Wednesday, January 9th 2013

ASUS Unveils World's First PCI-Express 3.0 Motherboard for AMD Processors

ASUS did the unthinkable yet simple, by innovating the first AMD platform motherboard that features PCI-Express 3.0 x16 slots, the Sabertooth 990FX/GEN3.0 R2.0. Long naming aside, the board provides you a couple of gen 3.0 slots by using PLX-made 48-lane PCI-Express Gen 3.0 bridge chip.

While the board features four PCI-Express x16 slots, only two similarly-colored slots can be used at a time, of which two are PCI-Express 2.0 x16, wired to the 990FX northbridge, and two slots being x16/NC or x8/x8-capable, being wired to a PLX 48-lane PCI-Express Gen 3.0 switch, which in turn takes two PCI-Express 2.0 x16 links from the northbridge.

Apart from this unique feature the socket AM3+ Sabertooth 990FX/GEN3.0 R2.0 features AMD SB950 southbridge, eight SATA 6 Gb/s ports, two eSATA 6 Gb/s, 8-channel HD audio, six USB 3.0 ports, and a zesty ASUS-exclusive feature-set. The new board could be released to the market very soon.
Add your own comment

43 Comments on ASUS Unveils World's First PCI-Express 3.0 Motherboard for AMD Processors

#1
cadaveca
My name is Dave
So why are there these users with SB showing PCIe 3.0? Nearly all are using boards without bridges, too. I've seen both ASRock and Biostar with that now in the past few months.

That's the weirdness that gets me. I understand all the stuff about switches.

What made sense to me si that PCIe 3.0 encoding was used, but like you say, SB shouldn't be able. Yet...
Posted on Reply
#2
neliz
MSI Rep
by: cadaveca
So why are there these users with SB showing PCIe 3.0? Nearly all are using boards without bridges, too. I've seen both ASRock and Biostar with that now in the past few months.

That's the weirdness that gets me. I understand all the stuff about switches.

What made sense to me si that PCIe 3.0 encoding was used, but like you say, SB shouldn't be able. Yet...
Are those are software readings? I can't even trust those for measuring voltages.
Posted on Reply
#3
cadaveca
My name is Dave
Honestly, I dunno. I just try to keep an ear to the ground on all things motherboard, and this has been an issue that's popping more often now, and I have no idea why. Not just here on TPU, either.

I mean, we can generally say, that if IVB is used, no matter the board, PCIe 3.0 is possible, link width is questionable though, and dependant on switches used. No problem there, of course, very straight forward.

Using PLX PEX8747, on PCIe 2.0, giving PCIe 3.0, does add some boost to multi-GPU, even though to the CPU is still just PCIe 2.0. So this ASUS board, to me, makes sense for multi GPU users.

Those PCIe 3.0 cards on this board, they'll show up as PCIe 3.0, I am almost willing to bet.

Are they really running 3.0? I guess the PLX PEX8747 makes this possible...or does it?

:roll: products like this one don't make this any easier. :roll:
Posted on Reply
#4
fusionblu
Hi everyone this motherboard is nice that it is the only other alternative to have to run PCI-E 3.0 opposed to fully shelling out for either a new Ivybridge system which isn't a great improvement over Sandybridge in terms of processing power alone or a Sandybridge-E system which is generally pricy (although prices are coming down now) and fairly inefficient for the normal tasks that most desktop users do (gaming included, but there probably is a significant improvement though).

I can see there is some sort of argument here. Put simply a normal Z68 or Z77 won't run Gen 3 without a Ivybridge CPU, that is simply fact. There is only one motherboard which is an exception to this rule and it is a ECS Z77H2-AX which has an additional chipset which allows Gen 3 with the use of a Sandybridge CPU; however, the motherboard in question is expensive and hard to get and you are looking to pay around £280 for it on ebay. There is a review of this particular motherboard on here for that.

Also to add, yes this particular ECS motherboard does make use of a PLX PEX8487 chipset.

Here is the review I mentioned: http://www.techpowerup.com/reviews/ECS/Z77H2-AX_review/13.html

There are other Z77 motherboards with the PLX PEX8487 chipset
Posted on Reply
#5
cadaveca
My name is Dave
by: fusionblu
Also to add, yes this particular ECS motherboard does make use of a PLX PEX8487 chipset.

Here is the review I mentioned: http://www.techpowerup.com/reviews/ECS/Z77H2-AX_review/13.html
Yes, I wrote that review. There are boards from every brand with PLX PEX8747 on Z77 that'll do that, the last review I did, the Z77X-UP7, is just the same.
Posted on Reply
#6
Jizzler
by: fusionblu
Hi everyone this motherboard is nice that it is the only other alternative to have to run PCI-E 3.0 opposed to fully shelling out for either a new Ivybridge system which isn't a great improvement over Sandybridge in terms of processing power alone or a Sandybridge-E system which is generally pricy (although prices are coming down now) and fairly inefficient for the normal tasks that most desktop users do (gaming included, but there probably is a significant improvement though).
I can't think of any situation that would make this the preferred choice, is there one?

The current Sabertooth 990FX is around $180. Due to the PCIe switch, the new model is going to cost more and it's questionable at this point if there will be anything gained by it's use. Shelling out less for a better performing Ivy Bridge system is looking really good right now.
Posted on Reply
#7
fusionblu
by: Jizzler
I can't think of any situation that would make this the preferred choice, is there one?

