Announcing the PCIe® 7.0 Specification: Doubling the Data Rate to 128 GT/s

Topic

New Reply

Announcing the PCIe® 7.0 Specification: Doubling the Data Rate to 128 GT/s for the Next Generation of Computing

..At the PCI-SIG Developers Conference 2022, we celebrated our 30-year anniversary with the announcement of the next evolution of PCIe technology: PCIe 7.0 specification. The forthcoming PCIe 7.0 specification is planned to once again deliver a speed increase in three years, expanding the data rate of the recently released PCIe 6.0 specification to 128 GT/s. The PCIe 7.0 specification is targeted for release to members in 2025.

PCI-SIG technical workgroups will be developing the PCIe 7.0 specification with the following feature goals:

Delivering 128 GT/s raw bit rate and up to 512 GB/s bi-directionally via x16 configuration
Utilizing PAM4 (Pulse Amplitude Modulation with 4 levels) signaling
Focusing on the channel parameters and reach
Continuing to deliver the low-latency and high-reliability targets
Improving power efficiency
Maintaining backwards compatibility with all previous generations of PCIe technology..

Reply | Quote

Replies

great I just can’t wait for 2033.

Just because you don't know where you are going doesn't mean any road will get you there.

Reply | Quote

Having no recollection of what the subject of this conversation is about, and the short-short participants’ comments so far not helping me to know what it is, I’ve looked around and found that these “PCe” things are cards that plug on special slots on the motherboard of a computer and where on them, in turn, one can plug the connectors to peripheral devices. These PCe things make possible the fast transfer of data from the computer bits and pieces on the motherboard where the data take form, to the so-connected peripherals — for whatever purpose these peripherals are supposed to get such data and get it fast.

https://www.tomshardware.com/reviews/pcie-definition,5754.html

Excerpt:

“PCIe (peripheral component interconnect express) is an interface standard for connecting high-speed components. Every desktop PC motherboard … has a number of PCIe slots you can use to add GPUs (opens in new tab) (aka video cards aka graphics cards), RAID cards (opens in new tab), Wi-Fi cards or SSD (opens in new tab) (solid-state drive) add-on cards. The types of PCIe slots available in your PC will depend on the motherboard you buy (opens in new tab).“

Ex-Windows user (Win. 98, XP, 7); since mid-2017 using also macOS. Presently on Monterey 12.15 & sometimes running also Linux (Mint).

MacBook Pro circa mid-2015, 15" display, with 16GB 1600 GHz DDR3 RAM, 1 TB SSD, a Haswell architecture Intel CPU with 4 Cores and 8 Threads model i7-4870HQ @ 2.50GHz.
Intel Iris Pro GPU with Built-in Bus, VRAM 1.5 GB, Display 2880 x 1800 Retina, 24-Bit color.
macOS Monterey; browsers: Waterfox "Current", Vivaldi and (now and then) Chrome; security apps. Intego AV

1 user thanked author for this post.

Alex5723

Reply | Quote

@OscarCP : You have never used a desktop PC ?

Reply | Quote

I first got involved in SSDs on the PCIe bus back in 2012.

Prior to that year, it seems as if very few Prosumers realized the long-term implications of 2x and sometimes 4x PCIe video cards installed in a single motherboard’s x16 expansion slots.

So, I started commenting at Internet User forums about exploiting all of that raw bandwidth to connect SSDs directly to the PCIe bus.

This seemed like a natural evolution, particularly because Intel’s motherboard chipsets were somewhat STUCK with a narrow DMI bus that did I/O to slower peripherals like rotating HDDs.

Despite suggestions to make it wider and to up the clock rate, Intel was very slow to make any significant changes to their proprietary DMI bus.

So, in 2012 my paper was accepted by the Storage Developer Conference.

One of the suggestions we stressed in that paper was to “sync storage devices to the chipset” by using the same clock speed and the same frame layout.

That one suggestion ushered in an 8G PCIe clock and a new 128b/130b “jumbo frame”, both of which were standard in PCIe version 3.0 .

And, of course M.2 NVMe SSDs adopted the same specs, thus “syncing” NVMe storage to the latest PCIe 3.0 chipsets.

The rest is history:  all kinds of solid-state storage started showing up in PCIe expansion slots.

And, of course, x4 M.2 NVMe sockets became an industry standard in almost every new motherboard.

My favorite STILL is the 4×4 add-in cards that support 4 x M.2 NVMe SSDs.

4@4 = 16 is a truly elegant engineering symmetry!

Intel tried to make their implementation proprietary with a “dongle” that was DOA.

Other implementations required a “bifurcation” feature in the motherboard’s BIOS.  This feature added to the cost of motherboards and was usually found on expensive server motherboards.

Other companies like Highpoint developed their own on-card controllers which circumvented bifurcation and enabled bootable RAID-0 arrays with a single 4×4 add-in card like their model SSD7103.

Another parallel development has been the growing availability of multi-core CPUs.  This evolution ended up making otherwise idle CPUs cores available to do Input-Output processing that had previously been done by dedicated IOPs installed on add-in cards.

That multi-core evolution quickly rendered such dedicated IOPs almost entirely obsolete.

10 years ago, I doubt I would have even believed, back then, that the clock speed of the future PCIe bus would continue to increase by leaps and bounds.

But, as history has shown, it continues to DOUBLE with every new generation.

And, solid-state storage continues to exploit the two key benefits of syncing storage to the chipset and the 128b/130b “jumbo frame”.

TRULY AMAZING!!

Reply | Quote