Will there be Rampage V Formula, Gene and Impact ?

[Orduu]

Will there be Rampage V Formula later or there will be only Extreme also will there be Rampage V Gene and Impact if yes then when ? I am also interested if there will be X99 Deluxe variant with bundled NFC2 , WLC and Thunderbolt 2 like happened with Z97 Deluxe.

[Krista_Wolffe]

@Skunk:

M.2 is pcie 3.0, not 2.0 compatible, hence mPCIE (which isn't mSATA, either)

AFA pcie lane allocation, quite a lot of motherboards work similar to what I proposed anyway; take a look at how the x99-deluxe does it.

WRT networking, 1GbE has to go, and it has to start somewhere... why not here, if we are dreaming 😉 1 gig e is at least 9 gigs too slow. I prefer infiniband, like mentioned in my other post, but realistically, gig e must go!

[Krista_Wolffe]

@kkn: USA, partially to my dismay. 😞

[Skunk]

Krista Wolffe wrote:
M.2 is pcie 3.0, not 2.0 compatible, hence mPCIE (which isn't mSATA, either)

PCI SIG just specifies the differential pairs through the connector and the tolerance needed to be able to handle the signalling rates at the highest supported rate. M.2 can and does have backward support for PCI-e 2.0...for example, the Samsung XP941 used in Apple's computers uses 4 PCI-e 2.0 lanes and the upcoming XP951 uses 4 PCI-e 3.0 lanes. mPCIe only supports up to two lanes of PCI-e 2.0 and wireless technologies are now beginning to exceed that bandwidth. WiGig can offer up to 7 Gbps, although I think they are currently seeing speeds in the 4-5 Gbps range.

Infiniband is nice, but I don't think Mellanox and others making infiniband products are going to move down into the consumer space anytime soon. BTW, while QDR "only" has 40 Gbps, it still has lower latencies than FDR does.

I'm a fan of PCI-e for high speed interfaces, as Thunderbolt has. With remote NVMe M.2 drives, you don't have to push the data through a network stack; you just access it directly. There is also DDIO, which is a pseudo-DMA function, that can allow for some high performance peripherals in the system (I know of the security issues, but you run that risk with any device you plug into a PCI-e connector on your computer).

[jnv11]

Krista Wolffe wrote:
You won't need the networking coming from processor PCI-E lanes; that's usually handled by the chipset, which is connected to the processor on a DMI 2.0 4x link at 20Gb/s. Plenty to handle GigE and WiFi.

2 M2 slots would be nice, but they'd have to be vertical. Probably won't get that, though.

I'd like to see 4 16x PCI-E slots, in some combination of (2-16x,8x), (16x,3- 8x), (16x, 2- 8x, 2- M.2), (2- 16x, 2- M.2), and have *all* running from the CPU, not a crappy pseudo 2.0 8x from the chipset.

If there's room, add 5he crappy 2.0 8x somewhere as an additional mPCIE.

If wishes are being granted: make ram extenders so I can use 8 dimms somehow, or allow for 32gb dimms on a Haswell-e.

KILL the ps/2 port!

2 on board usb3 headers would kick ass.

And as always, I'd recommend a 10GbE, as schizat needs rolling on that front.

I would have to disagree about killing the PS/2 port. First, I read the USB HID device class definition at http://www.usb.org/developers/hidpage/HID1_11.pdf and found out that keyboards by default will interrupt the CPU every half second according to this quote from page 53 from the HID device class definition that I linked to:

The recommended default idle rate (rate when the device is initialized) is 500 milliseconds for keyboards (delay before first repeat rate) and infinity for joysticks and mice.

