Idf 2011

[Ao1]

Looks like there will be a few good technical sessions on SSD. Hope they are available as podcasts at some stage.

SSDS001 Optimizing Solid-State Drive (SSD) Performance for Data Center Applications
SSDS002 Multi Level Cell Based Solid-State Drives: Meeting the Demands for High Endurance
SSDS003 The Business Implications of Solid-State Drives (SSDs) – Proof Points and Case Studies
SSDS004 PCI Express* Solid-State Drives: Coming to Enterprise and Client Environments
SSDS005 The Next Frontier in SSD Performance
SSDL001 Hands-on Lab: Intel® Solid-State Drives: Tools and Optimizations for Client and Enterprise
SSDC001 Poster Chat: SSDs and Mainstream Awareness
SSDL001R Repeat: Hands-on Lab: Intel® Solid-State Drives: Tools and Optimizations for Client and Enterprise

[Anvil]

Lets hope the sessions will be available in some way.

Most sessions looks interesting, SSDS002 could be very interesting.

BTW, IDF starts tomorrow. (13-15 September)

[alfaunits]

SSDL001 - wish I went for this one.

[Computurd]

Man its my favorite time of year They at least let out PDFs last year to the presentations, so we could learn a little bit
i still miss when they livestreamed it and hosted the video on 2008...
oh well will still be great, cant wait

[Computurd]

Vexs005

itle: Build more immersive games with intel® ssds and 2nd generation intel® core™ f...
Length: 50 min
experience level: Introductory
abstract: Optimizing games to exploit the capabilities of the latest platform technology allows game developers to create games featuring innovations that result in greater immersion. Spend your engineering time making the game better instead of working around storage delays, and unleash performance for more physics and ai intensive games. With this idea in mind, intel consulted with zombie studios on the creation of fundamentally new features for blacklight: Retribution, a highly immersive aaa first person shooter. The new features are enabled by exploiting the performance of intel® ssds, and blacklight: Retribution exemplifies the content quality made possible by designing for the latest platform ingredients.

Topics in this session include:
• blacklight: Retribution project overview, new features and demo
• hear from project lead jared gerritzen of zombie studios about his vision for future content made possible by new and emerging storage technologies

[zads]

I'll be there..
Hopefully I'll win an Intel SSD again this year ;D

[Computurd]

some of the SSD PDFs are up
https://intel.wingateweb.com/us11/scheduler/catalog.do

[Computurd]

CHSW001 Client PC Responsiveness

this one is looking good ties into SSDs and smart connect, of course

[Ao1]

PDF's are up for SSDS001 to SSDS004.

AHCI to NVM Express
http://www.nvmexpress.org/

[Wendy]

I wonder if there will be any mention of why X79 SATA3 was cut from 10 ports to 2, and why the dedicated x4 PCIe storage lanes have been disabled on SB-E.

[Computurd]

That is an excellent question wendy, it really is a buzzkill some of the SB-E info that is coming down the pipe. I have been waiting a long time for an upgrade, so its between the Dozer and SB-E, and storage is definitely gonna be a huge part of this for me. At least the PCH and forty pcie lanes will be present on the SB_E...small consolation i guess. But the gimped PCIE is why i skipped 1155/6

@ A01 excellent link! reading now

[Ao1]

^ Interesting that Intel chose to use DRAM to show how fast the NVMe interface will be. It looks like the low hanging fruit for storage system improvements is now focused on interfaces and driver systems, although Intel also talk about storage technologies that may in the future have sub microsecond latencies.

Untitled copy.jpg


123.png

[Anvil]

I wonder if there will be any mention of why X79 SATA3 was cut from 10 ports to 2, and why the dedicated x4 PCIe storage lanes have been disabled on SB-E.

From what I´ve read there will be up to 6 6Gbs ports on the X79, 6 ports will still be awesome.

[Wendy]

From what I´ve read there will be up to 6 6Gbs ports on the X79, 6 ports will still be awesome.

I had heard 6 SATA 3 ports about a month ago, but according to Tom's, it's now 2.
http://www.tomshardware.com/reviews/...ce,3026-3.html

[Anvil]

A pity if this is true, as there have been several misinterpretations/misunderstandings I guess I'll wait for some more official info.

