PCI Express 2.0: Scalable Interconnect Technology, TNG
by Kris Boughton on January 5, 2008 2:00 AM EST- Posted in
- CPUs
PCI Express Backwards Compatibility
PCI Express is a layered protocol composed of a three distinct partitions: physical (PHY), data link layer (DLL), and transaction layer (TL). This modular approach is what allowed the fundamental change in the physical layer data rate from 2.5GT/s to 5.0GT/s to go unnoticed by the upper layers. The PCI-E bus speed remains unchanged at 100MHz; the only feature that changed is the rate at which data is transferred across the board. This suggests that there is significant signal manipulation required on both the transmit and receive ends of the pipe before data is available for use. In the end, the change is akin to selecting a higher multiplier for a CPU: although the processor operates at an increased frequency, it continues to communicate with the interface component at the same rate. This concept will help us to introduce the concept of a "PCI-E Multiplier." This number is static and cannot be changed (except by the motherboard, which occurs during the training process). PCI Express 3.0 should further increase the PCI-E Multiplier to 80x, which will bring the base link frequency very near the maximum theoretical switching rate for copper (~10Gbps).
As we mentioned before, installing a PCI Express 1.x card in a PCI Express 2.0 compliant slot will result in PCI Express 1.x speeds. The same goes for installing a PCI Express 2.0 card in a PCI Express 1.x compliant slot. In every case, the system will operate at the lowest common speed with the understanding that all PCI-E 2.0 devices must be engineered with the ability to run at legacy PCI-E 1.x speeds.
Unfortunately, there have been some reports of new PCI-E 2.0 graphics cards refusing to POST (Power On Self-Test) in motherboards containing chipsets without PCI Express 2.0 support. We can assure you this is not by design as the PCI-SIG PCI Express 2.0 Specification is very clear on the issue - PCI Express 2.0 is backward-compatible with PCI Express 1.x in every case. With this in mind, and knowing that some PCI Express 1.x motherboards have no problems running the new graphics cards while others do, we have no choice but to blame the board in the case of no video. Unfortunately, this makes it difficult to determine if problems await during the upgrade process; you must first consult with others that use the same video card/motherboard combination. Thankfully the cases of incompatibility seem to be few and far between.
Facebook
Twitter
RSS
21 Comments
View All Comments
PhotoPrint - Sunday, May 11, 2008 - link
????? ?? ????????? ??????
????
?????
????? ??????
nubie - Monday, January 14, 2008 - link
I would like to point out that since the link auto-negotiates you can plug x16 cards into x8, x4, x2, and x1. The problem of physical connection is easily solved. I have done this two ways, one by cutting the back out of the motherboard connector, (seen here: http://picasaweb.google.com/nubie07/PCIEX1">http://picasaweb.google.com/nubie07/PCIEX1 ), and also by cutting the connector off of the video card down to x1 (sorry, no pics of this online). I did this to get 3 cards and 6 monitors on a standard (non SLI) motherboard. You can also purchase stand-offs from x16 to x8-x1, or modify a x1-x1 standoff (or "wearout" adaptor) to allow the x16 card to plug in.The throughput was more than enough, depending on your video cards on-board ram it can even play newer games fine. The utter lack of multi-head display support in DirectX and most games is just mind-boggling. Tell me why PC games won't allow multi-player, while consoles do?
cheburashka - Monday, January 7, 2008 - link
"and there is no obvious reason as to why 2x8 CrossFire on a P965 chipset should not work"It has a single LTSSM thus it can not be split into multiple ports.
cheburashka - Monday, January 7, 2008 - link
"and there is no obvious reason as to why 2x8 CrossFire on a P965 chipset should not work"It only has a single LTSSM thus it can not be split into multiple ports.
fredsky - Monday, January 7, 2008 - link
sorry guys to be not as enthusiast...http://www.fudzilla.com/index.php?option=com_conte...">http://www.fudzilla.com/index.php?optio...mp;task=...
there ARE a lot a issues here, especially with RAID cards pluged into PCIe 2 slots. LSI, 3Ware, Areca and so on.
anand can you make some tests ?
I read that Gigabyte GA-X38-DQ6 is compatible with Areca at least.
regards
fredsky
decalpha - Monday, January 7, 2008 - link
I am not sure who is the guilty party but my new and shiny 8800GT refuses to POST. And if you search the user forums it's clear that most of the problems are faced by socket 939 systems with nvidia chipset. In the end it's the user who suffers.Comdrpopnfresh - Monday, January 7, 2008 - link
Does PCI-E also increase the available current supplied to the card by the slot? Doubling seems to be a theme here... Maybe from 75-150 Watts? I skimmed, so I apologize if it was written or already mentioned...kjboughton - Monday, January 7, 2008 - link
Although we didn't discuss this in the article, I can certainly answer the question: no. The slot still supplies up to a maximum of 75W per the specification; however, the PCI Express Card Electromechanical interface spec will allow for an additional 150W power delivery via external power cables for a total of 225W. Anything above this number is technically out of specification.AndyHui - Sunday, January 6, 2008 - link
Didn't seem all that long ago when I wrote the first PCI Express article here on AT.... but looking back, that was 2003.Good article.... but I thought the official abbreviation was PCIe, not PCI-E?
saratoga - Saturday, January 5, 2008 - link
"PCI Express 3.0 should further increase the PCI-E Multiplier to 80x, which will bring the base link frequency very near the maximum theoretical switching rate for copper (~10Gbps)."This will be quite a surprise to the Ethernet people who can do 100 Gbit/s over 2 ethernet twisted pairs on their prototype systems! 500% of the theoretical maximum for copper is pretty good.
Theres no theoretical maximum for copper since in theory the SNR can be infinite, and thus you can keep coming up with better codes. Theres a practical limit, set by just how high you can get the SNR in a real circuit, but thats also unbelievably high. The real limit for a PC is how much power you're willing to commit to your increasingly complicated transmission system.