The current Sabertooth 990FX is around $180. Due to the PCIe switch, the new model is going to cost more and it's questionable at this point if there will be anything gained by it's use. Shelling out less for a better performing Ivy Bridge system is looking really good right now.
That is true and maybe only a AMD diehard user would go for this as their preferred choice, but the idea is that it is an option that will be available and that Intel isn't the only choice if you want PCI-E 3.0.
Posted on Reply
#8
Bp_968
Question pertaining to this discussion

I noticed a MSI hardware guy on here and had a question similar to this discussion. I've been looking at a device that will (passively) switch a single x16 slot to two x8 slots using a flexible cable and a PCB with two x16 slots on it. It requires support for pci-e bifurcation but doesn't explain if thats a chipset, MB, or CPU feature or if it requires the BIOS to allow it (or if it just automatically splits the first port into whatever the other two ports need).

Any ideas? It would make it possible to turn my mATX board into something with a bit more expansion possibilities with lower bandwidth cards (stuff thats plenty happy with a couple lanes or less).
Posted on Reply
#9
RejZoR
I wish they'd offer this in a microATX form with the same color scheme. And for god sake without the prehistoric PCI slot... It seems there are no proper high end boards for AMD processors in a microATX form, which i need for my tiny box...
Posted on Reply
#10
Prima.Vera
Why so much hate for the old PCI slot?? I have an old Leadtek tuner on PCI, I also have an USB 2.0 adapter for PCI with 4 ports, also an old X-FI PCI sound card, still better than any integrated...
Posted on Reply
#11
seronx
http://ark.intel.com/compare/65693,65694,65692,65690,66168
http://ark.intel.com/compare/63696,70845

I don't know, I'm not going to bother with benchmarks I'll be back with more information with Ivy Bridge being PCI-E 2.0. Ivy Bridge and Sandy Bridge-E have the same thing they have PCI-E 2.0 that can run at 8 GB/s. Sandy Bridge-E has 40 lanes of PCI-E 2.0(@3.0 Speed) while Ivy Bridge has 16 lanes of PCI-E 2.0(@3.0 Speed) going through one PLX switch or several PLX switches. (variations: 16x/16x/16x or 8x/8x/8x)

True native PCI-E 3.0 won't be supported till Haswell, hopefully. PCI-E 3.0 isn't just speed.
Posted on Reply
#12
cdawall
where the hell are my stars
by: seronx
I'll just wait for the new LGA platforms next year from AMD with native PCI-E 3.0.
AMD isn't releasing anything, but it's server chips on LGA. Everything in normal desktop is sticking with a socket. :laugh:
Posted on Reply
#13
seronx
by: cdawall
AMD isn't releasing anything, but it's server chips on LGA. Everything in normal desktop is sticking with a socket. :laugh:
AM3+ is the last socket for consumer/performance desktop of this generation.

The next generation is either Socket GC36 or this unknown socket which is based on Socket F and Socket G3.
Posted on Reply
#14
cdawall
where the hell are my stars
by: seronx
AM3+ is the last socket for consumer/performance desktop of this generation.

The next generation is either Socket GC36 or this unknown socket which is based on Socket F and Socket G3.
Not according to the last batch of roadmaps that put steamroller on a new socket supporting ddr4 with am3+ backwards compatibility.
Posted on Reply
#15
seronx
by: cdawall
Not according to the last batch of roadmaps that put steamroller on a new socket supporting ddr4 with am3+ backwards compatibility.
Show me this roadmap.

I've only seen the APU one and the server one.
Posted on Reply
#16
cdawall
where the hell are my stars
by: seronx
Show me this roadmap.

I've only seen the APU one and the server one.
Google it and remember amd servers are not amd desktops. G34 and C36 along with LGA1207 are server sockets something AMD will not release desktop parts onto. G3 was abandoned in 2008, GC36 does not exist nor is it on any roadmap. So wtf are you talking about.
Posted on Reply
#17
seronx
by: cdawall
Google it and remember amd servers are not amd desktops. G34 and C36 along with LGA1207 are server sockets something AMD will not release desktop parts onto. G3 was abandoned in 2008, GC36 does not exist nor is it on any roadmap. So wtf are you talking about.
I did google it.

28-nm FMx = Bolton <-- Socket (Mainstream)
28-nm LGA1 = Riverside <-- Socket (Performance)
28-nm LGA2 = Riverside <-- Socket GC36 (Server)
^(Only the FMx socket is compatible with the previous chipset since the infrastructure for LGA1/2 is substantially different than Socket C32/G34 and AM3+)
Posted on Reply
#18
cdawall
where the hell are my stars
by: seronx
I did google it.

28-nm FMx = Bolton <-- Socket (Mainstream)
28-nm LGA1 = Riverside <-- Socket (Performance)
28-nm LGA2 = Riverside <-- Socket GC36 (Server)
^(Only the FMx socket is compatible with the previous chipset since the infrastructure for LGA1/2 is substantially different than Socket C32/G34 and AM3+)
I have seen nothing about LGA1/LGA2 and LGA1 is an Intel Xeon socket.



Server side still shows a split 2P and 4P side through 2015 along with AM3+ through mid 2014.

Like I said before
What wasn't known up until recently, however, was that AMD intends to stick by the AM3+ socket for at least one more major processor iteration, which is highly likely to be Steamroller. One of the compelling reasons to stand by AMD is typically cost, knowing that an investment in an AM3+ motherboard will most likely survive an extra processor generation is a compelling cash saver and, in fact, a portion of users running AMD computers are likely to already have AM3+ motherboards, as the socket has been available since mid-2011.
They also mention AMD moving to a unified socket after FM2+ and AM3+ run there course.

source
Posted on Reply
Add your own comment