The purpose of this guide is to give the reader a general understanding of what RAM is, what types of RAM exist, and how to decide what RAM would make the best purchase choice.

What is RAM?

The acronym RAM stands for Random Access Memory. RAM is where data is stored within your PC so that said data can be accessed by your PC’s processor, or CPU. RAM should be considered as temporary memory within your PC, in order for RAM to maintain its data storage it must receive a continual pulse of electricity. Any data that exists solely within a PC’s RAM will be deleted when the PC is shut off or loses power.

PC’s utilize several common technologies to store data internally on a more permanent level than RAM. These storage mediums include ROM - Read Only Memory, Hard Drives – which store data magnetically in a permanent – or depending on your usage – semi-permanent fashion, CDRWs, DVDRWs, Floppy Disks and Tapes. All of these data storage methods can be considered more reliable than RAM data storage, however none of them offer the speed of data access that RAM does.

RAM stores data that is essential for a PCs immediate operation and does so in an extremely fast and reliable way. RAM allows for a PC to operate at speeds that would be unachievable if it were replaced by more permanent data storage mediums. Think of RAM as a “fast lane” on an expressway. RAM allows for a function or application’s most important data to have the fastest possible access to the processor. RAM exists as a “fast lane” for data that needs immediate attention by your PCs CPU. As such, the more RAM your PC has accesses to, the more “fast lanes” that are available for your PC’s functions and applications to access your PC’s CPU, the faster your PC will perform.

Types of RAM

SDRAM, DDR-SDRAM and RAMBUS are the three major types of RAM, or PC memory, in use today.

Modern RAM

Before the introduction of SDRAM, PC memory operated asynchronously from a PC’s Clock Speed. This asynchronous operation would create data bottlenecks within a PC itself and slow overall performance. Clock speed is the speed at which a microprocessor executes instructions; every computer contains an internal clock that regulates the rate at which data is processed and synchronizes a PC’s individual components. As it stands to reason, the faster a PC’s clock speed, the faster a PC’s CPU can process data. Before RAM was synchronized to operate at the same speed as the rest of a PC’s components, a PC’s CPU would be would be forced into the occasional delay while waiting for the RAM to be available to accept data. In theory, as long as SDRAM operated at the same speed as the system clock, it would be available to the system on a regular and consistent basis – thus eliminating data bottlenecks. By regulating RAM and tying its performance to the system clock, memory manufacturers have been forced to increase memory performance to match PC clock speeds.

SDRAM:

SDRAM – Synchronous Dynamic Random Access Memory - was the natural result of the PC industries search for better RAM and PC performance.

SDRAM is available 66, 100 and 133MHz speeds, known as PC66, PC100 and PC133 respectively. 66MHz RAM, in theory, would perform 66.6 million cycles per second and would be compatible with a 66MHz clock speed. Generally memory is considered to be backward compatible, so a pc with a clock speed of 100MHz could accept a memory module with a clock speed of 100MHz. The drawback being that the memory will only operate at the 100MHz speed.

SDRAM is no longer the memory of choice for modern PC’s, however, due to the many PCs still on the market that utilize SDRAM, it is certain that memory manufacturers will continue to produce this memory for some time. SDRAM has been in the marketplace for some time and as such is widely available for purchase as a used yet guaranteed product. This used availability offers the buyer the opportunity to save a great deal of money on its purchase, yet sacrifice a bare minimum with regard to reliability as RAM has no moving parts and is generally very durable and long lasting.

DDR–SDRAM:

DDR and DDR2 – Double Data Rate SDRAM – was again a natural result of the PC Industries search for better RAM and PC Performance. DDR memory has been available since the late 1990’s and is a great leap forward in RAM performance. Essentially, DDR RAM achieves its improved performance by transferring data to the processor twice, instead of once in the case of SDRAM, per clock cycle. Theoretically, a RAM module that refreshes the processor twice per clock cycle should equate to twice the performance offered by SDRAM. In reality DDR does not truly offer twice the performance of SDRAM, however it is a signification improvement over the older standard.

Types of DDR and DDR2 memory that are available include PC1600 – 200MHz, PC2100 – 266MHz, PC2700 – 333MHz, PC3200 – 400MHz, PC3500 – 433MHz, PC3700 – 466MHz, PC4000 – 500MHz, PC4200 – 533MHz and PC4400 – 566MHz. The first number represents the maximum memory bandwidth, in megabytes, that a RAM module can provide per second. The second number, MHz, is the clock speed that the module is compatible with. As with SDRAM, the memory is backward compatible, a PC2100 chip with a 266MHz clock speed with work with a PC with a clock speed of both 266MHz and 200MHz.

DDR and DDR2 memory are the current standard in the PC industry and will continue to be manufactured for some time. As with SDRAM, DDR and DDR2 types have been in the market for some time and are available as used or refurbished. Both used and refurbished DDR memory can offer substantial savings when purchased while giving the user similar reliability to new product.

RAMBUS:

RAMBUS – RDRAM – was developed by the RAMBUS Corporation and can be considered to be a proprietary version of RAM as only the RAMBUS company manufacturers it. RAMBUS is a high-performance version of RAM generally found in high-end business class PCs. Today very few manufacturers use the RAMBUS standard as DDR and DDR2 memory offers similar, and in some cases better performance. RAMBUS memory can be found in speeds of PC800, PC1066 and PC1200. Generally you can purchase RAMBUS PC800-() the () will contain a number that references the chips speed in nano-seconds, i.e. PC800-45.

Memory and Performance:

While adding more memory does not ensure faster performance, not having enough memory will guarantee slow downs. Having plenty of memory installed on your PC helps to ensure that your PC operates at its peak speeds and efficiency. Adding memory almost always results in a performance boost, especially if you run larger applications or multiple applications simultaneously. It is important to note that if you triple your current installed memory you will not see a three-fold boost in performance. You will almost always see some gain in performance, but you will go a long way in eliminating slow downs.

It will always be my contention and a good rule of thumb that you can never have too much memory. Maxing out you configuration with memory will help to guarantee the peak performance of your PC.

How to Choose Memory:

Choosing memory is dependent on several factors. First is the consideration of compatibility. Whether you are building a new PC or simply adding memory to an existing system, it is of key importance that the memory you purchase is compatible with your motherboard. Most motherboards accept a specific standard of memory, SDRAM, DDR, DDR2 or RAMBUS. To determine what type of memory that your PC will accept, you can either consult your motherboards owners’ manual, or, if this is not available, inspecting the motherboard for brand and model number will allow you to go online and determine the compatible memory type.

Generally a motherboard that accepts SDRAM will accept memory modules that have a higher rated speed in MHz than what is specified by the board’s manufacturer. For example, if your current motherboard has a clock speed of 66MHz and accepts PC66 SDRAM, you can install either PC100 or PC133 RAM chips. The board will only utilize the memory at its maximum speed however, so a PC133 SDRAM module will only operate at 66MHz speed. It is important to determine what RAM speeds – MHz – that your board is compatible with before making a purchase.

You should also consult either you owners manual or online documentation to determine the maximum memory that your board will support and physically inspect the board to determine how many available memory slots there are. Generally you will want to use the largest and fastest RAM chips that your board will support and fill all available slots with matching memory speeds. For example, your board has three available slots and currently one is in use by a DDR PC2100 266MHz RAM module. You discover that your board will accept DDR RAM up to PC2700 333MHz speeds. If you were to fill the remaining to slots with PC2700 memory, your memory would only function at the speed of the slowest RAM module, in this case 266 MHz.

New vs. Used:

Due to the sheer amount of memory manufactured within the last years, you will find an abundance of used memory for sale. When it comes to getting the most for your money, it should be noted that purchasing used memory is a great way to save money while getting similar reliability and performance as compared to new. RAM has no moving parts to speak of and as such is highly durable and reliable.

Spend some time researching prices between new and used memory modules. If you can purchase used or refurbished modules from vendors who will offer warranties, you may be making a major mistake spending that extra money on new RAM.

A little research on your part can help you to not only ensure that you see the maximum performance out of your PC set up, but also go a long way toward saving you money both in the short and long term.

Copyright 2006 www.hcditrading.com, Brad Calli

Article Source: http://EzineArticles.com/?expert=Brad_Calli
http://EzineArticles.com/?A-Guide-to-PC-Memory&id=310134

After careful consideration, it has become apparent to me that putting myself in jeopardy of commenting on active litigation that (potentially) involves my employer isn't a smart thing to do. That being said, I've decided to leave this content here, but with the following disclaimer:

The opinions expressed here are my personal opinions. Content published here is not read or approved in advance by EMC and does not necessarily reflect the views and opinions of EMC. The information I've presented below is through personal research through publically available news sources (news.google.com and other media outlets) and does not represent anything but a high level overview of the potential consequences of this lawsuit. 

-

As I was holding my 11 day old daughter last night (about 3am or so), I kept going back to this whole lawsuit issue between Seagate and STEC. There were a few things that were bugging me about the nature of this suit and, I thought I'd pose them here.

a.) Why STEC? In my mind, STEC represents the first successful ENTERPRISE foray for SSD drives in the storage market. BiTMICRO and mTron have done an excellent job of being the performance leaders for consumer drives, but, never quite reached that pinnacle of performance and reliability that is needed in the enterprise storage space. As such, STEC represents the single greatest danger to Seagate and their ability to continue to push FC disk as the performance leader in enterprise (SAS is another great challenger to that concept, but we'll leave that for another discussion). Additionally, as pricing parity is reached between SSDs and magnetic disk, there becomes an even greater divide between price and performance.

The other aspect of "why STEC" has to do with the who else is in the market. Why didn't Seagate sue Samsung, Intel, BiTMICRO, mTron, Crucial, et al. in addition? Each of these companies have SSD devices (either of their own design or OEM'd from others) that they're pushing into the market...

b.) Why sue vs. purchase? Interesting concept, at least to me. Knowing that they were somewhat behind the times on this new technology, why didn't Seagate investigate the potential benefits of purchasing STEC? It makes sense, really. Companies routinely purchase IP in order to gain advantage within the market. With Western Digital buying Komag (and locking up platter manufacturing), why not go down the route of buying your greatest threat? All in all, Seagate easily could have read the writing on the wall and seen where SSDs were going (especially after the EMC Symmetrix announcement!!!!). If Seagate truly wants to protect their shareholders, etc. it would make sense to get into a market segment that promises to be upwards of 8 billion dollars within the next year.

These are my initial thoughts for this morning. Let me know if you have any other ideas.

EDIT: 4/15/08 @ 12:21pm EST

c.) Is Seagate turning into the next Rambus? Sitting in a meeting this morning, I was again trying to review some of the peculiarities of this case (which have been somewhat validated by the statement issued by STEC below). Seagate's main contention is that STEC violated the following four baseline patents within 3 discrete categories: error correction, memory-backup systems, and storage interfaces with computers. If we follow this particular logic, then, we must assume the following:

• Seagate developed, controls, and receives royalty payments for: SAS, SATA, Fibre, SCSI interfaces within a given open system AND the signaling technology. (i.e. both physical and electrical connectivity)
• Error-correcting algorithms are EXCLUSIVE to Seagate and as such, SMART, sector remapping, etc. are exclusive domains of Seagate's IP.
• BBUs (battery back up) devices specific to cache within a storage system are proprietary to Seagate and thus subject to oversite and/or royalties, etc.

What's not really clear here (and mind you, I don't have access to the patent #'s in questions and their technological backend), is the role of each of the governance boards in this (Fibre Channel Industry Association, SCSI Trade Association). What I find very fishy is that Seagate, being a part of each of these groups, would be allowed to patent something that is an open format (that I am aware) and a trade standard (i.e. Fibre connectivity). If you recall, Rambus did the EXACT same thing by sitting on the DRAM design boards and then backend patenting the technology that was developed. Not saying that Seagate is ultimately a mini-Rambus, but the sheer ferocity in which Seagate seems to be going after STEC is quite suprising.

oh, and in case you missed it, STEC released their counter-statement this morning:

STEC is one of the first companies to build SSDs, having designed, manufactured and shipped SSDs as early as 1994, long before any of the suggested patents were issued to Seagate. Given the effect SSDs are having on the HDD market, STEC believes that Seagate's lawsuit is completely without merit and primarily motivated by competitive concerns rather than a desire to protect its intellectual property. STEC believes that Seagate's action is a desperate move to disrupt how aggressively customers are embracing STEC's Zeus-IOPS technology and changing the balance of power in enterprise storage. Seagate is sending a clear signal that it recognizes STEC as the leader in the SSD business and is attempting to slow down part of the growth that STEC is gaining through its SSD offering, particularly in the enterprise segment. STEC will aggressively pursue its defense to this infringement action.

In addition, STEC will also closely examine the patents asserted by Seagate as STEC believes it held such technology including prior patents, dating more than a decade prior to any of Seagate's patents. Although STEC is in the process of analyzing the claims in this lawsuit, STEC believes that Seagate's asserted patents pertain to technologies where STEC has years of prior experience and/or patents. STEC has significant patents related to SSD which have been developed through the decades of experience STEC has with developing, manufacturing and shipping SSDs. Beyond that long history, STEC also believes that many of Seagate's claims are not relevant to SSD. For example, STEC was one of the originators of stacking technology with patents dating back to the mid-1990s, while Seagate's patent on this matter was issued in 2005.

Through this process, STEC will determine if Seagate is misappropriating any of STEC's core technologies; STEC will take appropriate action to protect its interests, including seeking the invalidation of Seagate's patents.

(emphasis mine)

cheers,

Dave

W tej chwili Elpida udostępniła układy o pojemności 512 megabajtów. Wyprodukowano je w technologii 70 nanometrów.

Kości wykorzystują kilka technologii Rambusa, takich jak DRSL (DIfferential Rambus Signaling Level), która minimalizuje zakłócenia sygnału, czy ODR (Octal Data Rate), umożliwiająca przesyłanie 8 bitów danych w jednym cyklu zegara. Dzięki temu układ taktowany zegarem o częstotliwości 600 MHz pracuje z częstotliwością 4,8 GHz.

Wydajność najnowszych kości Elpidy jest sześciokrotnie wyższa niż standardowych układów DDR2-800.