Rambus DRAM (RDRAM)
A reader responds to comments about RDRAM in the Northwood P4 article 

Posted: 03nov2001
Updated: 08nov2001

Page 1 of 1



This page contains a reader's response to comments I made about Rambus RAM 
in the Intel Northwood Pentium 4 CPU article. John S. is an engineer who does 
system-level, rather than IC design. He lives in Florida, and has no connection 
with Rambus, other than owning their stock. 

His reason for responding was simple frustration. He says, "I hate seeing conniving 
organizations roll over the little guy and marketing inferior versions of their ideas 
just to avoid paying their due."

Admittedly, I haven't researched Rambus RAM very much. All I know is that it 
provides performance comparable to that of DDR RAM, yet costs ~4X more per 
MB. This reader makes compelling points, and includes a brief history of RAM 
development:

                                                  ***

"The problem with Rambus DRAM is that it's significantly more expen$ive 
than DDR  RAM, without providing a correspond increase in performance. 
It's rumored that Intel cut a deal with Rambus, which awarded Intel mega-
buck$ if they helped Rambus sell a crapload their RAM." 

Rambus RAM is more expensive than DDR RAM, but prices are dropping almost 
daily. Rambus is to SDRAM (including DDR) what SDRAM was to DRAM (including 
EDO RAM). A little background info:

When asynchronous Dynamic Random Access Memory (DRAM) became too 
slow for growing processor speeds, Synchronous DRAM (SDRAM) was created. 

DRAM makers used various tricks to squeeze out more performance, but in the 
end, Extended Data Output DRAM (EDO) max'ed out asynchronous memory 
performance at ~66MHz.   

In the beginning, SDRAM wasn't much faster than EDO RAM, because they were 
both running at 66MHz. But SDRAM was at the beginning of its design envelope. 
Before long, SDRAM went to 100MHz, and then to 133MHz. There is no longer 
any comparison. 

By using a trick borrowed from Rambus [heck, SDRAM itself is a trick borrowed, 
or stole, from Rambus, but that's another subject entirely. See the end of this
page for more insights along these lines] they managed to push the effective 
clock-rate to 266MHz. But now SDRAM is against a wall, making further speed 
increases difficult, if not impossible.   

Enter Rambus. Right out of the block, RDRAM runs equally as fast as DDR. And 
it's just getting started. Rambus has already announced plans for the next few 
years to crank up clock speeds and quadruple bus width to yield 64-bit with a 
throughput of 9.6GB/s. 

And that was before they announced their new Yellowstone technology, which, 
even with current 16-bit widths and clock speeds, is capable of doubling the 
data throughput rate!   

As for Intel, Rambus does not pay Intel to use RDRAM. Rather Intel pays 
Rambus for the right to use its technology. In fact, Intel recently expanded 
their license with Rambus to include more of their technology & extended it 
to 5 years.

                                                  ***

"I haven't researched these rumors enuf to know if they're true, but I know that 
there are a lot of people who don't like Rambus. It's a matter of principle with 
them *not* to buy Rambus DRAM."

Most people who don't like Rambus subscribe to the propaganda being 
disseminated by memory manufacturers. They don't like selling Rambus  
memory because they must pay a royalty on each unit they sell that uses 
Rambus' technology.

You see, Rambus doesn't manufacture memory themselves. They merely license 
the technology to others (manufacturers). 

In the past, new techniques for improved memory performance were designed 
and patented by the memory makers themselves. Since most companies were 
responsible for one innovation or another, no one paid any royalties, because 
of cross-licensing. 

[Cross-licensing is like: I invent a new suspension, you invent a new steering 
system. We agree not to pay each other, & both sell a better handling car.] 

But Rambus doesn't manufacturer memory. The company merely consists of 
very smart engineers, who design memory. If they don't charge a price for 
others to use their design, they have no way to profit from their efforts. 

Perhaps you can see how it might be objectionable for memory manufacturers 
to pay royalties when they've never had to do this before. So they badmouth 
Rambus, promote DDR, and pray the industry will embrace DDR long enough 
for them to come up with a design that gets around Rambus' patents.

                                                  ***

"The price of Rambus DRAM has dropped dramatically over the last few months, 
but it's still significantly more expensive than DDR RAM, which performs 
comparatively, and perhaps even better (lower latency)."

Lest you think that the royalty they charge is responsible for the higher cost, 
Rambus receives, on average, less than $3 for every $100 worth of RDRAM 
sold. And the fee-per-part goes down with the price. 

RDRAM prices are still more expensive than DDR, but continue to fall .. for the 
very same reasons that the first P4's sold for close to $800. Current, state-of-
the-art, 2GHz P4s, less than 1 month old, sell for half that much now. 

New designs like RDRAM are *very* expensive to implement (think Billions). 
Tweaks to existing designs are relatively cheap (say, tens of millions to yield 
a faster version of an existing part). 

That extra expense has to be recaptured somewhere. If volume continues to 
rise for RDRAM, in a few more months, they will be cost-competitive with DDR, 
royalties and all.

As for performance, similar data rates yield similar results. But because of the 
different interfaces involved, results do vary. DDR does appear to be marginally 
better at "traditional" business applications (such as word-processing). 

Games and audio are pretty much a dead heat. Compute-intensive and linear, 
repetitive operations like video compression favor RDRAM, sometimes by a wide 
margin.

                                                  ***

"The future is a Northwood Pentium 4 with DDR SDRAM and that's where you're 
money is invested wisely for the long run. RDRAM systems, if they still exist after 
Intel's current contract with Rambus expires, will do better with MPEG content 
and video .. "

It should be noted that this ZD article was written *before* Intel announced 
the expansion and extension of their contract with Rambus.

                                                  ***

One last thing...

"...value is not my #1 consideration..." "...I'm far more interested in system 
stability than saving a hundred bucks..."

"'Another factor is the stability and product quality of a system: while all Athlon 
processors suffered from occasional instability in our tests, the Pentium 4 
platform ran without a glitch.'"

I would note that *all* Athlon processors were tested with DDR-SDRAM memory 
while *all* Pentium 4 machines were running RDRAM! Coincidence?

                                                  ***

If you wanted to drive out into the middle of nowhere, hundreds of miles from 
any help, would you rather go in a car with an engine tweaked for every last 
ounce of performance? .. one that had to crank the engine to 6000 RPM, just 
to do 60 MPH?

Or would you rather go in one that has an engine with power to spare? .. 
one that does 60MPH @2500RPM, while asking, no begging you to do 100?

Yeah, me too. =) 
Enjoy your new machine.

John S.
Florida

                                                  ***

Note: Stock prices can sometimes tell us something about how well a company's 
products are doing in the industry. Here is a 5-year graph of Rambus' stock price. 

After climbing like a rocket early last year, it's gone from a high of ~US$120/share 
in the middle of last year, to its current price of ~$8 per share, losing over 90% of 
its market value in the process. It bottomed out in late August at ~$5 per share.

To be fair, I've included comparisons with two of the world's largest memory makers. 
The first is Micron. Here is a 5-year graph comparing Rambus to Micron. The second 
is Infineon. Here is a 5-year graph comparing Rambus to Infineon. Both Micron and 
Infineon look similar to Rambus. 

So the point could be made that the problem isn't with Rambus, but rather the whole 
industry. John says, "The fall of RMBS share price is due entirely to SDRAM! Sales of 
RDRAM are up this year.'

                                                  ***

This section includes John's follow-up comments. In addition to comparing RMBS 
stock price with that of two other major RAM manufacturers, he notes: