"On the motherboard, this means we need the M.2 (formerly NGFF) standard to be prevalent: we have seen it so far on the ASUS ROG line with mPCIe Combo II enabled motherboards, where it uses a PCIe 2.0 x1 lane from the PCH to connect to the system."
Wait, the M.2 port was slower than standard SATA III? What's the point in that?
The PCIe advantage comes in the form of less overhead, usually resulting in higher overall throughput and IOPS - at least in the reviews I've seen so far. Obviously, PCIe 2.0 x4+ and PCIe 3.0 x1+ will give us MOAR SPEAD!
Or they can do like ASRock did on this one and give PCIe 2 x2, which is 10 Gbps. Not significantly faster, but still better. I assume it is going to migrate to PCIe 3.0 x1 or x2 before too much longer.
I think it happened because they can do that on AMD board without hurting the PCI-e lanes that would go to the GPU... "This allows 990FX to be a platform for this type of expansion without hurting any GPU bandwidth limitations or reducing the number of extra ports on the board. For Intel products, the lanes have to be taken from the chipset (up to 8x PCIe 2.0 lanes) which could arguably introduce a delay due to the extra routing, but until we get two similar systems side-by-side for comparison, we cannot hold any data to that claim." I would love to see a surprise FX processor... I want so badly a new FX processor!!!
"ASRock are now preparing to release an AMD AM3+ motherboard based on the 990FX chipset with access to PCIe 2.0 x2, direct from the CPU, which should be suitable for 10 Gb/s peak speed (this around 1 GB/s) in SATA Express mode and SATA 6Gb/s in SATA mode."
The main PCI-e controller isn't on-die with the CPU but rather the chipset. This still has a discrete both and south bridges.
With the North Bridge, rather than the South Bridge. The latter being referred to more as the Chipset in modern terminology, and the former being associated closer with the CPU (hence the Intel and AMD APU method of combining them). But you are technically correct :)
I understand that part of that is some dude's nickname, but seriously did the marketing people think "Hmmm, how much more childish can this naming get?" when they decided to add Killer onto that?
What's next? the Fatal1ty Killer Turbo Xtreme edition?
YES!!! It hath begun....hope this prods asus and everyone else to start putting m.2 (ngff) slots on their mobos too. Better yet, two m.2 slots that can be raided. Sure there are some pcie cards with m.2 slots but I'd rather have them native...and raided. I can't wait til LSI/SandForce finish tweaking (and qualifying) the SF-3700 firmware. I'm pretty sure Intel is playing a big role with fw engineering and qualification. SF let Intel overhaul the uber proprietary fw for the SF-2281 when it was all screwed up...Intel knocked it out of the park on the first try. Their fw for the 520's was rock solid perfect to the point they haven't had to do any updates at all. LSI/SF + Intel = WIN!
I'm having a bit of a hard time understanding why the Fatal1ty branding is still around. Maybe I just haven't been paying attention, but hasn't it been a while since the dude was relevant?
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
22 Comments
Back to Article
Gigaplex - Thursday, December 12, 2013 - link
"On the motherboard, this means we need the M.2 (formerly NGFF) standard to be prevalent: we have seen it so far on the ASUS ROG line with mPCIe Combo II enabled motherboards, where it uses a PCIe 2.0 x1 lane from the PCH to connect to the system."Wait, the M.2 port was slower than standard SATA III? What's the point in that?
nathanddrews - Thursday, December 12, 2013 - link
The PCIe advantage comes in the form of less overhead, usually resulting in higher overall throughput and IOPS - at least in the reviews I've seen so far. Obviously, PCIe 2.0 x4+ and PCIe 3.0 x1+ will give us MOAR SPEAD!A5 - Thursday, December 12, 2013 - link
Or they can do like ASRock did on this one and give PCIe 2 x2, which is 10 Gbps. Not significantly faster, but still better. I assume it is going to migrate to PCIe 3.0 x1 or x2 before too much longer.Torashin - Thursday, December 12, 2013 - link
Why bother releasing this an an AM3+ board? Do they know something we don't? (Surprise steamroller FX...)Powerbaconzitos - Thursday, December 12, 2013 - link
I think it happened because they can do that on AMD board without hurting the PCI-e lanes that would go to the GPU..."This allows 990FX to be a platform for this type of expansion without hurting any GPU bandwidth limitations or reducing the number of extra ports on the board. For Intel products, the lanes have to be taken from the chipset (up to 8x PCIe 2.0 lanes) which could arguably introduce a delay due to the extra routing, but until we get two similar systems side-by-side for comparison, we cannot hold any data to that claim."
