RDRAM is cheaper than DDR

[Raystonn]

Straight from www.pricewatch.com, the cheapest of the best performers:
[Note: CAS2 PC2100 DDR at 64MB and 512MB was not found on pricewatch]

(Sorted by price)
$56 - RDRAM 64MB PC800
$102 - RDRAM 128MB PC800
$106 - SDRAM DDR 128MB PC2100 CAS2
$243 - RDRAM 256MB PC800
$264 - SDRAM DDR 256MB PC2100 CAS2
$619 - RDRAM 512MB PC800

(Sorted by type, then size)
$56 - RDRAM 64MB PC800
$102 - RDRAM 128MB PC800
$243 - RDRAM 256MB PC800
$619 - RDRAM 512MB PC800
$106 - SDRAM DDR 128MB PC2100 CAS2
$264 - SDRAM DDR 256MB PC2100 CAS2


Discarding DDR ram that's not CAS2 and PC2100, and discarding RDRAM that's not PC800, we have the above list.

It would appear that the best RDRAM memory solution is cheaper than the best DDR memory solution. The complaints about RDRAM's prices don't seem to be warranted.

-Raystonn

-- The center of your digital world --

 

[kurokaze]

how can you tell the speed of the RDRAM and the cas of the DDR SDRAM from pricewatch?

 

[Kelledin]

Entering "CAS2" or "PC800" or "PC2100" in the search box helps.

Interesting to notice that CAS2.5 PC2100 is still far less expensive than PC600 or PC800. And if the difference between CAS2 and CAS2.5 is comparable to the difference between CAS2 SDR and CAS3 SDR, why pay twice as much for CAS2 anyways?

Kind of changes things once you see the whole picture...

Kelledin

bash-2.04$ kill -9 1
init: Just what do you think you're doing, Dave?

 

[kurokaze]

silly me.. I just use the popups..

 

[Guest]

your probly right in your own way but who cares about price watch i can get cas 2 pc-2100 and i can get cas 2.5 256 for less then $100 as for 512 if you realy want to buy it for that over sized cost whatever. and even iof the cost isnt a big deal with rdram. the praformace is good enough. and last time i checked them your buying 2rdram vs 1 ddr so getting a multi buy discount on rdram vs a single quantity price on ddr isnt a comparison. how ever i havent seen yet that a good stick of rdram(if u wana call it good) is cheaper then a good stick of ddr whitch even at cas 2.5 256 is way faster then teh laggy rdram for under $100. maybe not there but thats not my problem if u wana waist time shopping online.

Computer Shop owner and Head tech.

 

[Raystonn]

"laggy rdram"

Actually, RDRAM has less latency than SDRAM (SDR or DDR) at medium and high memory use situations. SDRAM suffers from dead wait cycles whenever switching from reading to writing and back again. RDRAM does not.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[ejsmith2]

I'm thinking the exact same thing myself......

But for the other poor souls who don't know, I tell them to use RDRAM.

 

[Guest]

i know i was reading your articals before and i realy get annoyed with your posts. i will assume rdram has less lag. and i know it adds more per stick. but the point is that rdram has more lag ok dont jsut look at one point of a rdram system when u add in account the whole system it gets more lag so what if the modual has less lag if the interface to cpu blah blah blah what ever adds more. look at the whole picture not jsut the modual. that would be like saying the cpu is 2gig so it has super fas processing. but the bus is 66mhz so it gets lag. rdram even though i dont realy care if this vary poorly built tech has less lag when the system as a whole gets more lag just for using it.

Computer Shop owner and Head tech.

 

[Guest]

and before u blame it on the p-4. STOP! im well aware of the p-4 rdram will be gone long before sdram in any verson is gone. its always going to cost more for far less praformance. and if u bring up memory bandwidth bench marks from sandra aka crap dont bother. lets look at the bench marks for the end user not the high end business (whitch it didnt get a vary good rating in also.) ddr systems from what ive seen win in all bench marks except a couple meening like 2-4 outa like 15. pc-800 rdram 1.6 or so gigs of band(not dual channel) vs a 266mhz ddr stick that gives 2.1gigs of band ok now dual channel rdram 3gig dual channel ddr 3 gig. and i bet no matter how you pair it up ddr or a sdram system will vastly out praform a rdram system. in most benchmarks. im sure somewhere there is a application that rdram and *hrm p-4 would be better. but for almost every user im willing to bet rdram meens junk. my main point here was to figure how this dude could say rdram is cheaper then ddr. i have yet to see it. so one place sells it cheaper(is thats even true) so they decided to lose money for a while its vary unlikly it will be cheaper.

Computer Shop owner and Head tech.

 

[Guest]

im not trying to sound like im downing you but i was reading your articals or posts. and like some other articals i read, they only paint a part of the full picture.

Computer Shop owner and Head tech.

 

[Kelledin]

Woah...I think someone's having a Jeff K moment here...

Kelledin

bash-2.04$ kill -9 1
init: Just what do you think you're doing, Dave?

 

[Raystonn]

Whatever are you talking about? RDRAM has less latency adding in component and system latency in medium and high memory use situations. In low memory use situations, the latency of memory doesn't matter that much for overall system performance. Also, the P4's frontside bus is quad-pumped to an effective 400MHz. I suggest you read the post on "Bandwidth and Latency: FAO 1 and 2". What you are saying is not technically accurate.

-Raystonn

"Bandwidth and Latency: FAO 1 and 2":
http://forumz.tomshardware.com/hardware/modules.php?name=Forums&file=viewtopic&p=11830#11830


= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[Guest]

this is going to be my final thought on the whole rambus thing. im sick of talking about it and people keep saying "it has less lagg then sdram" and my reply is so what i dont care if it has less lag. tell me why a rambus system is slower in benchmarks tell me why it lagg when it has less lagg. let me put it this way. only one thing counts, the result! and no matter how much we argue about this it makes no difrence who is right or who thinks there right. the only thing that counts is the result and that is that ramdus systems are slower, be it the ram be it the cpu be it both. Result = slower.you can argue this all you want and yes i already read that artical. makes no difrence in any rait when the result is that there slower. why pay more fro a slower system when u can get a faster system with what you say is laggier ram. quad pumped 100mhz, well again a 100mhz double pumped system still beets it,seems to me the design of a ddr system is much better if its suposidly slower and yet still gets better result outa benchmarks. you can and im sure will argue your point again my point was system as a whole not one componant.

Computer Shop owner and Head tech.

 

[Guest]

“tell me why a rambus system is slower in benchmarks”

What benchmarks?? Comparing just memory benchmarks , Rambus wins hands down. Show me a AMD/DDR memory bench that’s faster than a P4/Rambus bench. OH! You mean in overall benchmarks and trying to compare clock for clock, well I don’t think RAMBUS is the issue. Check Raystonn’s analogy in the CPU section that might help you out a little in understanding the P4.

“you can and im sure will argue your point again my point was system as a whole not one componant.”

Well of course. For some reason you are solely blaming RAMBUS for the reason a P4 needs to run @ 40%-50% faster than a T-BIRD?? You keep comparing overall benchmarks, when this post about RDRAM vs. DDR SDRAM, and then associate the overall lack of benchmark performance of a P4 system to RAMBUS which is completely false.


<font color=blue>The #1 reason to upgrade your PC - to run faster benchmarks...</font color=blue>

 

[Raystonn]

We were discussing memory, not CPUs. The proper forum for CPUs is over in the CPU forum. The overall performance of a system is total total performance of all of its components. As far as the memory component goes, RDRAM is superior to SDRAM. Other factors will of course affect overall system performance.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[jg38141]

Doesn't matter much either way since the court in VA today just threw out all remaining cases against infineon by Rambus- which means (even though rambus appeals, blah, blah, blah) once the precident gets set, which it just has, rambus no longer will get money for patents on sdram technology which means they loose 50% of their income. Rdram and the company that markets it is rapidly fading. Good ridance.



My Jesus is whiter than your Jesus.

 

[Raystonn]

RAMBUS is just the company that originally designed RDRAM. They do not manufacture it. The dozens of memory manufacturers are the ones responsible for producing and marketing RDRAM. As more is produced, the prices will continue to drop.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[ksoth]

"Also, the P4's frontside bus is quad-pumped to an effective 400MHz."

Actually isn't the fronside bus only 100 mHz and the processor bus is what is quad-pumped to an effective 400 mHz. Kind of like how Athlons run off a 133 mHz frontside bus but have a DDR processor bus to an effective 266 mHz.

"We put the <i>fun</i> back into fundamentalist dogma!"

 

[Bud]

Nothing personal man, but I've read several of ur posts, and while you seem very knowledgeable in a technical sense, you seem obsessed with rdram and that's probably because you work for those people (intel). I guess what I'm saying is I don't trust you Ray. You seem to be running a commercial for rambus. All I know is that rdram is expensive and unless it's used w P4 there's no overall system speed gain. Tom posted the benchies for that stuff ages ago. And it's my understanding (limited though it is) that more rdram means more lag, And I hate lag.

Is anyone else getting tired of the less lag thread??

