Please register or login. There are 0 registered and 1709 anonymous users currently online. Current bandwidth usage: 326.30 kbit/s November 30 - 07:51pm EST 
Hardware Analysis
Forums Product Prices

  Latest Topics 

More >>


  You Are Here: 
/ Forums / Rambus PC-1066 and PC-1200, Pushing the Envelope

  Observations on Rambus (in retrospect).... 
 Date Written 
Robert Kropiewnicki Nov 15, 2001, 04:02pm EST Report Abuse
I wanted to keep this particular discussion separate from the actual discussion about the technology so as to keep opinions about Rambus (the company) separate from Rambus (the technology). As Sander mentioned, this was a distinction too many in the past had either been unable or simply unwilling to make.

Sander, you are right about the fact that there were numerous people who unfairly criticized the technology because of a dislike of the company, however I think you glossed over a number of issues.

The MTH of the i820 in many respects wasn't the biggest issue people had with using Rambus memory as it was the final straw where the i820 was concerned.

The first big screw up was the recall of the i820 (Camino) chipset the day before it was originally supposed to be released because of instabilities when all three RIMM slots were occupied. There was a big stink made of the fact that the eventual solution was that the third slot just be disabled.

Issue two was the price of the RIMM's themselves. When i820 was first released, overpriced wasn't the word to describe Rambus memory.....rediculously priced was almost being kind. A single 128 MB RIMM of PC800 was around $1,000 retail. Heck, it was a good thing that only two slots worked because normal people couldn't afford a third anyway.

Issue three was the lack of any major real-world performance gain over the famous 440BX boards. You were correct about one of the problems being the speed of the FSB. The other was a lack of hardware pre-fetch in the P3 which is found in the P4. That same problem reared its head when DDR systems for the Athlon were first released.

Issue four was the additional latency penalty for each RIMM in the system due to the serial nature of Rambus. This issue went by the boards with the advent of dual channel memory controllers but now you had to buy RIMMs two at a time.

By the time the MTH debacle had occurred, there was more than enough bad press regarding the Rambus situation to go around that was deserved to outweigh that which was purely anti-Intel and/or anti-Rambus.



  Topic Tools 
RSS UpdatesRSS Updates

  Related Articles 

A weekly newsletter featuring an editorial and a roundup of the latest articles, news and other interesting topics.

Please enter your email address below and click Subscribe.