Quite a few of the new boards have been shown having 10-14 ports (GB X79 UD5/UD7 @legitreview), that does of course not mean that they are all chipset ports.

edit:

There is an update to the review at toms

"Update: After this piece went up, several credible sources confirmed that Sandy Bridge-E's PCIe actually is 8 GT/s-capable; Intel just hasn't gone through the validation process yet. From what I'm told, there will be a second batch of processors in '12 that follows what we see at launch, and those will receive the official third-gen blessing. Until then, it'll have to be up to motherboard vendors to claim (or not) PCIe 3.0 compliance."

The potential lack of PCIe 3.0 was most likely the thing that made them think there would be only 2 ports, still, this is not official.

[Wendy]

A pity if this is true, as there have been several misinterpretations/misunderstandings I guess I'll wait for some more official info.

Quite a few of the new boards have been shown having 10-14 ports (GB X79 UD5/UD7 @legitreview), that does of course not mean that they are all chipset ports.

edit:

There is an update to the review at toms

"Update: After this piece went up, several credible sources confirmed that Sandy Bridge-E's PCIe actually is 8 GT/s-capable; Intel just hasn't gone through the validation process yet. From what I'm told, there will be a second batch of processors in '12 that follows what we see at launch, and those will receive the official third-gen blessing. Until then, it'll have to be up to motherboard vendors to claim (or not) PCIe 3.0 compliance."

The potential lack of PCIe 3.0 was most likely the thing that made them think there would be only 2 ports, still, this is not official.

From page two in the article that you link to.

The GIGABYTE GA-X79-UD7 features the same orange and black color scheme used on the X58A-OC. The GA-X79-UD7 has four PCIe X16 slots and three PCIe x1 slots. The GA-X79-UD7 has fourteen SATA ports, but we know eight of them have to be removed as Intel removed support for all those in the chipset.

This is such a pity, so I'll hold onto my Z68.

[Anvil]

Clipped from the first page from that same article I linked
"If you are like me and though that the 12 SATA ports on the GA-X79-UD3 was a bit on the excessive side. Then you will no doubt think the same about the 14 SATA ports on the GIGABYTE GA-X79-UD5 are equally so. If I were a betting man, I would suspect that the four black SATA ports as well as the pair of white SATA ports on the end are SATA III 6Gbps"

So, they clearly are guessing, at best.

Another link that shows 6x 6Gb/s ports Xbitlabs

Too much speculations but unless it ends up having more chipset-based storage ports than the Z68/P67 then there is not much point in moving from the SB to the SB-E, unless one needs better raid-card compatibility.
(which has proven to be hit or miss on SB)

[Wendy]

I guess we will have to wait and see what actually appears on X79, when the boards become available.

[Computurd]

the lanes...the lanes...more pcie lanes! great for cards

[Ao1]

Is this going to be a SF based drive?

"The Cherryville SSD will be based on 25nm multi-level cell (MLC) NAND memory and, according to Troy Winslow, Marketing Manager of Intel's NAND Solutions Group, will be "the highest performing SSD in the marketplace." "The Cherryville product will be a no compromise solution," said Winslow. "It will have better sequential and better random than any prior SSD from Intel and will lead the industry."

Toms HW states:
Initial performance information on "Cherryville" shows sequential and random access performance figures are expected to be up to 530 MB/s read, 490 MB/s write sequential performance; and 40,000 IOPS reads and 45,000 IOPS writes random performance.

Those IOPS don’t seem to fit an SF drive or the claim of better random performance that any prior ssd.

Untitled.jpg

[bluestang]

I hope not! But if so then maybe Intel can kick SF into getting their crap together.

[Computurd]

there have been reports on fudzilla that Intel is releasing a SF ssd, and not only that, but they are looking at buying them.

[One_Hertz]

It would be epic if Intel bought sandforce. The quality of SF SSDs would rise substantially.

[Christopher]

I don't think that it's out of the realm of possibility for Intel to purchase SandForce -- it's not like they can't shake some change out of the couch.

SandForce controllers with Intel's firmware and validation would make for an interesting combination.

[Computurd]

Yup, interesting stuff. A big thing here for intel is the patent portfolio of SandForce, that alone is enough of a catalyst for purchase. Too bad SF is private, i would own a little chunk right now LOL, they are gonna sell BIG.
as long as they drop that durawrite BS this will be great!