I would love to see a surprise FX processor... I want so badly a new FX processor!!!
anubis44 - Friday, December 13, 2013 - link
Ask and ye shall receive...Notmyusualid - Sunday, December 22, 2013 - link
Ha ha.Kevin G - Thursday, December 12, 2013 - link
"ASRock are now preparing to release an AMD AM3+ motherboard based on the 990FX chipset with access to PCIe 2.0 x2, direct from the CPU, which should be suitable for 10 Gb/s peak speed (this around 1 GB/s) in SATA Express mode and SATA 6Gb/s in SATA mode."The main PCI-e controller isn't on-die with the CPU but rather the chipset. This still has a discrete both and south bridges.
IanCutress - Thursday, December 12, 2013 - link
With the North Bridge, rather than the South Bridge. The latter being referred to more as the Chipset in modern terminology, and the former being associated closer with the CPU (hence the Intel and AMD APU method of combining them). But you are technically correct :)Kevin G - Thursday, December 12, 2013 - link
*discrete north and south bridges.Powerbaconzitos - Thursday, December 12, 2013 - link
The image has three "espress"... it should be "express", right? (not being ironic, i'm just checking if I'm missing something! =D)IanCutress - Thursday, December 12, 2013 - link
You're right - it's meant to be express. It's the AMD produced chipset layout I used back during 990FX coverage. Ha, I never noticed that.steven75 - Thursday, December 12, 2013 - link
"Fatal1ty Killer"?I understand that part of that is some dude's nickname, but seriously did the marketing people think "Hmmm, how much more childish can this naming get?" when they decided to add Killer onto that?
What's next? the Fatal1ty Killer Turbo Xtreme edition?
B-Unit1701 - Thursday, December 12, 2013 - link
In this case, 'Killer' is a reference to the Killer NIC product which is integrated.bobbozzo - Thursday, December 12, 2013 - link
Hi,It is unclear from the pictures/article where the m.2 or SATA Express connector is on the board.
Thanks!
WT - Thursday, December 12, 2013 - link
I'm guessing the M2 is located about 1.5" above the CMOS battery, just to the right of the white lettering (Killer 2200 silkscreening)DanNeely - Thursday, December 12, 2013 - link
The text says it's between the first two PCIe slots, and you can see a connector between the Killer 2200 label and the south bridge heatsink.Movieman420 - Thursday, December 12, 2013 - link
YES!!! It hath begun....hope this prods asus and everyone else to start putting m.2 (ngff) slots on their mobos too. Better yet, two m.2 slots that can be raided. Sure there are some pcie cards with m.2 slots but I'd rather have them native...and raided. I can't wait til LSI/SandForce finish tweaking (and qualifying) the SF-3700 firmware. I'm pretty sure Intel is playing a big role with fw engineering and qualification. SF let Intel overhaul the uber proprietary fw for the SF-2281 when it was all screwed up...Intel knocked it out of the park on the first try. Their fw for the 520's was rock solid perfect to the point they haven't had to do any updates at all. LSI/SF + Intel = WIN!fluxtatic - Wednesday, December 18, 2013 - link
I'm having a bit of a hard time understanding why the Fatal1ty branding is still around. Maybe I just haven't been paying attention, but hasn't it been a while since the dude was relevant?5thaccount - Sunday, December 22, 2013 - link
Why is noone capable of making 900 series chipset mATX motherboards?R3MF - Monday, December 23, 2013 - link
good question - i drop cash on one in a heartbeat.