If there was really flexible ways to bi/quad-furicate the 16 PCIe 4 lanes to what peripherals like 10Gbit Ethernet can actually handle, it could do wonders in workstation space.
Wish it was as flexible as the base architecture of Inifinity Fabric actually seems to be, but don't know how OEMs could achieve that with the hard wires they need into the motherboards.
Never understood the obsession with 10G. Unless you're moving VMs to other machines constantly or editing raw 4k over the network it's really overkill even for the enthusiast consumer. Your speed is 4GBps, until recently with NVME it was almost impossible to feed the network card.
To properly schedule network traffic to minimize jitter and loss, you need about 25% more bandwidth than provisioned. If you expect to transfer up to 1Gb/s of traffic, you should really had 1.25Gb/s of bandwidth if you don't want networking hickups.
I am less concerned with 10Gb than 2.5/5.0Gb, but those two new protocols are based on 10Gb, and pretty much everyone who makes nbase-t, supports 2.5/5/10. May as well just say "10".
What are you actually talking about?! The Ryzen 3000 CPUs provide 24 PCIe 4.0 lanes. 16 go to the GPU(s) in a 1x16 or 2x8 configuration dependant on the MoBo. Then there are 4 lanes to a dedicated M.2 NVME slot (which is a PCIe SSD) and 4 lanes to the X570 (and B550 etc.) chipset, which then further allocates lanes to SATA/USB/M.2(PCIe/SATA)/Controllers etc. dependant on the actual MoBo.
My question is, what generation of PCIe lanes come out of the X570 to feed the lower slots? With X370 and X470 it was PCIe 2.0 and the lower slots were starved of lanes. Does X570 give out PCIe 3.0, which means the number will be restricted and the bandwidth wasted, or does it give out twice as many PCIe 2.0 lanes? I hope it is the latter.
Leaked block diagrams showed all the chipset lanes to slots are gen4 pcie. MSI also is using a pcie multiplexer on some boards. So both the top slots can get x16, what I wonder is if its smart enough to mux two x16 3.0 devices into a single x16 4.0 signal back to the cpu. Anyone know more about how pcie multiplexers work?
AMD hasn't had a good chipset since irongate so I'll be interested in hearing more. It's too bad about third party chipsets going away though. Socket 7 had a ton of chipset competition and now we just have to put up with whatever turds Intel or AMD foist on us.
The only third party chipset I ever used was nForce4 (Socket AM2) and that was irredeemably dreadful. It contained major flaws in its PCI and SATA implementation that were never/could never be fixed. In fact, I seem to remember that to avoid SATA data corruption you had to kludge the hard drive's own firmware! It should never have been launched.
Oh yeah, I remember the nForce chipsets. Those were some awful stuff. SiS and VIA made competitive products, but that was long before Socket AM2 existed. The field was a lot more broad around and before the time of the Pentium 4.
Irongate on my Abit KG7-Raid was night and day over the KT133 debacle. I know 133A fixed a bunch of things, but I never wanted another VIA chipset again after my KT7-Raid disaster.
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.
28 Comments
Back to Article
abufrejoval - Sunday, May 26, 2019 - link
If there was really flexible ways to bi/quad-furicate the 16 PCIe 4 lanes to what peripherals like 10Gbit Ethernet can actually handle, it could do wonders in workstation space.Wish it was as flexible as the base architecture of Inifinity Fabric actually seems to be, but don't know how OEMs could achieve that with the hard wires they need into the motherboards.
StevenD - Tuesday, May 28, 2019 - link
Never understood the obsession with 10G. Unless you're moving VMs to other machines constantly or editing raw 4k over the network it's really overkill even for the enthusiast consumer.Your speed is 4GBps, until recently with NVME it was almost impossible to feed the network card.
bcronce - Tuesday, May 28, 2019 - link
To properly schedule network traffic to minimize jitter and loss, you need about 25% more bandwidth than provisioned. If you expect to transfer up to 1Gb/s of traffic, you should really had 1.25Gb/s of bandwidth if you don't want networking hickups.I am less concerned with 10Gb than 2.5/5.0Gb, but those two new protocols are based on 10Gb, and pretty much everyone who makes nbase-t, supports 2.5/5/10. May as well just say "10".
