• Happy holidays, folks! Thanks to each and every one of you for being part of the Tom's Hardware community!

Rumor: Intel 9-Series Chipsets Won't Feature SATA-Express

Status
Not open for further replies.
"Intel is encouraging manufacturers to implement the technology by themselves"

So they would need to implement there own chips and such? I won't begin to speculate why they wouldn't add it on but I am assuming at first this is only going to be on the higher end boards, however I don'tt hink thats really a bad thing per say as most people who NEED the extra bandwidth of sata express is going to be using a higher end board anyhow. (Or atleast has the money to afford one.)
 
Not much point in adding higher-speed IO to Intel's IO hub until Intel upgrades the DMI bus between the CPU and IO hub. Otherwise, things will simply bottleneck there unless motherboard manufacturers start "borrowing" lanes from the CPU's PCIe 3.0 controller.
 
I care, southernshark. One of the biggest bottlenecks in most modern day desktop computers is the hard drive. and with solid state drives getting faster exponentially as time goes by it is almost certain that the current SATA interface will be saturated very soon (See http://www.anandtech.com/show/7520/lsi-announces-sandforce-sf3700-sata-and-pcie-in-one-silicon). In fact SATA-e is one of the biggest reason for me to upgrade from my current build. Without SATA-e I will have a lot less reason to upgrade from my current build.
 


Yeah yeah, but it doesn't need to be an on chip solution. And on most computers it won't matter. Where it will matter is on higher end machines where it doesn't need to be on the chip.

 
I'm not exactly in the market for a new system this go around in the first place, but I still find this a little surprising and unfortunate. It is not so much that there is a particular 'need' for SATA-e so soon after the release of SATA3 (It would be like the update from ATA-100 to ATA-133 with higher burst speed, but generally the same performance level), but 'need' is not why high-end enthusiasts upgrade; It is all about having something better, and squeezing out every last bit of performance possible from your system.

I mean let's face it, a 2 year old i5 or i7 is going to be about as good as a new chip, and because the older chips can be OC'd higher, it really is a performance wash, and is expected to be with next years chips as well. The main motivating factor to upgrade is for chipset, peripheral, and standards support. I moved from z68 to z77 for the sake of TRIM via RAID support. Many others have upgraded for more USB3 or SATA3 ports, while others have upgraded for PCIe3, or even faster native memory support. But it is always an upgrade for these connectivity options rather than a boost in raw CPU performance. If you need extra CPU performance then you simply have to upgrade to a faster platform (LGA2011 i7 or Xenon) rather than a newer chip of the same tier. I fear this will be a theme for the future as well.

But in a world where connectivity is the motivating factor for upgrades I find it very surprising that Intel would delay the release of such a hotly anticipated connectivity standard. It makes me wonder if Intel knows something that we don't know about yet.
 
I guess I'm sticking with my X79 for a while longer. Still no reason to upgrade. There wasn't much of a reason to upgrade to my current chip over my previous Core i7 920 other than the fact the SATA ports on my P6T quit working. Intel just isn't innovating, and it is all because AMD can no longer compete. The damage Intel wrought on AMD back in the early Athlon days (denying motherboard manufacturers access to the 440BX chipset if they made an Athlon motherboard) laid the foundation to what we're facing today. I doubt we'll see any real progress in PC CPUs for another decade, at least. It's going to take a competitor to push Intel into actually innovating again. I doubt AMD will ever be that competitor again.
 

Intel is still using the DMI bus which provides only 20Gbps between the CPU and chipset for SATA, USB, PCIe 2.0 and all other stuff either built in the chipset or attached to it.

No point in implementing higher-speed peripheral connectivity when the middle-bus lacks the bandwidth necessary to properly support it.
 
So what's the point in switching to Z97?
-SATA Express: a good feature I must say, but it won't be included based on the rumor.
-USB 3.1: There will always be somebody who needs more speed. However, this won't be include too.
-DDR4 RAM: Obviously NOT.
-PCIE 4.0: ...are you kidding...
-VRM wil be moved back to mainboard: New feature or just a good-old feature that Intel has messed it up recently?
-----
Basically EVERY GOOD FEATURE you can think of will be included in Sky Lake (or later) then why should one buy Boardwell (unless he/she has a lot of money any doesn't know what to do)?
Intel, do you have anything to say or convince us?
 
So what's the point in switching to Z97?
-SATA Express: a good feature I must say, but it won't be included based on the rumor.
-USB 3.1: There will always be somebody who needs more speed. However, this won't be include too.
-DDR4 RAM: Obviously NOT.
-PCIE 4.0: ...are you kidding...
-VRM wil be moved back to mainboard: New feature or just a good-old feature that Intel has messed it up recently?
-----
Basically EVERY GOOD FEATURE you can think of will be included in Sky Lake (or later) then why should one buy Boardwell (unless he/she has a lot of money any doesn't know what to do)?
Intel, do you have anything to say or convince us?
 
with non-volatile storage (like ssd) getting faster and faster, maybe even future SATA spec doesn't cut it anymore. maybe they should be treated as DRAM, and be given dedicated bus to the CPU.
SATA Express is just a temporary solution, that Intel might be skipping over?
 
I couldn't care less about the extra bandwidth of SATA-E, but I want that new low latency PCIe. About 1-2 magnitudes difference. I'm sure that will help with random IO performance.

Won't know for sure until benchmarks come out, and is also dependent on controllers.
 
hey this is great for me in 6 months i can get haswell and a z79 mobo for cheap as an upgrade. also does anyone know when the 800 series will come for nvidia or are they going to expand the already stretched out gk110 and the 700 series?
 

The DMI bus between the CPU and chipset does not have enough bandwidth to comfortably support USB3.1, SATA-Express and other IO upgrades so those won't happen until Intel upgrades it.

HDMI2 and other video output upgrades would require a much more powerful IGP to back them up if you want anything beyond 2D so this is probably wishful thinking. Even if they do put it in, you might not be able to use it until Broadwell-K and by the time Broadwell-K comes out, Skylake might be just around the corner: on Intel's latest server roadmap, Broadwell-E/EP/EX and Skylake-E/EP/EX are still on-schedule at least in terms of calendar year.
 
And Intel wonders why desktop sales are slipping? The only innovating is being done on phones and tablets.

We need some more cool technology on the desktop to make it worth an upgrade. In most cases 4 or 5 year old machines are fully adequate.
 

Are they really innovating? For the most part, they are simply catching up with PC specs: more RAM, more RAM channels, more MLC flash, more cores, higher clocks, bigger IGP, etc. It is easy to "innovate" when you start from so far behind due to severe power draw, space, weight and cost constraints. Phones and tablets will slow down once the bulk of their catching up is done and at the current rate of nearly doubling overall performance each year through a combination of extra cores, improved IPC, improved (re)compilers, improved OS and increased clocks, there are 3-4 years of that left.

What would a hypothetical N8v3 be like? 12-16GB DDR4 RAM on a 192-256bits bus, 3.5GHz 8-core CPU, IGP probably in a similar performance class as the R7-270, 10MP camera, 1600p screen, the usual Android sensor complement and that's pretty much it. Mere evolution just like PCs - nothing particularly innovative.
 
Status
Not open for further replies.