In this quote, the idle rate is the rate at which an HID device will respond with a report even when nothing is changing. When there is no report to make due to the lack of a change and no expiration of an idle timer, an HID device will respond to a polling request by the USB hardware with a NAK, or negative acknowledgement. The polling is done by the USB controller with no intervention from the CPU except for setting up the polling. Each time the CPU is interrupted by one of these null reports that are sent every half second, it must waste time and power to process it. PS/2 does not interrupt the CPU unless it has an event to deliver, which saves CPU utilization and power. Fortunately, all HID devices other than keyboards are recommended to default to having an infinite idle timer, so they do not waste CPU time or power.
The second reason I prefer PS/2 for the keyboard is security. When USB was defined, there had to be a way for USB mice and keyboards to be able to interface with operating systems that are entirely unaware of USB and only expected the traditional keyboard controller. The solution chosen was to use the System Management Mode (SMM), a benign rootkit that is built into the 386SL, 80486, and all later x86 CPUs according to https://en.wikipedia.org/wiki/System_Management_Mode. It was invented to handle power management for laptops whose operating system (DOS) could not handle power management and had no room or extra battery power for an embedded controller to handle power management for the laptop. When no USB-aware operating system loads, a process called USB Legacy Support converts USB input into legacy input using SMM by having the USB controller trigger the SMM line when either the mouse moves or the keyboard sends input. The SMM code will then translate and load the input into the keyboard controller. The OS then processes the input as if it came from the keyboard controller. When a USB-aware operating system loads and initializes the USB driver, it will tell the USB controllers to quit triggering the SMM line so that the OS will handle future USB mouse and keyboard input. Unfortunately, malware called SMM rootkit that corrupt the built-in SMM code into becoming a malicious rootkit exist, most notably by the NSA in various SMM rootkits like SOUFFLETROUGH, SCHOOLMONTANA, DEITYBOUNCE, and IRONCHEF as seen in the Wikipedia article I linked to. Therefore, an SMM rootkit can easily become an ideal keylogger along with malware that tells the USB controller to start triggering the SMM line each time input from the keyboard or mouse comes in. I would like to not give any SMM rootkit an entrance to my keyboard. I just wish that all motherboard manufacturers would secure their BIOS, firmware, and driver downloads with HTTPS so that performing a man-in-the-middle attack against someone downloading one of these can be assured that these downloads are genuine and not tampered with so that we can ensure that there is no SMM malware in the download.

[Skunk]

I mocked up what a µATX board could look like:



I got tired at the end and didn't want to try to fill everything in. It's not that great of a pic, just a hack job, quick and dirty.

I took the concept of the daughter board and modified it. It isn't possible to use the ones on the Z87 and Z97 boards because of the memory, so I rotated it parallel to the board where it would be mounted on stand-offs and plugged down into the connector below (behind the USB connectors, here). That allows for full 7.1 and a line in while freeing up space on the board for other components. I did forget that a nice feature to have would be an audio-only HDMI (mini and micro work, too) for Dolby TrueHD and DTS-MA (plus a number of other formats coming, like Dolby Atmos, Auro3d, etc.)

I put the M.2 slots on top in slot positions 2 & 4. These would be PCI-e only (no SATA)...and remember that M.2 are intended to be full blown expansion cards, not just for storage. So, you therefore can place two slot PCI-e cards in each slot on the board, and then use the board space underneath them for additional expansion cards or storage, instead of having an unused slot.

There is also the M.2 slot that is vertically oriented (like on the X99-Deluxe) by the rear panel for the wireless card.

The heat sink over the VRMs should be no taller than the tops of the chokes right next to them...the heat pipe will move a lot of the heat to the side. This should keep a nice open airway for cooling to the processor.

Other than that, I think the rest is evident. Just what I think an ideal board could look like...

[FreakinApplePie]

Whoa! Wonder if you could "mock" ROG Rampage V Gene :)?

[kkn]

and for 10gb you need to raid up to get the speed up too.
as of now its the ssd and hdd between each pc that holds it up.
unless you start using raid arays to transfer from pc to pc.
but in the end it would be cheeper to get pcie cards and link them together.
or a fiber card.
and the normal man in the streed cant afford a fiber connector on the mobo it self. let alone the cost of the equipment.

[Krista_Wolffe]

10GbE maxes out around 1GBps, so a good M.2 SSD or a pair of sata 6g SSDs in raid 0 will do this without a problem.

1GbE maxes out around 100MBps, and even a decent platter spinner will do this.

I just went FDR infiniband at home (dual port cards, so 3 nodes in a switchless configuration) has me hauling at ~56Gbps line speed, although since I'm not running E or EP class rigs yet, I'm limited to about 22Gbps in all practicality, as I'm out of pcie bandwidth due to GPUs, IB cards, and SAS adapters.

1GbE needs to go. This is the longest time I can remember where a single drive could max out a commodity network interface. This needs to change, we've been stuck at GigE far too long.

[kkn]

sweden?

[pathfindercod]

My next build will be matx or itx so I hope they do release a smaller version.