Please consider a donation to the Higher Intellect project. See https://preterhuman.net/donate.php or the Donate to Higher Intellect page for more info.

TechWorks PowerRAM

From Higher Intellect Vintage Wiki
Jump to navigation Jump to search
Powerram.gif
Simm2.gif

The cornerstone of TechWorks' product line is our add-on memory upgrades for today's most popular computers and laser printers. In the rapidly changing memory market, you need solutions fast. TechWorks has sold over four million memory upgrades and has the design and manufacturing expertise in-house to bring the latest technology to your desktop.

TechWorks continues to be the market leader in selling memory upgrades for the Macintosh personal computer. We have achieved this position by focusing on the customer's needs and providing memory solutions for virtually every Macintosh ever made. As a Certified Apple Developer, we are provided with the latest technical information that allows us to build products exactly to Apple specifications. TechWorks' memory products provide solutions for everything from desktop and portable Macs to Workgroup Servers and Mac OS compatible computers.

The TechWorks PC Memory product line is the fastest growing piece of our business. Over the past several years, we have expanded to cover virtually every PC on the market. Whether it's for your office or home desktop system, your department's PC server, or your faithful notebook traveling companion, TechWorks has the memory solution for you.

SDRAM Memory

In the ever changing world of personal computers, one of the few things that has remained constant in recent years is DRAM technology. As a result, memory performance has become a bottleneck in many high-end systems. Now, a new kind of memory known as Synchronous Dynamic Random Access Memory (SDRAM) is being used increasingly as an effective solution.

What is SDRAM?

SDRAM is a new technology designed to match memory functionality to the increasing speeds of high-performance microprocessors and motherboards. The technology is revolutionary because all operations that an SDRAM chip is instructed to do are synchronized to a clock signal provided by the computer. This clock is usually the same clock that the microprocessor bus uses. The timing coordination between memory, the microprocessor, and other support chips permits more efficient memory access and eliminates system wait states. That means your system performs better (up to 20% improvement over EDO DRAM for certain applications).

SDRAM has a two-bank (two separate areas in which data can be stored) architecture that allows pipelining of information through the SDRAM. Pipelined data access means when one bank is outputting data, a second bank can be charged and begin decoding the next information. SDRAM pipelines memory accesses between the two banks, which allows memory accesses to overlap between the two. Design complexity is also greatly reduced because of the pipeline.

Burst output (a retrieval of consecutive instructions) is also possible with SDRAM, and the burst output length is programmable. SDRAM burst data access rate is almost four times faster than that of standard page mode DRAM.

SDRAM technology finally provides the ability to reach high-performance bandwidths. Using SDRAM, system users can now close the performance gap between high-speed microprocessors and memory functions.

What does all that mean?

That means an SDRAM-compatible computer is an absolute state-of-the-art machine offering up to a 20% performance improvement over EDO DRAM machines.

That also means SDRAM is a technology that will be tomorrow's standard for memory, and many new user applications are being developed with SDRAM in mind. Those applications will run slower on non-SDRAM systems.

Do I need SDRAM in my system?

Because SDRAM is a new technology, SDRAM-compatible computers and workstations are just now being placed on the market. Some of those machines are engineered to operate with either EDO DRAM or SDRAM installed. Therefore, if you have an SDRAM-compatible computer or workstation that currently has EDO DRAM installed and you use your system for CAD, multimedia, 3-D, or any other application that requires intense memory performance, then an SDRAM upgrade is for you.

Are SDRAM modules easy to install?

SDRAM modules are as easy to install as conventional DIMMs. There is no single rule that covers all installations, but for the most part, installing DIMM modules is quick and easy. As with any static-sensitive device, remember that it is important to be properly grounded. Some DIMM modules are held in place by friction only, and therefore should fit firmly into the socket. Other modules are secured to the socket by clips.

EDO Memory

What is EDO Memory?

Extended Data-Output (EDO) is a DRAM technology that improves the performance of the memory subsystem while maintaining backward compatibility with previous systems.

