Bookmark and Share


After Advanced Micro Devices cancelled code-named Komodo processor in favour of Vishera, the company’s high-end desktop platform lost any ability to obtain support for PCI Express 3.0 till at least late 2014. However, Asustek Computer has managed to make impossible possible with its Sabertooth 990FX/Gen3 R2.0 mainboard, which enables PCI Express 3.0 support on AMD FX platform.

Asus Sabertooth 990FX/Gen3 R2.0 motherboard is based on AMD 990FX/SB950 chipset and is designed for microprocessors in AM3+ packaging, such as AMD FX, AMD Phenom II and other. The mainboard is made for performance enthusiasts, hence it has robust 8+2 digital phase power design, highly-reliable components, advanced overclocking capabilities, as well as support for 4-way multi-GPU graphics technologies, including AMD CrossFireX and Nvidia SLI.

Since current-generation AMD FX microprocessors do not support internal PCI Express connectivity, while AMD 990FX only supports PCI Express 2.0, Asustek had to install a special PCI Express 3.0 bridge chip made by PLX (according to TechPowerUp web-site). The 48-lane PCIe 3.0 PLX chip lets two PCI Express x16 slots to operate at PCIe 3.0 data-rates, which may boost performance of high-end multi-GPU configurations.

While Asustek Computer has already published all the specifications and features of the Asus Sabertooth 990FX/Gen3 R2.0 mainboard, it is unknown when exactly the mainboard is supposed to enter the market. Since the board was demonstrated at the Consumer Electronics Show this week, it is logical to expect mass availability this quarter. The regular Sabertooth 990FX motherboard costs $179 online at the moment, the version with PCI Express 3.0 will probably cost $200 - $210 when available.

AMD’s high-end FX platforms will gain official support for PCI Express 3.0 along with the release of next-generation high-performance desktop and server central processing units, which is expected to happen in late 2014.

Asus Sabertooth 990FX/Gen3 R2.0 mainboard features extensive coolind mechanisms onboard

Tags: ASUS, AMD, FX, PCI Express, Vishera, Piledriver, 32nm


Comments currently: 17
Discussion started: 01/11/13 02:15:00 PM
Latest comment: 12/19/15 05:20:11 PM
Expand all threads | Collapse all threads


In ASUS we trust, they are about to add support even for the intel only thunderbolt on their AMD hardware. There were nothing to stop them from adding and a PCIe 3.

Last week I installed Sabertooth R2.0 on my system, unbelievable piece of hardware.

I wanted to hang it on the wall to be able to see this motherboard all the time.

Keep it up ASUS!
6 0 [Posted by: nitro912gr  | Date: 01/11/13 02:15:00 PM]

So they're converting 32 PCIe 2.0 lanes into 32 PCIe 3.0 lanes and this magically improves performance? Yeah right. You're still limited by the PCIe 2.0 speeds that feed the 3.0 plx chip this is marketing for suckers and stupid people
2 0 [Posted by: sollord  | Date: 01/11/13 02:41:44 PM]
- collapse thread

maybe they used more pcie2.0 lanes per pcie3 lane . but the gpus can communicate whit each other at the PCIe 3.0 speed.
1 0 [Posted by: massau  | Date: 01/12/13 10:33:53 AM]
Exact. But here's the marketing goal to return extra profit with this product. It's pretty much the same as those oveclocking championships that puts a bug in Joe's mind to keep upgrading his system at least every 12 month for sake of greedy OEMs.

Asus did the same with their first and only s775 X38 mainboard which featured first ddr3 support on thta chipset but DIDNT use 2x16 lanes coming from MCH-X38 but JUST ONE of available lines was split up to two x8 pcie 2.0 lanes to support CFX (Well they were two x16 pcie 1.1 lines after PLX arbitration connected to ONE x16 PCIe 2.0 coming from X38)... it's good ... Much better than regular P45 which was first mainstream PCIe 2.0 intel platform and having only x4 PCI 1.1 coming from SB if you'd like to use CFX on that board.

Duh, it's strange marketing to cripple AMD this way ... well i know it's enhancement -- a marketing enhancement.
0 0 [Posted by: OmegaHuman  | Date: 01/14/13 04:55:00 PM]

I'm ashamed to publicly admit this, but I don't really understand how can two PCIe x16 3.0 slots work at 3.0 speeds if the chip that manages them is routed through 2.0 PCIe lanes.

I would gladly accept any mocking in exchange for even a patronizing, but logical and easy-to-understand explanation.

On the other hand, the guys at TechPowerUp might actually find themselves in an error as the bridge chip from PLX (if it's from PLX) is not using PCIe 2.0 lanes to connect to the "northbridge."

After all ... there is no northbridge in a IMC enhanced CPU.

I know for a fact that PLX has been a member of HyperTransport Consortium for a decade now and there is a possibility that the PLX controller actually uses a HT connection to the CPU and that would indeed provide the necessary bandwidth to enjoy real PCIe 3.0 speeds.
1 0 [Posted by: East17  | Date: 01/11/13 04:59:33 PM]
- collapse thread

Well it's about bandwidth, two PCIe 2.0 x16 ports combined will obviously bring much more bandwidth, so from what I understand it basically translates your PCIe 3.0 x16 to PCIe 2.0 x32
1 0 [Posted by: Medallish  | Date: 01/11/13 07:14:03 PM]
First one that obviously fall onto that cheap trick. Do you see marketing is doing miraculous things?
(Hint: You CAN NOT improve data trnsfer rate over the maximum 990FX performane, just you couldnt do the SAMRE TRICK if you use SandyBridge and if some *kreyzi mobo maker* like Asus would put PLX chip on Z77 boards ... It wont do the trick that SB provide PCIe 3.0 x16(+x4) data rate)
0 0 [Posted by: OmegaHuman  | Date: 01/14/13 04:58:04 PM]
You shouldn't be ashamed as this is just dirty cheap marketing trick.

Simply, overall system performance is always capped (or as overclocker boyz would say : bottl'nekd) with maximum NB performance -- in this case with 990FX chipset
1 0 [Posted by: OmegaHuman  | Date: 01/14/13 04:56:32 PM]

Anyone who with a good working knowledge about PC operation already knows that PCIe 3.0 is of zero performance value for any desktop PC - at the moment. Maybe in a few years there may be a need for it, but right now it's a solution to a problem that doesn't exist, though you can't convince the technically challenged because if a technology is advertised as better, then it must be better, even when it isn't - at the moment. The minor useable features of PCIe 3.0 right now, make no real difference in GPU performance as the x16 lanes are not saturated most of the time.

Asus is guilty of bogus promotion of PCIe 3.0 as a performance enhancement and that is why they developed this hack-patch of questionable merit. Asus knows there are a lot of technically challenged consumers who will buy Asus mobos because they falsely believe that PCIe 3.0 will magically provide a performance advantage, which it doesn't at the moment or probably in the next few years even. This is just another disingenuous marketing ploy IMO, until proven to actually deliver improved system performance in real applications, not synthetic benches. I can't wait to see the reviews with real apps not synthetic benches...
1 1 [Posted by: beenthere  | Date: 01/11/13 07:24:35 PM]
- collapse thread

Zero performance boost? You are forgetting about GPGPU (CUDA and OpenCL). Those do benefit from PCIe 3.0. Beside the bandwidth, there is significantly less overhead from PCIe 3.0

At this point average user who games won't realize it, but performance user who crunch number all the time will have speedup from PCIe 3.0.

However it is pointless in releasing PCIe 3.0 on AM3+ platform, since HT 3.0 still bottlenecks PCIe performance. At some point AMD will have to dump HT (and AM3+ platform), if AMD wants to be successful on HSA.
2 1 [Posted by: trumpet-205  | Date: 01/12/13 09:15:22 AM]
I'm not forgetting anything. PCIe 3.0 does not offer anything in actual PC system performance for current desktop PC use, as the PCI lanes are not saturated for more than a few seconds at a time. In theory PCIe is great but until there is enough actual data throughput to use the attributes of PCIe, it's all marketing gimmicks.
2 2 [Posted by: beenthere  | Date: 01/12/13 11:12:41 AM]
Depends on how many highend gpus you are running at the same time two GT 680's or 2 7970's can fully saturate pci 2.0e lanes and 3 of them could over saturate it. as toms hardware showed, but that's highend stuff and many people don't have 2 highend gpus and esp 3 highend gpus like that. So in generel avg use you are right, but fro really highend to extreme highend use pc 3.0 is the way to go.
0 0 [Posted by: SteelCity1981  | Date: 01/14/13 11:48:04 AM]
And really extreme computer user would become a donkey after dealing with this Asus motherboard. And everybody once again will say how AMD sucks just because som cheap greedy manufacturer designed some alchemist dream and call it a motherboard
0 1 [Posted by: OmegaHuman  | Date: 01/14/13 05:08:12 PM]

The CPU to Northbridge bandwidth is still heavily bottlenecked by HyperTransport 3.0, which can only handle about 10 to 11 PCI-E 3.0 lanes at full load. There should be only a marginal improvement, if any. It's mainly marketing gimmick.
1 0 [Posted by: r3velin  | Date: 01/11/13 10:19:43 PM]
- collapse thread

But you still neglecting that PCIe 2.0 lanes in 990FX cant provide any extra bandwidth just because someone clamp some external chip on those lines ... It simply doesnt matter how much bandwidth HT3.1 could provide
0 0 [Posted by: OmegaHuman  | Date: 01/14/13 05:11:00 PM]

It looks like in single GPU mode, the PLX switch might be able to route 32 PCIE 2.0 lanes into the single PCIE 3.0 x16 lane slot to give it full bandwidth.
1 1 [Posted by: efok  | Date: 01/14/13 11:54:13 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture