We store cookies, you can get more info from our privacy policy.

GameCube FAQ


by the NWR Staff - March 4, 2001, 1:25 am EST




What’s MoSys about? What is the company’s History?

Since MoSys’ foundation in 1991, it has worked to establish itself as a premier supplier of leading-edge memory products and building blocks for communication, workstation, desktop and mobile markets.

MoSys currently has been granted 20 patents to protect its ever-growing roster of new technologies. Of these, it is the development of revolutionary DRAM architecture, called "Multibank.” The “Multibank” architecture greatly improves performance while simultaneously lowers the cost of high-performance platforms. Building and selling DRAM and SRAM products based on the “Mulitbank” architecture, the Sunnvale-based MoSys is truly redefining the standards in regards to memory cost and performance.

MoSys achievements haven’t gone unnoticed. In September 1999, MoSys announced its partnership with Nintendo. On January 31, 2000, MoSys & fellow-Nintendo ally NEC announced another alliance. It seems NEC’s 1st LSI Memory Division will use MoSys’ 1T-SRAM technology in high-density application-specific memories, including Nintendo GameCube's.

What is MoSys’ Involvement with Nintendo and the GameCube?

MoSys’ unique 1T-SRAM technology will hook up Art-X’s “Flipper” chip to IBM’s Gekko CPU and enable them to work together at incredible speed.

“The incredible performance of MoSys’ 1T-SRAM Memory, with the proprietary custom graphics chip designed by ArtX is the perfect match for IBM’s custom, copper-based CPU,” said Howard Lincoln.

What is this “1T-SRAM” exactly?

Let’s break it down… 1T-SRAM is MoSys’ unique Static RAM, hence the “S” (unlike dynamic, or DRAM,) based off of their “Multibank” architecture. The static nature of the SRAM will give the Gekko chip instant access to all necessary information, giving the GameCube its incredible speed.

The "1T" stands for the technology’s single transistor cell, which is used to save space and decrease cost. The “S” stands for static. This describes the refresh-free interface and the resulting low latency time typical of traditional six-transistor SRAM cells. 1T-SRAM is available in densities up to 128Mbits.Unlike embedded DRAM, 1T-SRAM uses simpler circuitry. This allows costs to be kept down.

Embedded 1T-SRAM, as used on the GameCube console’s graphics chip (“Flipper”), enables designers to further decrease latency times. The designers of the GameCube feel that latency is one of the biggest problems game designers face. 1T-SRAM was also used because embedding six-transistor SRAM or DRAM causes other problems as well. For example, the increased chip complexity results in higher production costs. 1T-SRAM can be fabricated in either pure logic or embedded memory processes using as little as one ninth of the area of traditional six-transistor SRAM cores. In addition to the high performance and density, this technology offers dramatic power consumption savings by using under a quarter of the power of traditional SRAM memories.

In laymen’s terms, 1T-SRAM is fast. Imagine you’re looking for the ace of spades in a deck of cards. If you’re using SRAM, the cards are all visible on a table and you can immediately grab the ace of spades. If you’re using dynamic RAM, you simply flip through the deck of cards one at a time until you get it. As you can see, SRAM is much faster. In addition to this, 1T-SRAM only has one transistor for every six that traditional SRAM has. You end up with the speed of SRAM with the low complexity and size of DRAM; the best of both worlds.

Got a news tip? Send it in!