In basic terms, EDO can be viewed simply as a very fast standard DRAM. In technical terms, the essential advantage of EDO is shorter Page Mode cycle times than standard Fast Page Mode (FPM) DRAM. In FPM DRAMs, the output buffer is turned off with the rising edge of CAS (column address strobe) and the output data is no longer available. In EDO memory, this is not the case҆the data is not turned off by the rising edge of CAS. This effectively means that the data is available for a longer period, thereby enabling the system to read output data while setting up for the next cycle. In achieving this, EDO saves one clock cycle for every page access, resulting in a significant increase in system performance.

In addition to this fundamental difference in operation, EDO has two other key features. The memory design offers both increased peak memory bandwidth and simplified constraints on access timing, which help to increase memory performance. As an example of the performance increase, while a 60ns standard FPM DRAM offers a 40ns cycle time, a 60ns EDO DRAM offers a 25ns cycle time. EDO typically results in a 10-15% overall performance increase in systems that are designed to take advantage of this technology.

Where is EDO appropriate?

While the logical application of this fast DRAM is to squeeze every bit of performance out of systems with the very latest processors, there is also an important place for it with the low-end of Pentium systems. Because these technologies narrow the gap between DRAM and SRAM performance, the use of it with a high speed processor will enable system vendors to offer cacheless systems that still deliver high-speed capability. Since the incremental cost of EDO is virtually zero, vendors will be able to move to lower price points without giving up much performance. This approach is particularly attractive for notebook systems. Gaining performance without having to incorporate external L2 cache lowers power consumption (lengthens battery life), lowers heat dissipation needs, and minimizes board real estate demands.

What does this mean for the installed base and future memory purchases?

Fortunately, since EDO memory acts very similarly to conventional FPM DRAM, the impact of the growth of this technology is minimal. EDO uses the same signals and packages as FPM DRAM. In most cases, EDO will work in systems that were designed to work only with FPM. This even includes Apple Macintosh systems that use industry standard 32-bit, 72-pin SIMMs. As would be expected, the increase in performance typically associated with EDO would not occur. Most new chipset designs will allow any combination of both technologies (FPM and EDO) to be used. Placing standard DRAM modules in a system equipped to utilize EDO will also not cause problems. The system will function properly, just without the performance boost the faster memory could have provided. If you mix EDO and non-EDO modules in a system, function will be unaffected and performance will be at standard FPM DRAM levels. It is recommended that modules using different types of DRAM not be mixed within a memory bank.

EDO modules are available in the same configurations as FPM modules and use the same board design. Because the only differences between Fast Page Mode and EDO are within the DRAM chip, it is difficult to distinguish between these two modules by sight. The only visible difference will be the DRAM part numbers.

DIMMs (Dual In-line Memory Modules)

Modules.gif

The acronym DIMM stands for Dual In-line Memory Module. The DIMM architecture is an evolution of the SIMM, or Single In-line Memory Module. In its simplest form, the DIMM can be thought of as two SIMMs built into one unit. Each of the SIMM pins (shiny teeth on the bottom edge of the SIMM) is tied to the pin on the other side of the module and therefore the two act as a single pin. On a DIMM, however, the pins on each side of the module act independently for addressing information to the memory chips. DIMMs in their various forms are gaining acceptance in numerous applications, including notebooks, the latest Apple PowerMac systems, and high-end PC systems. They will continue to grow in popularity, likely replacing both SIMMs and some proprietary notebook modules.

What are the different types and sizes of DIMMs?

For desktop systems, 168-pin DIMMs are becoming popular in a variety of configurations and sizes. For notebook computers, 72-pin, 88-pin and 144-pin small outline (SO) DIMMs are rapidly gaining ground. Examples of common DIMM types and configurations are shown in Table 1. The 168-pin DIMMs are available in non-parity (64-bit) as well as parity and Error Checking and Correcting (ECC) (72-bit) versions. Used primarily in desktop computers, these modules are 5.25" wide and are available in 1.1" and 1.25" heights. The computer manufacturer specifies which sizes and types are appropriate. As with SIMMs, the pins can be either gold or tin. For example, Apple specifies gold pins for its DIMMs.

SO DIMMs are most typically seen in non-parity configurations for use in notebook computers. The standard SO DIMM design allows for 4MB, 8MB, 16MB, 32MB, and 64 MB module capacities. The SO DIMM design is gaining popularity because of its small size, low cost, and ease of upgrade for the user. One of the key differentiating factors in SO DIMMs is the operating voltage, which is either 3.3 Volts or 5 Volts. The SO DIMM connectors and modules for the two voltages are designed with different sized notches on the side of the DIMM (see Figure2). This prevents a 3.3V module from being installed in a 5V system. When ordering SO DIMMs for your PC notebook, be sure to confirm the proper voltage modules.

