PROMOS/Micron D9?

[Donkers]

Thus, is it fair to say that at one point in time the Corsairs were a 'steal', and that right now it's the Buffalo Firestix that has what we all want? Man, isn't competition just great? Redbeard, are you listening?

 

[Tracker]

Well, you are seriously confused my friend. You have some odd logic. Anyways, I was just trying to show you your error in thinking, but I think I will stop. I think you pretty much showed you have no idea what you are talking about if you think Buffalo Firestix are crappy modules because they are "rated" at cas5. It is ok to be wrong though.

One of these days you will get some experience and do your own testing and realize, timings actually make little difference in realworld performance anyways. Im not sure why you are so obsessed with low timings. If you think the difference in speed from cas4 to cas5 is a dramatic difference in speed, you are mistaken. I think you will find these days, that most people buy memory that can overclock higher, and less people that focus on timings. There is a reason for that.

i like how u can convince urself of anything, no matter how wrong it is.. its quite funny.

but yeh.. im not sure why i want cas4..

 

[Tracker]

Thus, is it fair to say that at one point in time the Corsairs were a 'steal', and that right now it's the Buffalo Firestix that has what we all want? Man, isn't competition just great? Redbeard, are you listening?

the firestix are 6400C5.. they suck.

 

[Donkers]

the firestix are 6400C5.. they suck.

They are all Micron D9 chips.. logic dictates they will all reach the same clock speed.

Are you contradicting yourself now? If your logic is right, then you must be confused?

 

[(cf)Eclipse]

Those are all Micron D9 chips. Logic dictates (as ive just said) that they will all be able to do 8888C4.
<snip>
Thus ur claim is reffuted.

you make me laugh, seriously. no two memory sticks are the same, even if they have "micron D9" chips on them

you do realize that ALL micron FBGA chips are marked with "D9xxx" unless they're engineering samples, right? even the mobile SDRAM..
ontop of that, you're telling us that all those chips clock identically? from the revA to revE DDR2 dies? even though i've yet to touch two kits of D9GMH (revD, in 64x8 form fwiw) sticks that clock the same? give us a break, and leave this peaceful memory forum alone

 

[Tracker]

Are you contradicting yourself now? If your logic is right, then you must be confused?

Not all firestix are Micron D9. The ones that are must be the worst of the D9.
My logic claims that if a company can sell a chip as CAS 4, they arent going to sell it as CAS 5 because then they will make less money off of it.

 

[Tracker]

you make me laugh, seriously. no two memory sticks are the same, even if they have "micron D9" chips on them

you do realize that ALL micron FBGA chips are marked with "D9xxx" unless they're engineering samples, right? even the mobile SDRAM..
ontop of that, you're telling us that all those chips clock identically? from the revA to revE DDR2 dies? even though i've yet to touch two kits of D9GMH (revD, in 64x8 form fwiw) sticks that clock the same? give us a break, and leave this peaceful memory forum alone

Did i hit a nerve? Hmm..

Micron D9 chips will overclock to the AROUND the same area, or are capable of this.
Sure, things like revision and type (i.e. GKX or GMH or GCT) come into play, but the differences between them are not all that big.

 

[Tracker]

Here is the basis for my claims on how well each module can overclock.

 

[(cf)Eclipse]

Did i hit a nerve? Hmm..

Micron D9 chips will overclock to the AROUND the same area, or are capable of this.
Sure, things like revision and type (i.e. GKX or GMH or GCT) come into play, but the differences between them are not all that big.

thanks for completely missing the point of my post
so you're saying that anything that starts with the "D9" label will hit about the same speed?

 

[Tracker]

so you're saying that anything that starts with the "D9" label will hit about the same speed?

Prove me wrong.

 

[Manny Calavera]

Actually Geil is better. If u want the best price/performance ill share the list I made.

Put down the crack pipe..... Geil is horrible if one takes into account CS,and after sale support...

BTW: your sig,fits very very well given your outragous posts.

 

[Tracker]

Put down the crack pipe..... Geil is horrible if one takes into account CS,and after sale support...

BTW: your sig,fits very very well given your outragous posts.

I think u are the only one making outragous posts.

I have no idea what CS is, but i cant call them anyway cause i dont live in the US.

To the ppl in this forum - if u want to refute something i have said, ur going to have to do better than "ur wrong IMO".

P.S. You did NOT overclock to DDR2-1000 PC2-8000 with 3-3-4-9 latencies, so whats with the lie?

 

[(cf)Eclipse]

Prove me wrong.

If you had read what I said, you would see that I already have.

