Intel Please tell us more information about Conroe such as FSB clock speed and other feaurtures besides 4mb of L2 cache for last year or so thats all been adding to your products.
Ask intel how they feel about this.
"Rollins acknowledged that Dell, in response to customer demand, had begun selling stand-alone processors from Advanced Micro Devices, a chief rival of chipmaker Intel Corp., on its Web site. Dell has long relied exclusively on processors made by Intel, and Rollins gave no indication that the company would introduce computers with AMD chips anytime soon."
from
http://biz.yahoo.com/ap/051110/earns_dell.html?.v=13
"Rollins acknowledged that Dell, in response to customer demand, had begun selling stand-alone processors from Advanced Micro Devices, a chief rival of chipmaker Intel Corp., on its Web site. Dell has long relied exclusively on processors made by Intel, and Rollins gave no indication that the company would introduce computers with AMD chips anytime soon."
from
http://biz.yahoo.com/ap/051110/earns_dell.html?.v=13
Questions about onboard sound. :thumbsup:
Too bad NVIDIA won't be there. We could grill them about not having SoundStorm anymore :|
Too bad NVIDIA won't be there. We could grill them about not having SoundStorm anymore :|
THere is nothing insufficient about SATA/SAS for solid state drives. Using old RAM as a storage methoer is infeasible, because it's volatile and thus requires constant power. And these devices already exist on the market see here.
The real question should be, is Intel going to apply 65nm process to flash ram? How much will this increase spead and reduce power consumption/costs versus current flash RAM production? (info here)
This really isn't an issue for Intel. All current Asus boards and most of the other current motherboard can boot from USB devices no problem.
This would be a support nightmare.
You can't standardize everything...if you do you'll just end up with a Mac, and nobody wants that.
the question at the top of my list is: when can we expect ASUS's Crossfire boards to hit the market, such as the A8R-MVP?
I suppose I'm most curious about the timing, features, and pricepoint of the first quad core chips (desktop or workstation).
Yeah, you have to be careful in what you ask, things about specifics of products like Conroe etc will probably not be answered, since they are still confidential facts, and if Intel were willing to let people have that information, it would already be posted up as new on many websites. e.g.
Did you read the i-ram review?? SATA is DEFINITELY a limiting factor for ram, most modern ram 1500MB+MB/sec, how is SATA 'fast' again? ughh... Seriously, where is your head?
As for old ram, well there's two ways these companies could go about it: Shrink and make it a product or let other companies like gigabyte pick up the slack with their i-ram, etc.
I'd like to know if that 'old tech' ram, could be made super small (and therefore very dense) and you could make solid state storage devices out of them, sure power consumption would be there but it's not like it would be very different from modern hard disks. I know they would never be as big as hard disks, but my god they could do a lot better then the i-ram if it was designed *as a product*, instead of as a 2ndary memory slot add-in card, ala the i-ram.
But 'current tech' RAM is essentially what you describe, old tech RAM made denser and repackaged to be compatible with a faster interface. What would be the point of revisiting old designs with new tech?
Where's your head? Volatile RAM as a storage medium isn't a solid state drive. If it requires power to keep it's data, it's not a drive. Period. And even the fastest of current non-volatile RAM wouldn't be bottlenecked by SATA 3G or SAS. And what desktop user would be limited by 300MB/s? I don't know of any Battlefield 2 maps or Sims 2 neighborhoods that are in excess of 300MB. As for iRAM, who cares if its limited to 300MB/s either? Its max capacity is far more limiting than it's STR.
Old RAM is just that. Not manufactured anymore. That means there would be no reason to apply a die-shrink to it.
Asus: Are there any plans to release any additional nForce 4 Ultra boards, such as one with a heatpipe like the A8N-SLI Premium?
ASUS: Do you plan on changing your cooling solution for NF4 chipsets? It appears the failure rate of the fan is high. Is the issue a quality issue with your supplier or a poor cooling design?
bob4432
Lifer
- Sep 6, 2003
- 11,726
- 45
- 91
AMD: please start pushing your mt turion series/future line of mobile cpus to the desktop. a 25W cpu would be great with power costs going through the roof, and you could also push the marketing into being more "earth friendly".
ASUS: when are you going to start offering Turion/AMD mobile support on your desktop boards for low power and low heat pcs with good performance? with the state the country is in and with power costs going through the roof, a cpu like this would be a welcome addition especially to those that leave their machines on 24/7. this would not only help the enthusiast but also business by keeping their server rooms more economical to keep cool vs the "normaly" cpus either amd or intel. an additional benefit would be that you could push the marketing into being more "earth friendly"
ASUS: when are you going to start offering Turion/AMD mobile support on your desktop boards for low power and low heat pcs with good performance? with the state the country is in and with power costs going through the roof, a cpu like this would be a welcome addition especially to those that leave their machines on 24/7. this would not only help the enthusiast but also business by keeping their server rooms more economical to keep cool vs the "normaly" cpus either amd or intel. an additional benefit would be that you could push the marketing into being more "earth friendly"
*** A old designs can be done cheaper, do not have to run at high clock frequencies (i.e. DDR, DDR2's, vs PC66 SDRAM) to be a lot faster then modern hard disks (i.e. PC100 is something like 1200MB/sec, thats a hell of a lot faster then any kind of flash or hard
disk drive)
*** Higher yields due to not having to increase clock frequencies,
*** Less interface/trace complexity/interference, the higher your clockspeeds on ram, the more expensive everything gets.
I do not know the finances of such a thing but I don't think anyone has seriously looked into it, i.e. Gigabyte thought that a "Ram drive" device would have a market, but the problem is they are trying to use modern ram sticks, if I were to design an "i-ram" I'd revisit older PC 66/PC100 SDRAM, less complexity, clockspeeds, etc, and with the right high bandwidth interface (i.e. 1200MB/sec versus 300MB/sec ATA), one could do hybrid systems much better then today.