Why are DIMMs being used?

Computer manufacturers have adopted DIMMs for two main reasons: 1) to allow a wider data path with a single module, and 2) to place more memory on a smaller physical width module. The 64-bit data path of the 144-pin or168-pin DIMM is twice as wide as the 32-bit data path of a 72-pin SIMM or DIMM, meaning fewer modules are needed for a typical upgrade. For instance, a machine that requires adding SIMMs in pairs may require DIMMs to be added only one module at a time. In addition, DIMMs support higher capacity modules, increasing the maximum capacity to 512MB for a 168-pin DIMM compared to 32MB for a JEDEC standard 72-pin SIMM. The size benefit is best illustrated by the example that a 72-pin DIMM is almost half as long as a 72-pin SIMM.

Like the currently popular SIMMs, the DIMM standard is flexible enough to accommodate a variety of DRAM designs, including Fast Page Mode (FPM), Extended Data Out (EDO), Burst EDO, and Synchronous DRAM (SDRAM), as well as multiple configurations including non-parity, parity, and Error Checking and Correcting (ECC).

Are DIMMs easy to install?

There is no single rule that covers all installations, but for the most part DIMMs are very easy to install. As with any static-sensitive device, remember that it is important to be properly grounded. Unlike SIMMs, 168-pin DIMMs are held in place mainly by friction and therefore should fit firmly into the socket. For SO DIMMs, the installation procedure is almost identical to that of a standard SIMM, except that the final orientation of the module is typically flat against the system board, instead of perpendicular to it.

Level 2 Cache Memory

Diagram2.gif

Why do I need TechWorks cache?

A cache (pronounced "cash") is a special type of memory utilizing ultra-fast static RAM (SRAM) technology, which dramatically enhances your system's performance. System memories composed of dynamic RAM (DRAM) alone have not been able to keep up with the dramatic increases in CPU speeds over the years. In order to optimize the memory performance in these systems, designers are implementing architectures using cache memory, resulting in speed increases of 5% to 45%. Cache isn't actually a new idea, as mini and mainframe computers have been using this technology for years. TechWorks offers many different types and sizes of cache cards which are compatible with a wide variety of systems.

How does it work?

The cache card acts as an intermediary between a computer's CPU and its ordinary DRAM. As you run your applications, a copy of frequently used data is stored in the cache. When the CPU needs to read data, it first looks in the cache. If it finds the data there, it doesn't need to look in the computer's DRAM. Because the cache is built with high-speed SRAM, the CPU accesses the cache faster and more efficiently than it accesses the computer's DRAM. Moreover, the cache has a direct line to the CPU, unlike devices installed in a bus architecture like PCI, NuBus, ISA, or SCSI.

What are the different types and sizes?

There are two different types of memory caches: Level 1 and Level 2. A Level 1 cache is smaller (typically 256 bytes to 32K bytes) and is integrated into the microprocessor. All modern day processors from companies such as Intel, Motorola, and AMD come with built-in Level 1 cache. A Level 2 cache is separate and is typically installed on the motherboard or in an expansion slot external to the processor. This cache typically ranges in size from 128K to 1MB. The Level 2 cache works in conjunction with the microprocessor's internal cache to provide maximum performance.

How much cache is right for me?

Obviously, with more cache the CPU has a better chance of executing instructions sooner. The maximum amount you can have depends on your particular system. Most recent Apple Macintosh and Intel Pentium systems and many 486 machines are designed for cache cards. The new Apple Macintosh computers utilizing the PowerPC chip, in particular, realize significant performance benefits from adding cache.

How you use your computer will decide how much cache is needed. Anyone can benefit from adding cache, but the question is how much do you need to obtain maximum performance without going overboard? The following guidelines should help you decide:

If you are an average user with applications including word processing, spreadsheets, or games, a small to medium sized cache is probably fine for you. If you regularly use high-end applications such as desktop publishing, video editing, and sophisticated databases then you will probably want to get as much cache as your system will handle.

See Also