ASUS Prepares P7P55D Premium, Equips SATA III

[btarunr]

Here's our ASUS P55 motherboard of the day: the P7P55D Premium. Positioned in the high-end segment of socket LGA-1156 motherboards, this board boasts of support for SATA III 6 Gbps. Associated with the SATA III controller, is also an interesting feature. To begin with, the CPU is powered by a digital PWM circuit with "32+3" phases. Sure there are that many chokes, but we do suspect choke redundancy here. The CPU is wired to four DDR3 DIMM slots to support dual-channel memory.

Unlike the P7P55D Deluxe, this board makes do with two PCI-E 2.0 x16 slots, each of which arrange to 8-lanes PCI-E when both are populated. Other expansion slots include two PCI-E x1 and two PCI. Connectivity options include two gigabit Ethernet controllers, 8 channel audio, and a number of USB and Firewire ports. From the looks of it, there's also a place-holder for the NEC µPD720200 USB 3.0 controller, and its marked blue-coloured USB ports, though it's MIA.



The P55 PCH provides six SATA II ports, and an additional Marvell 88SE9123 controller gives out two next-generation SATA III ports. ASUS used an interesting (and expensive) method of making sure the controller works to its full potential. A PCI-Express 2.0 switch was added to the board, that connects to all the four PCI-E 1.1 lanes the P55 PCH provides, and gives out 12 PCI-E 2.0 lanes, with support for three ports in all. One of these ports, configured as PCI-E 2.0 x1, connects to the Marvell 88SE9123 (which is a PCI-E 2.0 x1 device) to make sure it is provided with a 500 MB/s interconnect, reducing its bottleneck compared to most other motherboards that simply connect one of the four PCI-E 1.1 lanes from the PCH to it. We think the two PCI-E x1 slots make for the rest of the PLX PCI-E switch's two ports. It is believed that the P7P55D Premium will launch in the next couple of months, if not being part of the first LGA-1156 boards from ASUS.



View at TechPowerUp Main Site

 

[Axaion]

At last, PCI-e x1 slots placed properly. and a good 2 space between the x16's

 

[legends84]

sata III o'ready? whats the different between sata II?

 

[Mussels]

sata III o'ready? whats the different between sata II?

600MB/s

Again, this has failed like the asrock.


Even at PCI-E 2.0, its still only 250MB/s each way, making it capped below SATA II speeds.
(that said, its a hell of a lot better than boards capped at 125MB/s each way, using PCI-E 1.1 lanes!)

 

[btarunr]

Once again, PCI-E 2.0 is 500 MB/s per direction.

 

[Mussels]

Once again, PCI-E 2.0 is 500 MB/s per direction.

i found a link and quoted it, stating otherwise.

I'm confident that my numbers are right, i have a PCI-E 1x SATA II card with two E-sata ports, and no matter what i do, the total bandwidth in one direction is always 120MB/s or under. Two sata II drives (one on each port) copying to two internal drives, its either 90-110MB/s from one external to one internal, or 50-60MB/s each.

 

[btarunr]

i found a link and quoted it, stating otherwise.

I contradicted you using the same link.

 

[Mussels]

I contradicted you using the same link.

please see other thread.

 

[btarunr]

please see other thread.

After you.

 

[Sihastru]

SATA III for current HDDs and SSDs is like PCIe 3.0 for current video cards... which doesn't mean it's bad to have, it just means that for now it's just a marketing gimmick. For now I am more excited about USB 3.0 then any of these other unnecessary advancements.

[...]I have a PCI-E 1x SATA II card with two E-sata ports, and no matter what i do, the total bandwidth in one direction is always 120MB/s or under. Two sata II drives (one on each port) copying to two internal drives, its either 90-110MB/s from one external to one internal, or 50-60MB/s each.

This doesn't look like a good test. There's always overhead and there's always eSATA cards that won't do exactly what it says on the tin. In this case it looks like a limitation of the mechanical HDDs you used. Most 7200RPM drives today will go up to 120MB/s. The low end is 60MB/s depending on where the data is placed on the platter.

 

[Mussels]

so, we're updating our stance on this.

PCI-E 1.x: 250MB/s per direction
PCI-E 2.0 500MB/s per direction.

Hooray!

it at least gives you enough performance for one sata-III device to almost max out the lane.

 

[Sihastru]

Hooray! It at least gives you enough performance for one sata-III device to almost max out the lane.