They've already started talking about flash/hard disk hybrids, my question is why not hybrids with ram/battery + hard disks, etc? personally as an enthusiast I'd love to have a product like that )battery backedup) to run my OS on and keep 'all the storage' on traditional Hard disks / flash.
Where did the 300 MB/sec ATA come from? The I-RAM is a standard PCI card, which means that the bandwidth is limited to 133 MB/s. Of course, such an implementation lends itself easily to a PCI-E port with the addition of a bridge chip, and if implemented as an x4 solution, it should have plenty of bandwidth. The other thing to note is that with such drives, the latency improvement is far more critical than the bandwidth improvement, as seldom is it ever necessary to access 100+ MB of data all at once (especially in average user-level systems, which is what the IRAM targets), and even a relatively low bandwidth bus like PCI has a much better latency than accessing a standard HDD.
So the problem with using old RAM is twofold. First, a quick look at pricing indicates that with very few exceptions, DDR400 RAM is the cheapest per-GB memory currently available. This is due to the fact that all the major manufacturer's turn out tons and tons of the stuff. In order to get old RAM competitive again (regardless of the easier producibility), it would be necessary not only to increase the density as you say, but also get all the major players to start producing the high-density old-style RAM again, which is far easier said than done. I highly doubt that any single individual or company has the clout and/or resources necessary to do such a thing.
And then the second problem relates to the latency thing. Take a bunch of this older style RAM and shove it on a high bandwidth interface, and what do you get? A moderate latency system that will outperform the lower latency DDR400 based system *only* during those very rare instances of sustained sequential reads in the order of 100+ MB at a time. It just doesn't make sense to "optimize" the solution for that sort of thing, unless you make the target be enterprise class data-mining systems, which typically already come with vast quantities of standard RAM that they end up using just like a RAM-drive anyways.
I think that gigabyte made the right decisions when implementing the I-RAM, in that they provide something that has the potential to be a cost effective (relatively) and extremely low-latency storage solution for typical desktop users that's based on technology that's readily available, without having to resort to proprietary storage tech, or to investing potentially billions of dollars in the cost of R&D and getting the industry to start producing updated "old RAM" solely for use in RAM-based storage applications.
I understand what you're saying but the equipment they used to design the old ram *doesn't go away*, it doesn't just magically disappear, we're talking about IC's here, we're not talking about tool and or car manufacturing. I really REALLY doubt it would be expensive to take a look at older SDRAM, and turn it into a product (i.e. not an i-ram where people have go out and buy 'sticks' of it to put it into their IRAM). Have an actual small drive, ala something like Western Digital did with the Raptor (34GB/74GB 10,000 rpm) have it as an add-in card (like they did for 3D video cards, notice the amount of integrated RAM on modern video cards) or SATA device by itself, , or have it on SATA, IMO they could probably use older ram designs on modern SATA/PCI/PCI-E, and have a much better product then i-ram for a lot cheaper with no heat issues, trace issues, timing, etc, because the technology has been 1) Perfected and 2) ridiculously cheap because all the R&D has already been done, all it would require is 1) Sticking ram chips onto an PCB + controller. 2) Battery backup. This is not rocket science here.
Cerb
Elite Member
- Aug 26, 2000
- 17,484
- 33
- 86
1. Regular desktop users do not want or need a RAMdisk. A HDD with very large (128MB+), and 'smart' caches to hide latency and bandwidth for small amounts of data would be cool (a few companies are working on those, IIRC), but a RAMdisk is very much a low-volume, enthusiast and workstation only thing. There is no way, with such a small demand, to make the product both cheap and worthwhile to build.
2. The R&D has been done for DDR and DDR-II, as far as the RAM itself is concerned.
Gigabyte has to stick a controller that will convert it to a standard drive, as far as the SATA controller knows, and a SATA controller (straight PCI-E alone would could offer bettwer bandwidth now or pretty soon, depending on chipset), then make a BIOS for them. Using older RAM might make for a slightly simpler PCB, but otherwise, it wouldn't be any easier. They'd have to do as much R&D for SDRAM as DDR or DDR-II.
3. The old technology might not go away, but it also does not get cheaper. Right now, DDR-II 400 would be the most cost-effective RAM to buy. Why worry about SDRAM? It's more expensive.
4. As with #1, a product like the I-RAM does not, and will not, have enough of a demand to warrant special memory upgrade modules. It ceases to be cost-effective at that point.
Technically, what you're taling about would be pretty cool. But who will buy it? As it is, very few people are going to buy a product like the I-RAM, and it, as it currently exists, is a simple and awesome way to implement a RAMdisk.
Er, uh, sorry for the hijack. I actually did come in with a question
.Asus: With Intel transitioning to a cooler processor design, what do you think the future holds for BTX, will it die off, gain popularity, or continue to be a niche product mainly used by OEMs.
When can we expect to see the A8S-X (SiS based socket 939) board available? and are there any plans for a ULi based board?
Does Asus have any plans to produce a multimedia card for PCI-E?
When can we expect to see the A8S-X (SiS based socket 939) board available? and are there any plans for a ULi based board?
Does Asus have any plans to produce a multimedia card for PCI-E?
ASUS: When will we see XDR motherboards out there?
Intel: When will we see 128bit cpus?
Intel: The more cored cpu's that will be released the harder we're making it for software developers. True or false? Otherwise the cores go to waste and lay unutililized.
Intel: When will we see 128bit cpus?
Intel: The more cored cpu's that will be released the harder we're making it for software developers. True or false? Otherwise the cores go to waste and lay unutililized.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter