Cola!!!

(About 10 or 12 years ago, there was a little program sent around by email that claimed to be a Christmas present from Coke - not sure if they were really involved or not. When you executed it, it wished you a Merry Christmas from Coca Cola, said they hoped you enjoyed your free cup holder, and opened your CD-ROM drive drawer.)
 
I upgraded to XPx64 when I realized I couldn't use the 8GB of RAM I had put in my machine.

You could get a copy of Windows7 RC in 64 bit. It'll work for about a year before you would have to buy it. I've been using the beta up till now. Works pretty well and gotten a good rep so far.
 
Brent said:
nothing, it can't address it, have to get a 64bit OS
Click to expand...

Yes, there's an overly complicated reason to do with video memory why 32-bit Windows caps it at just over 3gb, but you will never be able to use more than 4gb with a 32-bit OS.

Memory addresses are stored as unsigned integers. The reason why your version of Windows is described as a 32-bit OS is because it uses 32-bit integers for those addresses. A 64-bit OS uses 64-bit integers.

This means that in a 32-bit OS the number of physical memory addresses you can possibly store is 2^32. Similarly, in a 64-bit OS it is 2^64. This is a basic design limitation of modern computers that cannot be changed.
 
Interesting, When it comes time for q format, i'd consider the 64 version, although by then there will probably be 128bit systems to content with.

My motherboard can support 16gig of memory so if I ever put in another 2g stick it will be a good time to upgrade.
 
ThunderAeroI said:
Interesting, When it comes time for q format, i'd consider the 64 version, although by then there will probably be 128bit systems to content with.

My motherboard can support 16gig of memory so if I ever put in another 2g stick it will be a good time to upgrade.
Click to expand...

We don't really have a pressing need for 128-bit architectures right now. The x86 line has been 32-bit since 1985 (80386). We're just now seeing wider adoption of 64-bit on the desktop. Windows 7 wasn't even supposed to be released for 32-bit systems, but Microsoft is doing it anyway because the market penetration for 64-bit is just not high enough.

Since the advantages of 64-bit over 32-bit are exponential rather than incremental, it should be quite a long time (a good 20 years?) before there's any kind of serious push for 128-bit on the desktop. Various kinds of CPUs do have 128-bit registers (and even bigger ones) but the general duties of a mainboard CPU don't change much over time.

Apparently current 64-bit CPUs don't even take advantage of the full 64 bits of memory addressing--the current generation can address 40 or 48 bits, which is up to 256TiB. Man, that'd be one sweet machine...

So, I hope there's nobody waiting around for the 64-bit generation to be replaced by 128. It's probably gonna take a while. :lol:
 
ThunderAeroI said:
Interesting, When it comes time for q format, i'd consider the 64 version, although by then there will probably be 128bit systems to content with.

My motherboard can support 16gig of memory so if I ever put in another 2g stick it will be a good time to upgrade.
Click to expand...

128 bit address buses are not gonna happen any time soon. Maybe never. 64 bit is itself overkill. If 128-bit addressing ever happens, there will definitely be no need for 256-bit. The universe doesn't have enough matter to build a memory chip that big!

128 bit data buses will happen, however, probably before 2020.

Don't confuse these two concepts. Every computer has these two ?-bit concepts. Even in the "8-bit" days, those machines still did 16-bit addressing, even though they were 8-bit processors.
 
Last edited:
wikipedia will tell you it.

http://en.wikipedia.org/wiki/Address_bus
http://en.wikipedia.org/wiki/Data_bus

But it's simple really. Imagine it like this:

The address bus governs the number of houses you can deliver a parcel to. Increasing this by 1 bit doubles the number of houses. So it does ramp up very quickly. 256 bits is enough to refer uniquely to every atom in the universe.

To put this figure into perspective:
  • The Z80, popular in 1980, used 15 bits to address memory.
  • The 68000, popular in 1990, used 23 bits to address memory.
  • The x86:Pentium 3, popular in 2000 used 32 bits to address memory.
  • A newish x64 system, released in 2010 will support upto 64 bit address, but the hardware is probably layed out only to 40 or 48 bit to keep things sensible.

The data bus governs the size of the parcel you can deliver. This is what we're referring to when we say "64-bit processor." Doubling this value, doubles the amount of data processed in each clock tick. So it's not a vast improvement considering it takes 10 years to double the bandwidth of the hardware. Increasing cpu clock speed has generally happened a lot quicker. But that's really slowed to a crawl now. We're having to rely on parallel processing to see improvement now, whether that's the width of the data bus, or multiple cores.
 
Last edited:
What is causing the growth in processor speed to a crawl? Is it just heat or is something else blocking that progress?
 
In a nutshell, yes it is heat.

There is the light speed barrier too, as it takes time for a signal to travel across the chip. And when we're dealing with GHz, light speed is significant.

To put that into perspective, 3GHz means that with each clock tick, light can travel a maximum of 10cm distance. Electricity is slower than light. So given that your computer's ram is about 10cm from the processor, you know it can't load data from main memory in one clock tick. Ye canna change the laws o' physics.

So miniaturization is necessary to make progress. That means finer conductors, which generate more heat for their size.

Those fine silicon semiconductors will burn out if they're pulsed with current too frequently. This is why core voltages keep dropping to try and counteract heating, 5v --> 3.3v --> 1.3v. There's a limit how low that can go though.

Superconducting semiconductors will revitalise the clock speed race if we ever get them manageable on the domestic scale.
 
Last edited:
ThunderAeroI said:
What is causing the growth in processor speed to a crawl? Is it just heat or is something else blocking that progress?
Click to expand...
I'm not sure what you mean by slowing to a crawl.

Basically, a processor pipeline (the queue through which processes are executed) can be though of as an assembly line. If you need to assemble 40 cars in an hour you can either have one line that can spit out one car in 1.5 minutes or 40 lines that can put together one car per hour. You could also use eight lines that can assemble five cars per hour, or any variation between.

Problem with the Pentium 4 days is that Intel realized that marketing power of having a high CPU clock speed, which essentially means Intel was going for having a few pipelines that run at fast speeds. This, as others have pointed out, creates a great deal of heat.

Current Intel chips no longer rely on the clock speed as a selling factor because Intel "came to its senses" and realized by having more pipelines and a slower clock speed they can avoid having processors that can be used to fry bacon.

There are of course other factors, such as larger cache, dynamic execution, and wider instruction sets that make procs more efficient, thus removing the need to have high proc speeds.

here is an article that explains this all far better than I.
 
You must log in or register to reply here.
If you are not already a member then please register an account and join in the discussion!

Sign up / Register


Back
Top