But fine, exhibit A.. Let's start with the DDR2 stuff only.
Ever hear of micron D9DQT and D9DQW? revA chips, commonly known as "fatbody", loves low low latencies, like 4-3-2 at 400-500mhz. Loosening the timings up won't get you much past what 4-3-3 can do with any but the most rare chips.
revD (which is the die in D9GCT, GMH and GKX that clocks so damn well) on the other hand scales well with looser timings, I've yet to own a set of D9GMH that couldn't break 600mhz
then there's revE. this is the highest I could get with a 2x1gb kit of D9HNL.
http://www.eclipseoc.com/image/memory/Crucial&#37;20DDR2-667 2x1gb/491.2.png
Far cry from what the revD chips can do, huh?

Now exhibit B: Your claim that ALL micron 'D9' chips clock the same. How can that be, when some aren't even DDR2?
http://www.micron.com/support/designsupport/tools/fbga/decoder
search for 'D9CKG', and notice how it turns up a 200mhz DDR1 chip? I HIGHLY doubt that it can do the >600mhz that D9GMH often does
or, how about 'D9GXM'? some RLDRAM. Exquiste

 

[jcll2002]

You create the Dominator 6400C4 with Micron D9 chips and sell it for an OK price.
Then u secretly switch the chips to the much crappier PROMOS.

If u had wanted to create a Dominator with PROMOS, then u should have created a whole new module and not decieve us by switching it under our noses and keeping it at the exact same price point so that u make a larger profit off of your unsuspecting customers.

Its like if Intel were to suddenly sell us Core 2 Duo's with a Pentium 4 core inside. Or if Dell were to switch out their S-PVA panel off of their 2407WFP for a TN panel without even thinking to tell anyone.

Im very sorry for everyone who bought a Dominator 6400C4 after the switch. It is truly un-fair.
And it hurts me to know that Corsair would do something like this. I respected them as a top memory manufacturer for so long..

that p4 c2d analogy is complete crap. They are advertising something and delivering it to the customer. Youre expecting more which was never guaranteed in the first place.

 

[Tracker]

If you had read what I said, you would see that I already have.

Not without proof, my friend.

But fine, exhibit A.. Let's start with the DDR2 stuff only.
Ever hear of micron D9DQT and D9DQW? revA chips, commonly known as "fatbody", loves low low latencies, like 4-3-2 at 400-500mhz. Loosening the timings up won't get you much past what 4-3-3 can do with any but the most rare chips.
revD (which is the die in D9GCT, GMH and GKX that clocks so damn well) on the other hand scales well with looser timings, I've yet to own a set of D9GMH that couldn't break 600mhz
then there's revE. this is the highest I could get with a 2x1gb kit of D9HNL.
http://www.eclipseoc.com/image/memory/Crucial DDR2-667 2x1gb/491.2.png
Far cry from what the revD chips can do, huh?

Now exhibit B: Your claim that ALL micron 'D9' chips clock the same. How can that be, when some aren't even DDR2?
http://www.micron.com/support/designsupport/tools/fbga/decoder
search for 'D9CKG', and notice how it turns up a 200mhz DDR1 chip? I HIGHLY doubt that it can do the >600mhz that D9GMH often does
or, how about 'D9GXM'? some RLDRAM. Exquiste

First off, let me say how much i appriciate u taking the time to write this.. it was very interesting

Now i can change my claim - All revD Micron D9 chips will reach similar clocks at similar latencies. And yes, i plan to buy revD's

 

[(cf)Eclipse]

Not without proof, my friend.
<snip>
All revD Micron D9 chips will reach similar clocks at similar latencies. And yes, i plan to buy revD's

yeah, sorry.. i've gotten used to people just taking my word for things around here

but for that last bit.. even then, not all revD ddr2 chips will clock the same. there are a few different bins:
400mhz at cas5 (D9GKX)
400mhz at cas6 (D9HCD)
333mhz at cas4 (D9HQT)
333mhz at cas5 (D9GMH)
266mhz at cas4 (D9GCT)
200mhz at cas3 (D9GCV)

while they're all the same die revision. no two chips clock identically, just like cpu's. you can have two cpus with identical stepping, week and batch codes and still get vastly different results. based on how fast each die can go, it goes in the appropriate bin. If there is more demand for a certain chip speed, chips that can make higher bins are sold as that bin that is in higher demand. Since D9GMH is so popular, many of the chips are capable of GKX speeds. Then further binning by the performance memory companies who sell the modules weed out the rest of the lame ones, and the 'bad' chips that actually belong in the bin because they don't clock all that well are often resold or sold as value ram that won't be overclocked. However, out there are DEFINITELY D9GMH chips that can't do 400mhz at cas5, or D9GKX chips that can't do 500mhz at cas5.. overclocking is NEVER EVER EVER a given. It's just that statistically, it is highly probable that high clock speeds will be seen with certain chips with certain levels of binning just because so many other results have shown this.

but this is straying way off topic now

guess the performance difference between 4-4-4-12 and 5-5-5-15 at a given clock speed

 

[Retsam]

I tried to explain to him he is putting WAY too much stock into worrying about timings on memory. He doesnt get it. You are arguing with a guy who thinks a $240 GEIL kit has a better price/performance ration than a $120 firestix micrond9 kit, because his cas3 beats cas4. rofl? Im thinking theres a possibility hes never actually owned any memory before.

