Computer Memory

Memory Speed and overheating

There are a number of upgrades that come with the new DDR2 technology over DDR1 memory. DDR2 memory is of course faster, and will ultimately reach clock speeds of 667MHz or higher as opposed to the top functional speed of DDR1 at 400MHz. The higher speeds of DDR 2 memory allow it to have up to 10.6GB per seconds of throughput - quite a bit more than DDR1 memory which maxed at 6.4GB per second. The feature that makes these higher speeds practical is the reduced power consumption of DDR2, which runs at only 1.8 volts. Heat is one of the real banes of IC's, and the faster you try and clock a chip the hotter it runs and the less reliable it gets. One of the speed limitations of DDR1 was that at 2.5 volts created too much heat and was simply too hot to run consistently at speeds greater than 400MHz. DDR2's 1.8 volts combined with a much smaller (FBGA) chip package goes a long way towards solving the heat issue at higher speeds.

Cas Latency issues

DDR2 is also reversing the trend of faster and faster CAS Latency. SDRAM was originally CAS 3 though CAS 2 became popular later on with PC100 and in some PC133 modules. The problem with putting the burst cycles so close together is that the faster you push the chips, the more unstable they become. This is why, after moving to CAS 2 in a lot of PC100 components, the standard for PC133 reverted to CAS 3. DDR1 came out with two different speeds CAS 2 and CAS 2.5, though there were some chips that were available with 3-3-3 timing as well. There were some problems when DDR was released getting the modules to work consistently, and when DDR 400 started becoming available there were problems again, some of them related to Latency. The industry seems to have recognized this, and the JEDEC standard for DDR 2 is CAS 3 and even CAS 4 in some cases. This will go a long way towards making the modules more stable. There will be a small decrease in number of functions a stick will be able to perform per second, but the decrease will be so minor as to be unnoticeable and will be more than made up for in dramatically increased stability.DDR1 MemoryDDR2 MemoryModule package184-pin DIMM/ 200-pin SODIMM240-pin DIMM/ 200-pin SODIMMSpeeds (in MHz)200, 266, 333, 400400, 533, 667Voltage2.5 volt1.8 voltChip packageTSOP/BGAFBGAModule BandwidthUp to 6.4GB/secondUp to 10.6GB/secondChip densitiesUp to 1GbUp to 4GbCAS Latency2/2.53/4

The memory bus

• Open in Google Docs Viewer
• Open link in new tab
• Open link in new window
• Open link in new incognito window
• Download file
• Copy link address
• Edit PDF File on PDFescape.com

There are actually two answers to this, depending on what method you want to use to determine the answer.

Standard definition The standard definition for kilobyte is 1,000 (10^3) bytes, just like there are 1,000 meters in a kilometre.

Likewise, the standard definition for gigabyte is 1,000,000,000 (10^9) bytes. Dividing, this means there are 1,000,000 (1 million) kilobytes (kB) in a gigabyte (GB).

This is the definition used for hard drives, portable memory drives (memory cards, USB drives), DVDs, Blu-ray disks, and most measures of performance. Some software (such as Mac OS X and the Linux kernel) uses this definition when displaying file and disk sizes. Memory manufacturer usage Memory manufacturers use a non-standard system that conflicts with the way everyone else measures. By their definition, a kilobyte (written "KB") is 1,024 (2^10) bytes, and a gigabyte (written "GB") is 1,073,741,824 (2^30) bytes.

The correct name for 1,024 bytes is "kilobinary byte" or "kibibyte". The correct name for 1,073,741,824 bytes is "gigabinary byte" or "gibibyte".

This is approximately equal to the above, but you may need to know it to figure out exact numbers:

By this definition, 1 "GB" = 1,048,576 "KB" More correctly, 1 GiB = 1,048,576 KiB

This definition is used by memory manufacturers and some software, like Microsoft Windows.