Thursday, January 10th 2013
Crucial Demos DDR4 DRAM Modules at CES
Crucial, a leading global brand of memory and storage upgrades, today announced its first DDR4 DRAM demonstration at CES 2013. As a global brand of Micron, the Crucial DDR4 DRAM demonstration is based on Micron's 30-nanometer (nm) technology, the 4-gigabit (Gb) DDR4 x8 part is the first piece of what is expected to be the industry's most complete portfolio of DDR4-based modules, which will include RDIMMs, LRDIMMs, SODIMMs and UDIMMs (standard and ECC).
The new Crucial DDR4 DRAM modules use up to 20 percent lower voltage than previous technology, enabling smaller, more efficient form factors and longer battery lives. The DDR4 DRAM will have lower voltage, operating at 1.2V in comparison to current DDR3 offerings operating from 1.35V to 1.5V. The new memory modules enable mainstream data rates that are up to twice as fast as DDR3 memory. This will enable systems to process data faster, load applications faster, improve system responsiveness, and increase the ability to multi-task and handle data-intensive programs.
"The initial readiness of our DDR4 product portfolio represents a big step forward in terms of power and performance improvements for our customers," said Robert Feurle, vice president for Micron's DRAM Marketing. "A variety of speed and density options enable leading edge modules for the most demanding applications."
The first public demonstration of the Crucial Ballistix DDR4 DRAM modules will be held at the 2013 International CES, January 8-11th
"With DDR4 data rates already twice that of DDR3 when it was introduced, Crucial Ballistix is the ideal showcase for Micron's latest generation of DRAM," said Jim Jardine, director of DRAM product marketing, Crucial. "Other benefits, including reduced voltage compared to DDR3, will make DDR4 an important advancement in memory technology that will benefit computing systems that range from the personal computer to the data center."
The Crucial DDR4 DRAM modules are backed by a limited lifetime warranty and are rigorously tested to meet or exceed the high-quality performance specifications customers have come to expect from Crucial. New DDR4 DRAM modules are expected to ship in late 2013.
The new Crucial DDR4 DRAM modules use up to 20 percent lower voltage than previous technology, enabling smaller, more efficient form factors and longer battery lives. The DDR4 DRAM will have lower voltage, operating at 1.2V in comparison to current DDR3 offerings operating from 1.35V to 1.5V. The new memory modules enable mainstream data rates that are up to twice as fast as DDR3 memory. This will enable systems to process data faster, load applications faster, improve system responsiveness, and increase the ability to multi-task and handle data-intensive programs.
The first public demonstration of the Crucial Ballistix DDR4 DRAM modules will be held at the 2013 International CES, January 8-11th
"With DDR4 data rates already twice that of DDR3 when it was introduced, Crucial Ballistix is the ideal showcase for Micron's latest generation of DRAM," said Jim Jardine, director of DRAM product marketing, Crucial. "Other benefits, including reduced voltage compared to DDR3, will make DDR4 an important advancement in memory technology that will benefit computing systems that range from the personal computer to the data center."
The Crucial DDR4 DRAM modules are backed by a limited lifetime warranty and are rigorously tested to meet or exceed the high-quality performance specifications customers have come to expect from Crucial. New DDR4 DRAM modules are expected to ship in late 2013.
18 Comments on Crucial Demos DDR4 DRAM Modules at CES
technewspedia.com/futurology-haswell-ep-will-have-14-cores-and-35mb-l3/
Just note:
Haswell/Ivy Bridge-E = DDR3 up to ~2600-2800 MHz
Broadwell/Haswell-E = DDR4 up to ~3600-4400 MHz
en.wikipedia.org/wiki/DDR4_SDRAM
You will see more low voltage (1.35v) DDR3 as DRAM makers try to stimulate DDR3 sales because DRAM prices are in the cellar and staying their with PC sales down and a rocky economy expected for years more.
www.nbcnews.com/technology/technolog/holiday-pc-sales-slide-first-time-5-years-1B7931427
APUs which are getting rapidly popular, and (almost all) IGPs are using the system memory as video memory, and the faster the RAMs are, the better GPU performance you will get (mainly because you can feed more to the ROPs). Besides the obvious GPU performance benefits, you also gain computational speed increase as well (e.g.: compressions, offline CGI rendering, etc), and that's something you usually also do on desktop workstations. Just my two cents.