 

[Tracker]

Im thinking theres a possibility hes never actually owned any memory before.

You got me there.

 

[Tracker]

guess the performance difference between 4-4-4-12 and 5-5-5-15 at a given clock speed

Its got to be over 20% because then otherwise why would ppl buy a 450$ 6400C3 when they can get 6400C4 for less than 250$? Although i realize its probably less, and then my answer to why i want CAS 4 is that its just my personal goal and challenge.. although if it costs me too much money ill just skip it and settle for 1066 @ CAS5.

 

[Donkers]

Not without proof, my friend.



Now i can change my claim - All revD Micron D9 chips will reach similar clocks at similar latencies. And yes, i plan to buy revD's

Ok! Have you ever thought about going into politics? With everyone else around you being your advisors... and you being in denial, you’d be perfect! Lol.

 

[Tracker]

Ok! Have you ever thought about going into politics? With everyone else around you being your advisors... and you being in denial, you’d be perfect! Lol.

Werent u the guy who asked wich module would get to 1200?

 

[HyperTension]

Wow.. Time to toss my limited knowledge into the forray here. Timings DO NOT matter as much as with DDR2 vs. DDR. Why?

Bandwidth. Simple as that. Your equating high RPM to high speed. Thats not always the case.

Maybe I'm looking into your post's too shallowly, but you seem to be REALLY caught up in timings.... OK.. Rock on.... Tighten those timings as much as you want, or buy a really tight maxed out kit. Congrats... You have just limited your overclock... Thats taking a few steps back huh?

With regards to Corsair.. Good God.. That horse is dead, rotting, and needs to be stuffed in a hole, never to be exposed to air again. The smell of rotting meat is killing open, logical thinking. Redbeard has gone (many, many, many times) over (and did so again here) but you then change routes and start on this timing rant with Eclipse, somebody who has a far stronger grasp on memory / timings vs. what you have shown us here in this thread.

Your thread initially starts off with CONSPIRICY! CONSPIRICY! CONSPIRICY!.. and then when finally exposed as incorrect.. ... ... ... TIMINGS! TIMINGS! TIMINGS!! .

As a gift, I'll email you a xanax and a prevacid.. Please take both, and feel free to repeat as necessary.


P.S. C.S. stands for Customer Support. I'm pretty sure that you have that in your country, or a number to call that will refer you to somebody there (Calling Microsoft, and getting somebody who sounds like the live in India).

Relax.. wow.. Just reading your posts gould give somebody Irritable Bowel Syndrome.

 

[Retsam]

He is fairly exposed from his last comment of guessing 20&#37; increase in performance from cas3 to cas4. I dont know how to respond to that, besides laughing. Try about 2% Maybe 3% or something.

What scares me is, hes been arguing back and forth saying everything I have stated is incorrect/wrong, but he doesnt know the performance difference from different timings? If you are going to argue, you really should have actually done some of your own testing and benchmarks before dealing with others who have. Just a suggestion.

As to your logic as to why it should have a 20% increase, is because of what other reason would people have to purchase $450 cas3 vs less than $250 cas4. This is exactly what I have been trying to explain to you. THERE IS NOT ONE.

This is why I laughed when you said a Geil $240 cas3 kit has a better price/performance ratio than a $120 Firestix micron d9 kit(which is cas5 at stock, but does cas4 easily). You seem to believe that the small jump from cas3 -> cas4 or cas4 -> cas5 yields some ENORMOUS performance difference, when it does not.

The only thing you gain from buying a $240 kit just for cas3, is knowing you just got reamed out of $100 or more.

Anyways, I hope you do not take offense to any of that, that was not my intention. I am just trying to understand your oddball logic on timings and price/performance ratio.

 

[HyperTension]

After re-reading all of your posts and scattered, confused "logic", i have to post this...

--> Flame on...
Your retarded. You CHOOSE to NOT listen to those who are smarter / more experience / whatever and create this fantasy world where Corsair has swindled the masses and timeings rule the world. You honestly thing that all D9 are the same, or will fall into a predictable, linear patten??

Hello, McFly? Are you in there?
How many of us here have bought processors with the expectation that they will overclock well based on the reviews, bin, week manufactured, etc only to get a measley 200 - 300mhz overclock or no overclock at all?

If you want an easier example, look at overclocking video cards... but that is probably expecting too much.

According to your "logic", that just can't happen, they all C2D, A64's, FX's, etc. will ALL overclock, based on their chipset, to a degree of predictability.

WRONG. And the fact that you continue to be abrasive, brash, rude and ignorant despite what others have told you, (minus what you could find out for yourself if you put a bit of effort) puts you on the intellectual level of a rounded brick.

You continue this level of silliness in another post with saying that with BF2, that...

Originally Posted by Hagrid

So yes, 2gb's will make a huge difference.

Then your knowledgeable post:

bullshit. they will make no difference.
graphics card will.

You are a fool. I'd love to waste more time and energy on this, but I think this is a wasted effort, and worthy of a locked thread, as well as a warning from a moderator with regards to your behavior.

 

[Mansize_tissue]

You are a fool. I'd love to waste more time and energy on this, but I think this is a wasted effort, and worthy of a locked thread, as well as a warning from a moderator with regards to your behavior.

I agree completely. This guy simply can't be educated on the matter, especially since he has no idea what he's talking about in the first place.

Tracker, you're asking us for "proof"; where's your proof to justify your compelte disregard on the advice everyone has been so nice (too nice, maybe, considering the attitude you've been giving in return) to give you?#

Edit: Wait, i have more to add. The 20&#37; increase in performance you're hoping for by having memory running at CAS4 timings, rather than CAS5, simply isn't going to happen; i doubt you'd even be able to notice such differences in synthetic benchmarks.

I can't seem to see where this discussion is going any more; are you actually going any where with this, Tracker?

 

[TheCur]

They are all Micron D9 chips.. logic dictates they will all reach the same clock speed.

I stopped reading this thread at this post. As an electrical engineer in the production test field you clearly have zero knowledge about semiconductor manufacturing processes, testing and parts binning.

Please, just stop.

 

[Donkers]

Werent u the guy who asked wich module would get to 1200?

Sorry to dissapoint you, but I even know what to get to reach 1300. Funny thing is, the more I know, the more I know that I don't know enough. Keep an open mind and you'll be ok. Cheers mate!

 

[skadebo]

I drew a diagram since I have learned from experience that sometimes words are hard to understand, even for me. I drew this diagram as much for myself, as for others!

Here you can see the timeline of releases. Why was the first quarter of the year micron based? Well! Redbeard has answered this as they were the only chips capable of running the advertised speeds! Whoa, why did they release a part if they didn't have the proper and cheap chips in the first place? An educated guess would perhaps be that they wanted to remain competitive, and it was essential to release a chip capable of X speeds in order to maintain market share.

Now, in Q2, promos chips are finally ready to be used! AWESOME, now they will use them. Why no price change? perhaps the micron based memory was already being sold at promos prices! Definitely maybe yes, no?

Anyway, in Q3, low end dominator parts are released with promos chips so that the ram is cheap, and the ram is cheap because promos is being used, so that the ram can be sold at a cheaper price. Also, the promos runs at the advertised speed, which is pretty cool.

Finally, this pie chart explains the distribution of the distribution of 6400C4 ram ( play on words intended rolf). You see, Micron based 6400C4 is only 1% of the TOTAL 6400C4 sold TO DATE. haha, to date is an exaggeration on my part. maybe yesterday perhaps? Thus leaving 99% promos.


capisce?

 

[-=Antimatter=-]

I stopped reading this thread at this post. As an electrical engineer in the production test field you clearly have zero knowledge about semiconductor manufacturing processes, testing and parts binning.

Please, just stop.

x2 from an electrical engineer. Don't try to argue what you don't understand, Tracker. The only people who got screwed from Corsair switching from Micron to PROMOS were the ones who forgot that Corsair doesn't have to guarantee anything but the performance indicated on the box.

 

[(cf)Eclipse]

Its got to be over 20% because then otherwise why would ppl buy a 450$ 6400C3 when they can get 6400C4 for less than 250$? Although i realize its probably less, and then my answer to why i want CAS 4 is that its just my personal goal and challenge.. although if it costs me too much money ill just skip it and settle for 1066 @ CAS5.

sorry bud, but it's more like 1% performance difference unless you're running memory specific benches. even when running synthetic bandwidth and latency tests, the difference at the same clock speeds between 3-3-3, 4-4-4 and 5-5-5 are pretty damn small. i can do some benching specific to this for you if you would like


Antimatter, it's not a matter of what profession you are. i'm a mechanical engineer, and i like to think that i grasp the idea pretty well

 

[-=Antimatter=-]

Antimatter, it's not a matter of what profession you are. i'm a mechanical engineer, and i like to think that i grasp the idea pretty well

Well yeah

 

[luihed]

This is one funny thread lol...... I feel like going to a Porche dealership tommorrow and just argue with the salesman on why they charge so much for a 911.....

" A mustang GT can go faster and its half the price"
" Flat 6? you guys charge that much and not even put a V8?"

Maybe Ill make a chart of some sort and try to talk them into a better way to run thier business.......

(Calling Microsoft, and getting somebody who sounds like the live in India).

.

Cos they live in India