Chaitanya - Monday, May 27, 2019 - link
Is that "lightning" m.2 a pcie 4.0 based M.2 slot?ArcadeEngineer - Monday, May 27, 2019 - link
Yes, Matisse/X570 does 16 lane GPU+4 lanes direct CPU-connected M.2+4 lanes chipset.blingon - Monday, May 27, 2019 - link
"as will the four PCIe lanes assigned to the first NVMe M.2 slot"What if you use a PCIe SSD? Plus a GPU?
blingon - Monday, May 27, 2019 - link
...I guess this was just a description of a typical MB layout, let's hope there's some variety.Death666Angel - Monday, May 27, 2019 - link
What are you actually talking about?!The Ryzen 3000 CPUs provide 24 PCIe 4.0 lanes. 16 go to the GPU(s) in a 1x16 or 2x8 configuration dependant on the MoBo. Then there are 4 lanes to a dedicated M.2 NVME slot (which is a PCIe SSD) and 4 lanes to the X570 (and B550 etc.) chipset, which then further allocates lanes to SATA/USB/M.2(PCIe/SATA)/Controllers etc. dependant on the actual MoBo.
John_M - Monday, May 27, 2019 - link
My question is, what generation of PCIe lanes come out of the X570 to feed the lower slots? With X370 and X470 it was PCIe 2.0 and the lower slots were starved of lanes. Does X570 give out PCIe 3.0, which means the number will be restricted and the bandwidth wasted, or does it give out twice as many PCIe 2.0 lanes? I hope it is the latter.12345 - Monday, May 27, 2019 - link
Leaked block diagrams showed all the chipset lanes to slots are gen4 pcie. MSI also is using a pcie multiplexer on some boards. So both the top slots can get x16, what I wonder is if its smart enough to mux two x16 3.0 devices into a single x16 4.0 signal back to the cpu. Anyone know more about how pcie multiplexers work?blingon - Monday, May 27, 2019 - link
I meant AIC form factor. It's a pity a form factor developed for laptops (M.2) is so prevalent on desktop given it's cooling headaches, vs U.2 or AIC.But looks like there are 2x8 motherboards ( aimed at SLI ) which offer the option of cpu lanes to a AIC rather than only M.2. So panic over : )
plonk420 - Monday, May 27, 2019 - link
dat guy staring intently at the Godlikecolonelclaw - Monday, May 27, 2019 - link
dat guy is none other than @gavbon86 the editor of this very web site!plonk420 - Monday, May 27, 2019 - link
ah, i've only mainly seen (some) of the guys in the twitcher panel to the rightjust4U - Monday, May 27, 2019 - link
Does his pic come with the Godlike board? If so.. better have RGB highlights!Ashinjuka - Monday, May 27, 2019 - link
The Godlike should come with a pack-in sticker for that piece of chrome with that very image on it. We love you, Gavin! 😘PeachNCream - Monday, May 27, 2019 - link
AMD hasn't had a good chipset since irongate so I'll be interested in hearing more. It's too bad about third party chipsets going away though. Socket 7 had a ton of chipset competition and now we just have to put up with whatever turds Intel or AMD foist on us.John_M - Monday, May 27, 2019 - link
The only third party chipset I ever used was nForce4 (Socket AM2) and that was irredeemably dreadful. It contained major flaws in its PCI and SATA implementation that were never/could never be fixed. In fact, I seem to remember that to avoid SATA data corruption you had to kludge the hard drive's own firmware! It should never have been launched.PeachNCream - Monday, May 27, 2019 - link
Oh yeah, I remember the nForce chipsets. Those were some awful stuff. SiS and VIA made competitive products, but that was long before Socket AM2 existed. The field was a lot more broad around and before the time of the Pentium 4.just4U - Monday, May 27, 2019 - link
I liked the Nforce chipsets and thought they were a step up over sis/via. Never had any problems with them.Infernus - Monday, May 27, 2019 - link
Irongate on my Abit KG7-Raid was night and day over the KT133 debacle. I know 133A fixed a bunch of things, but I never wanted another VIA chipset again after my KT7-Raid disaster.Assimilator87 - Monday, May 27, 2019 - link
LOL There's a dial next to the onboard power button that goes to 11. Anyone know what it does?just4U - Monday, May 27, 2019 - link
fan maybe?blingon - Tuesday, May 28, 2019 - link
10 triggers a tinny dragon roar. Go to 11 to make the fan spit out actual RGB LEDs.