It seems even if someone can go toe to toe with you on the technical stuff, all you do is wait a couple days and post the same thing again. That's called propaganda friend. Say it long enough and people will believe …even it's not true. Regardless whether ur right or wrong, your method smell like a commercial to me.

No offence….Bud



I thought I was wrong once, but I was mistaken.

 

[Raystonn]

All of this information is available at http://developer.intel.com.

The NetBurst micro-architecture features include a 400MHz system bus. This is a quad pumped bus running off a 100MHz system clock making 3.2Gbytes/second data transfer rates possible.

The Pentium 4 processor in the 423-pin package uses a new scalable system bus protocol referred to as the "system bus" above. The Pentium 4 processor system bus utilizes a split-transaction, deferred reply protocol similar to that of the P6 processor family system bus, but is not compatible with the P6 processor family system bus. The system bus uses Source-Synchronous Transfer (SST) of address and data to improve performance. Whereas the P6 processor family transfers data once per bus clock, the Pentium 4 processor transfers data four times per bus clock (4X data transfer rate, as in AGP 4X). Along with the 4X data bus, the address bus can deliver addresses 2 times per bus clock and is referred to as a 'double-clocked' or 2X address bus. In addition, the Request Phase completes in one clock cycle. Working together, the 4X data bus and 2X address bus provide a data bus bandwidth of up to 3.2Gbytes/second .

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[Raystonn]

"rdram is expensive"

It's cheaper than PC2100 CAS2 DDR ram.

"unless it's used w P4 there's no overall system speed gain"

It requires a processor with a system bus that has enough bandwidth to actually make use of the increased memory bandwidth. Driving a ferrari on the freeway isn't going to produce 180mph speeds. You must eliminate the other bottlenecks first. The Pentium III, as used in Tom's RDRAM benchmarks previously, suffered from system bandwidth problems and thus could not take advantage of the benefits of RDRAM's increased bandwidth.

"more rdram means more lag"

Yes, a bit more. It still has less lag (latency) than SDRAM under medium and high memory use situations.

"wait a couple days and post the same thing again"

Sorry, I haven't been doing this. I always respond to any comments, questions, or arguments. I never ignore them and repeat again later.

"your method smell like a commercial to me"

If the truth smells like a commercial, perhaps you need more commercials in your life. I'm not here to brainwash people. I'm here to present the facts. Feel free to avoid the facts if you do not like them.

-Raystonn

= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[Guest]

i dont remember compairing only rambus as the problem sinple one of them. and u can say its better al you people want to but i dout it will be system memory for much longer since even intel admits that rambus (of course they want to deflect off there cpus) as the problem. however i dont care and already knwo about p-4's problem and as far as im conserned when u need to dual channel and make it 800mhz just to get 3gigs? since when was that good since ddr can do 2.1gigs at more then half the clock and none dual channel. and only pc-800 single channel has more bandwidth then pc-133 sdram. like the p-4 it need to be clocked higher and dual channel to even compiet. and the point of this post was someone asking what memory to get i simply added my thought on it and somehow all these rdram people are jumping on me. well like intel people i dont care what is said. pc-800 ram is the only good rambus and it needs to be pumped way higher to reach more bandwidth. i meen quad pumped dual channel ddr gets 6.2gig of band at a 500mhz(about clock) rambus cant even do that at 800mhz. like i said before so what. maybe if the cpu was made better rambus wouldnt look so bad but the fact is that it dose and intel wont be making quality as long as they keep letting amd push them like they are. if they wouldnt worry about amd they would wait for quality and not have there rambus p-4 thing flop like it has. and i dont knwo one person that runs bandwidth only relyed apps. so wtf.

Computer Shop owner and Head tech.

 

[Kelledin]

If RDRAM has less latency than SDRAM, then why was the i820 defeated by a similarly configured i815? Both were restricted to the same bandwidth by the CPU. This was in a pair of Intel's own benchmarks, in case you don't remember.

Kelledin

bash-2.04$ kill -9 1
init: Just what do you think you're doing, Dave?

 

[Raystonn]

Would you mind giving a link to these benchmarks? I'd be interested to see if these were low memory-use benchmark apps. (The kind that would not do lots of switching between reading and writing.) Lots of switching from reading to writing or vice versa is the achilles heel of SDRAM.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =

 

[Kelledin]

Silicon Strategies still has the article:

<A HREF="http://www.siliconstrategies.com/story/OEG20000708S0001" target="_new">http://www.siliconstrategies.com/story/OEG20000708S0001</A>

The benchmark scores were originally on Intel's web site (linked in the article), but they've since been replaced.

Kelledin

bash-2.04$ kill -9 1
init: Just what do you think you're doing, Dave?