Yes, this is a "small" problem their marketing enhanced "Perfect SATA6Gb/s Solution" didn't account for

 

[Mussels]

Yes, this is a "small" problem their marketing enhanced "Perfect SATA6Gb/s Solution" didn't account for

lol true, its hardly perfect.

"better than what the other guys did!" doesnt work so well on the box.

 

[lemonadesoda]

2 paper cups and a piece of string.

 

[Mussels]

hey look, the picture covers their asses! nice one!

while its a copout, i do agree that 500MB/s is indeed enough for 2 current SSD's. (but then again, that would work out fine on sata II anyway)

 

[MrMilli]

so, we're updating our stance on this.

PCI-E 1.x: 250MB/s per direction
PCI-E 2.0 500MB/s per direction.

Hooray!

it at least gives you enough performance for one sata-III device to almost max out the lane.

And you believe PCI-E 2.0 is 100% efficient?
Well Sata rev3 isn't either, so i guess that evens things out.

It's not a perfect solution because that PLX chip has added latency.
That said, the fact that the P55 chipset doesn't have PCI-E 2.0 is a serious backdraw.

 

[AsRock]

Not much for the colours they used although like the layout of it.

Any one have a idea what the micro switch is\does next to the 24pin ?.

 

[JrRacinFan]

Any one have a idea what the micro switch is\does next to the 24pin ?.

Looks like a pwr on switch. "??"


I absolutely love the layout of this board. Think I may be watching this board for my i5 upgrade.

 

[Fitseries3]

i'll just wait till ICH11R comes out and sata3 is fully supported.

their "perfect solution" is quite a shit move.

native support is almost always better.

 

[Sihastru]

And you believe PCI-E 2.0 is 100% efficient?
Well Sata rev3 isn't either, so i guess that evens things out.

The 500MB/s theoretical bandwidth already accounts for 20% overhead in the PCIe 2.0 specification. The actual bandwidth is 20% higher, but it uses an error correction protocol.

The PCIe 3.0 specification eliminates the 20% overhead and it's rated at 1000MB/s per lane.

I don't have exact data for SATA, but SMART and other things like that will take some of the bandwidth away. But you shouldn't be so afraid of SATA2 (or 3Gb/s). 3Gb/s is 3000Mb/s that is 375MB/s. Considering a 20% overhead (larger then I think it actually is for SATA II) we still come up at 300MB/s. And usually you get 6 native SATA ports on a MB, so that's 1800MB/s just for hard drives (or SSDs).

SATA III promises 6Gb/s, so that's 750MB/s. Six SATA III ports should have a total of 4500 MB/s... that's 4.5 GB/s bandwidth... that's a lot for a current mainstream motherboard.

 

[btarunr]

i'll just wait till ICH11R comes out and sata3 is fully supported.

their "perfect solution" is quite a shit move.

native support is almost always better.

If that "ICH11R" continues using DMI (which it will, if it has to work with X58), then it's worse, because DMI is 2 GB/s max. Even two~three SATA III devices that top out its bandwidth (such as high-density SSDs) will choke it.

 

[Fitseries3]

then they should use a reliable source for implementation of sata 3 such as LSI.

 

[3870x2]

ummm...we havent reached the capabilities of SATA 1 yet have we? please forgive my ignorance, the way I understand it, even the best SSD gets 160 or so MBps, which doesnt stress SATA1?

unless, my alternative theory about how SATA is that it is PER CONTROLLER, so you have 1.5/3/6Gbps, so you might have 4 or 6 sata slots, but they share that bandwidth? it would make sense then. maybe I should visit wikipedia.org. . . brb. . .

 

[[I.R.A]_FBi]

bite vs bytes

 

[Mussels]

ummm...we havent reached the capabilities of SATA 1 yet have we? please forgive my ignorance, the way I understand it, even the best SSD gets 160 or so MBps, which doesnt stress SATA1?

unless, my alternative theory about how SATA is that it is PER CONTROLLER, so you have 1.5/3/6Gbps, so you might have 4 or 6 sata slots, but they share that bandwidth? it would make sense then. maybe I should visit wikipedia.org. . . brb. . .

Gbits vs Gbytes, to the controllers.


The best SSDs do around 220MB/s - and they're getting faster rapidly. Remember that this tech takes years to spread around in systems, better to release the ports before the drives so that people can actually use it when they need it.