Flag This Hub

Chosing Memory

By


Memory modules can be divided into three categories: modules "generic" unbranded modules, well-known brands like Kingston, Micron, Corsair, Samsung, etc.. modules and "premium" offering access times lower support clock frequencies higher and / or have oversized sinks, which facilitate overclocking.

In general, the difference in price between the generic modules and modules of brands is very small, less than 10% in most cases, so they wind up worth more to avoid the generic penalty.

Not that they are necessarily bad. In many cases, are used the same chips used in modules known brands, often with the same quality manufacturing and finishing. The market for the manufacture of memory modules is incredibly crowded, so that manufacturers that produce inferior products, or have defects margins above the average, end up not surviving very long.

The main point is that buying generic building blocks the possibility of giving the chance to buy a bad module is larger, since we are talking about working with manufacturers that profit margins are incredibly tight and pitted by any penny less in price. Known manufacturers work with prices a little above average and therefore can afford to have better systems of quality control and offer more guarantees.

Modules premium, in turn, are rarely a good option because the price difference is too large, reaching 100% or more for generic modules of same capacity and the small performance gain does not outweigh the investment.

Modules are composed of premium chips chosen during the initial phase of testing before they are welded to the modules. As with processors, some memory chips are capable of operating at higher frequencies than others. This is due to small differences in the focus of the lenses in lithography. Combining these modules with handpicked quality control more rigorous and more efficient sinks make these modules really stand out.

The big problem is that manufacturers take advantage of the hype and the limited volume of production to be sold at much higher prices. Buy premium modules is like buying designer clothes. The quality is usually a little higher, but the big difference in price does not pay.

OCZ Reaper, an example of module "premium"
See all 2 photos
OCZ Reaper, an example of module "premium"

You may feel better with a pair of modules Corsair Dominator or OCZ Reaper inside the case and perhaps even to show them to friends, but if the idea is to improve the good old cost-effective, then you get more from your money by investing in a better 3D card, processor a little faster, or a motherboard with more features.

Then there is the frequency and times of access. In cases where the seller is well informed, he can give you the information, otherwise you end up being forced to search for identification code modeled on the module or code identifying the manufacturer. In Kingston modules, for example, you is a code like "KVR533D2N4/512" on the label identification module:

A quick search on http://kingston.com shows the meaning of the code:
A quick search on http://kingston.com shows the meaning of the code:

The symbol "KVR" indicates the product line which is part of the module, in this case "Value", the line of low cost modules of Kingston. Then we have the frequency of operation supported by the module. For module of the picture, we have a "533", which indicates that it is a module DDR2-533.

After the "D2N", which indicates that it is a DDR2 module, without ECC, we have the number that indicates the latency time, that the module of the photo is 4 times. The letter "K" (if present) indicates that the module is part of a kit with two or four parts, intended for use in a dual-channel card, while the last digits indicate the capacity of the modules, as in "512" , 1G, 2G or 4G.

When buying, you should naturally give preference to modules with operating frequency and higher latency times lower, provided that this does not lead to a significant increase in price. Unfortunately, the low cost modules both go together, so that the modules with higher frequencies are often also latency times higher, since it is easier to increase the operating frequency of the module to reduce the time " real "access, represented by the latency time.

From the DDR2 modules, the access time has a more direct impact on the performance of the operating frequency. This happens not only because access times lower allow the processor to start the readings in less clock cycles, but also why they have a direct influence on the effective transfer rate for the module.

In a module with 5 CAS and RAS to CAS delay 5, for example, an access to address any of the module would take a total of 10 cycles. After this initial cycle, the controller can perform a burst of over 7 readings (8 in total), where each one takes only half a cycle. At the end of the 8 hits would take 13.5 cycles, 10 of which are related to the initial access.In a module with CAS and RAS to CAS delay of 4 days, starting the cycle would take only 8 cycles, thus enabling the 8 readings in only 11.5 cycles.


Comments

No comments yet.

Submit a Comment
Members and Guests

Sign in or sign up and post using a hubpages account.



    Like this Hub?
    Please wait working