APUs are the only place on desktop PCs where faster than 1600 MHz. RAM shows any tangible performance gains in real applications. Synthetic memory testing benches exaggerate the tiny gains from increased RAM frequency above 1600 MHz. and those who don't know any different think that will improve system performance, but it doesn't because the RAM isn't a bottleneck @ 1600+ MHz.
The IMC has nothing to do with the bandwidth per se of the system and isn't what determines the RAM frequency that is sufficient to eliminate system bottlenecks. The IMC just moves memory control inside the CPU instead of having a discrete chipset controlling the flow of data. Both Intel and AMD systems have been documented to have sufficient bandwidth with just DDR3 @ 1333 MHz. but 1600 MHz. gives you bandwidth to spare and no system bottleneck.
Folks can prove this by running real applications and timing them. There is no tangible gains for enthusiasts or anyone else. This has been proven over and over and I suggest people prove it to themselves with back-to-back test on their own desktop PC by running their DRAM at 1600 MHz. then upping the frequency to it's fastest stable speed and running the real applications again.
Most people will be shocked to find no tangible, i.e. meaningful system gains, unless they understand that DDR3 RAM @ 1600 MHz. is more than enough bandwidth to handle current CPUs and then some. That is why there is no tangible gains except for APUs where the GPU section is able to take advantage of higher frequency RAM. Those who aren't up to speed technically on system performance can easily be duped into buying faster RAM that does not deliver any useful performance gains at all in real apps.
I could agree with you about tighter memory timings and lower latency are nothing but marketing gimmicks for the syntetic benchers, but can't agree about memory speeds, which is indeed delivers more performance with the most popular integrated GPUs (HPGs/APUs and EPGs) and with many applications as well, and the sooner the DDR4 will become mainstream, the better is for everybody.
My parts trickle down to other systems, so even if it's giving me a marginal gain now, it may be more substantial later on.
IMC matters a TONNE, too. AMD's IMC sucks, and Intel's is far better. Most performance differences between the two can be summed up to differences in CPU cache, IMC, and memory performance.
Perfect example is how an FM1 APU will kick the snot out of an AMD 1100T, same ram speed, in CodeMasters F1 titles(Dirt too). An FM1 APU is 50% faster than an 1100T, even with the lack of L3 cache(same CPU cores). Boosting ram performance matters little here on the APU, as shown.:
Just so we are clear, the APU and the 1100T are both using an HD6950 here. So the APU's onboard GPU is not what's seeing the benefit here. It's purely about memory control(IMC).
Anyone who doesn't want to bother running extensive test to see for yourself that RAM speeds above 1600 MHz. do not provide any significant, i.e. tangible system gains, can go to AnndTech or Tom's Hardware and see independent confirmation on RAM frequencies from 1333 MHz. to 2000 MHz. on multiple AMD and Intel platforms. What you will see is slight variations depending on the software tested but the overall conclusion by AnandTech and Tom's based on their extensive testing, including rendering, is that "there is NO SIGNIFICANT system performance gains with the faster RAM."
Numerous people in several PC enthusiasts hardware forums have conducted their own test and even those who are proponents of using faster RAM had to admit after posting their own test results in RAM intensive software such as video games, that there was no significant system performance gains from testing at 1333 MHz. to 2000 MHz. That is why I suggest that people run their own test and see with their own eyes.
If you want to buy faster RAM because it makes you happy, that is perfectly fine. There is nothing wrong with that approach at all. If however you are buying faster than 1600 MHz. RAM for typical desktop use OTHER than for an APU system, then you won't see any significant, i.e. tangible system performance gains. It's been documented by many reputable sources yet some folks still refuse to believe reality.
And yes as noted by you, me and many others, tighter memory timings also do not make any tangible gains in system performance with DDR3 RAM with frequencies of 1333 MHz. or more.
As far as RAM pricing, if you can get 1866 MHz. for the same price as 1600 MHz. then go for it. $5. more for 1866 MHz. is an option, though you still won't get $5.00 more performance, but paying $30-$50 more for 2x 4 GB. of RAM is just throwing money away. You'd be WAY farther ahead to use that $30-$50 on a faster CPU or GPU, but it's your dime so you get to spend in whatever way makes you happy. :)