Discuss Quad Band Memory in the Memory forum on Dev Hardware. Quad Band Memory Memory forum for discussing chips and modules from brands such as Corsair, Patriot, OCZ, Mushkin, G.Skill, Crucial and Kingston. Learn to fix RAM problems, overclock memory, adjust latencies, change timings, and use Memtest86.
ASP Free and Iron Speed Designer are giving away $5,500+ in FREE licenses. Iron Speed's RAD CASE toolset can save up to 80% of your coding time. One free license per week, one perpetual license per month!
Open source technologies have proven to be extremely advantageous to businesses.
This adaptive and highly efficient kind of software is central to the infrastructure of most companies. Since the world of open source technology is constantly changing,
IT professionals need a resource to give them up-to-the-minute information about these enterprise level and open source technologies. Dev Shed is that resource .
The ASP Free website provides in-depth information on the latest developer tools available from Microsoft. Our cadre of writers, highly experienced industry experts, reveals the best ways to use established technologies as well as new and emerging technologies. Our coverage of Microsoft's development and administration technologies is among the most respected in the IT industry today. .
Posts: 207
Time spent in forums: < 1 sec
Reputation Power: 9
Quad Band Memory
So, it looks like OCZ is currently working on Quad Band Memory. Wow.
This shit is insane.
Remember SDRAM? And how it ran at 100/133MHz?
Then came DDR SDRAM, which pretty much doubled the speed. 266, 333, 400MHz and higher.
Now this Quad band stuff?
533MHz, 667MHz and 800MHz. Wow.
The 800MHz sticks run a bandwidth of 6.4GB/sec, put a couple sticks in dual mode, and you've got yourself a whopping 12.8GB/sec.
Of course, 12.8GB/sec would be pretty useless, as the Pentium 4 has a theoretical maximum intake of 4.2GB/sec
What do you guys think?
Oh and does anyone know the maximum intake of AthlonXPs?
wtf is VIA PT600??? I don't suppose Quad band memory will be supported on any of the current boards right? I mean it's probably backward compatible but can you really get that speed on any of the current boards?
Posts: 4,290
Time spent in forums: 2 Days 30 m 34 sec
Reputation Power: 21
You should let me add "QBM will never see the light of day" to your poll...
With only Via backing it from a chipset standpoint, it'll never become main stream, and get off the ground. Via doesn't know a good memory controller from their ass.
Regardless, once DDRII matures, and also moves dual channel, there will be no need for QBM. The only application is for the P4's, and intel has a solid solution allready. Two in fact, that will fill the bus, although RDRAM is beginning to fall off the map. AMD, since they are still running dual pumped busses, have no need for this technology. It's dead end before it even starts.
The max intake of the 133fsb athlons is 2.1GB/s. The 166fsb Athlons have 2.7GB/s max data throughput.
__________________ "There's no happy ending to cocaine. You either die, you go to jail, or else you run out." - Sam Kinison
Posts: 1,603
Time spent in forums: 4 h 11 m 31 sec
Reputation Power: 12
bye bye rambus
hello dual channel ddr
correct me if im wrong, but ddr II is basically just better ddr, right?
anyone know if the Athlon 64 supports dual channel ddr? i know the opteron does, which makes me think the athlon 64 doesnt, which kinda sucks =/. well hopefully it will remove the limit of the NB on the frequency the memory is run at. the memory will then be able to theoretically run at the same frequency as the processor! that makes ddr II useful, but where does dual channel DDR come into play w/ Athlon 64? does it?
[edit]
mmm, i like ukcuf's avatar!
__________________
Addicted to OCAddiction!!!
Last edited by cabe : March 13th, 2003 at 12:33 AM.
Posts: 207
Time spent in forums: < 1 sec
Reputation Power: 9
Aha, stupid me.
So basically since the 133 athlons use a Front Side Bus that is basically doubled, to 266MHz, it's the same bandwidth as 266MHz memory (which is of course 2.1GB/sec). This is simple, and makes complete sense, it's a wonder I never figured this out before.
So basically, in *theory*, using ddr333 in async mode on a 266fsb wouldn't improve performance over ddr266 in sync with the FSB. The same must go for running ddr400 on a 166(333) FSB.
Now when it comes to the P4, they have quad pumped FSB's, so if you have ddr266 running in sync with a 533 fsb, you're only providing half the maximum bandwidth the pentium 4 can handle right? And the only way to get more bandwidth out of the memory is to use a higher FSB, but in turn that also increases the processor's maximum bandwidth intake too, correct?
So this bring me to the conclusion, since ddr ran at double the speed of sdr, and now qbr will run double the speed of ddr (or 4x that of sdr) then that'd be perfect for Pentium 4's with their quad pumped front side buses!
Running a 533FSB and 533QBR memory, it'd bea perfect tandem! And later on, QBR800 will be released, which'll be perfect for the then released P4's with 800fsb.
This makes complete sense to me, am I correct about this? DMOS what's your opinion you big knowledgable bastard, you
Posts: 1,603
Time spent in forums: 4 h 11 m 31 sec
Reputation Power: 12
thats right, which is y i always say running sync is better, and yea you are only using half the badnwidht for the P4, which is y they are now coming out w/ dual channel since rambus is turning belly up, and yes, higher FSB is the only way you can get more memory bandwidht out of a P4.
and no, QBR will not be perfect, in theory, octouple band memory would be better, but i dont see such a thing, and until i see it getting more support, ..just think of it as a small possibility, that *might* happen. look at what happend to rambus.
Posts: 4,290
Time spent in forums: 2 Days 30 m 34 sec
Reputation Power: 21
It would be perfect, but thats theory. In practice, it just isn't going to work. With only one memory manufacturer and one chipset maker backing it, while all the others embrace DDRII, that alone is enough to put nails in QBM's coffin. Then add the fact that Intel doens't want anything to do with Via, and won't give them a bus license, and that AMD doesn't need QBM, well, you have a marketless product. Having engineers do the heavy work to create a new technology is only half the battle, the marketing and business types have to do the rest.
Posts: 7,892
Time spent in forums: 1 h 53 m 58 sec
Reputation Power: 218
Umm, because VIA sucks ass? Not to mention Intel doesn't like them taking money from their pockets by releasing Intel based chipsets. Nosir Intel does not like it one bit.
Posts: 3,185
Time spent in forums: < 1 sec
Reputation Power: 13
correct me if im wrong, but if using say a 133mhz fsb proc and a stick of ddr333 ram (like me), is there any gain in putting the mem frequency async to the cpu and having it at 166 (like me again).........
Posts: 106
Time spent in forums: < 1 sec
Reputation Power: 9
I think that Quad Band Memory (QBM) will really only be applicable in server environments running the new AMD 64bit CPU. I am guessing that will be the only CPU this year to take full advantage of the speed of QBM. Agree?
Posts: 182
Time spent in forums: 12 h 2 m 26 sec
Reputation Power: 9
Why are u all thinking about intel. DO you think that AMD will stay at 2x fsb. This is a perfect opportunity for AMD to catch up with intel on Mhz and leave intel in teh dust. As some of you said rearlier with via and the manufacturer backing it up it wont go far, but if AMD backs it up too, every mobo manufacturer that supports AMD proccesors is going to start making mobos that support QBM memory. This will also be faster that DDR2 since when it will hit the market it will have higher Mhz and more bandwith than ddr2.
Posts: 207
Time spent in forums: < 1 sec
Reputation Power: 9
Quote:
correct me if im wrong, but if using say a 133mhz fsb proc and a stick of ddr333 ram (like me), is there any gain in putting the mem frequency async to the cpu and having it at 166 (like me again).........
That depends. Like I was saying, P4's run on quad-pumped FSB's, turning that 133 number into 533MHz. This means that a P4 on a 533 Bus can use up to 4.2GB/sec of memory bandwidth.
AthlonXP's effectively double the FSB, from 133 to 266MHz. This means an AthlonXP can use up to 2.1GB/sec of memory bandwidth.
If you're using a P4, at 533MHz Bus, and you're running the DDR memory at 333MHz, theoretically you can provide the P4 more bandwidth than at 266MHz, enhancing performance.
If you're using an AthlonXP, at 266MHz Bus, and 333MHz DDR memory in async, there's not much point since the memory is providing 2.7GB/sec of bandwidth, but the processor can only use 2.1GB/sec of that available bandwidth.
Quote:
I think that Quad Band Memory (QBM) will really only be applicable in server environments running the new AMD 64bit CPU. I am guessing that will be the only CPU this year to take full advantage of the speed of QBM. Agree?
As it stands right now, there will only be one manufacturer producing QBR memory (OCZ), and one manufacturer providing the chipset(s), that's VIA. The chipsets will be used with the P4, not with *any* AMD processors. This could change in the future of course.
Quote:
Why are u all thinking about intel. DO you think that AMD will stay at 2x fsb. This is a perfect opportunity for AMD to catch up with intel on Mhz and leave intel in teh dust
Yes I do think that AMD will stay at 2x FSB. If you can show me an article suggesting otherwise, then please provide me with a link to it
Also, *ANYTIME* would be a perfect time for AMD to catch up with Intel in the MHz race, but AMD just can't match Intel. Simple as that. Yeah AMD XP's do more work per clock cycle (hence the xxxx+ processor names), so a 2.4GHz P4 is comparable to an AthlonXP 2400+, but MegaHurtz wise, Intel's ahead and will stay ahead in the forseeable future.
Quote:
actually it will have lower MHz but higher bandwidth
MHz directly relates to bandwidth, no? I think DDRII will debut at 533MHz, and QBR will debut at 533MHz and 667MHz. QBR 533 is 133MHz x 4. Then later will come QBR800MHz (mmmm...plenty bandwidth there my folks).
LOADING INFUSIONSOFTLOADING INFUSIONSOFT 1debug:overlay status: OFF overlay not displayed
overlay cookie defined: TI_CAMPAIGN_1012_D
OVERLAY COOKIE set:
status off