Amd Officially Confirms Fresh Next-Gen Zen 6 CPU Details

You say that, but DDR6 will double the memory bandwidth over DDR5. This means modern systems will go beyond 200GB/s memory bandwidth just for the CPU alone.

I rocked my Haswell i5 until last year when I built a brand new machine around the 9800x3d. Along the way I upgraded it from 8gb of ram to 32gb, got a gen 1 pcie3 NVME, and went through successive hand-me-down GPUs starting from a GeForce 770 to the RTX 2070 it has now.

In fact my wife is still rocking that machine - although her gaming needs are much less equipment intense than mine. After a small refurb I gave it (new case, new air cooler, new PSU) - I expect it to last another 5 years for her.

I really wish I would've bought 192G when it was less than a few thousand dollars!

Had been running a coffee lake refresh 4 core for several years and as interested as I was in new platforms, especially AM5, the work of replacing motherboard never felt worth it. Now with the ram wars heating up, I just committed to that buy picking up a used top-end 8 for coffee lake for $50 to cut a few seconds off my vulkan shader compiles with minimal replacement effort.

Depends wildly on what you're doing.

I'm a gamer, often playing games that need a BEEFY CPU, like MS Flight Simulator. My upgrade from an i9-9900K to a Ryzen 9800X3D was noticeable.

My understanding is they’re using the same process time for cpus and gpus so they may just be able to reallocate it for datacenter gpus. Sure they’re behind but some of the AI companies have already made deals with them as they just want compute, any compute. So I think the effect might be less than some hope for

A socket change does not correspond with a RAM change. Intel has had like 10 sockets in DDR4 including their newest socket.

They already do the latter with X3D.

You won’t be able to add RAM to the die itself there no room on the interposer really.

Yeah the only use of the large bandwith in Apple Silicon is for the GPU. I'm always amazed by the fanboys who keep hyping this trope.

Even when feeding all cores, the max bandwith used by the CPU is less than 200GB/s, in fact it is quite comparable to Intel/AMD CPUs and even less than their high-end ones (x86 still rules on the multi-core front in any case).

I actually see this as a weakness of Apple Silicon, because it doesn't scale that well. It's basically the problem of their Ultra chip: doesn't allow doubling of the compute and doesn't allow faster RAM bandwith, you only get higher RAM capacity in exchange for slower GPU compute.

They just scaled up their mobile architecture and it has its limit.

It’s really not that simple, the unpopulated memory slots will cause havoc with signal integrity 4 slot boards already suffer from this.

You are also overestimating how much room there is on the interposer.

As someone with a 9950x3d with direct die cooling setup I can tell you there is no room.

The M.2 form factor isn't that conducive to having lots of them, since they're on the board and need large connectors and physical standoffs. They're also a pain in the ass to install because they lie flat, close to the board, so you're likely to have to remove a bunch of shit to get to them. This is why I've never cared about and mostly hated every "tool-less" M.2 latching mechanism cooked up by the motherboard manufacturers: I already have a screwdriver because I needed to remove my GPU and my ethernet card and the stupid motherboard "armor" to even get at the damn slots.

SATA was a cabling nightmare, sure, but cables let you relocate bulk somewhere else in the case, so you can bunch all the connectors up on the board.

Frankly, given that most advertised M.2 speeds are not sustained or even hit most of the time, I could deal with some slower speeds due to cable length if it meant I could mount my SSDs anywhere but underneath my triple slot GPU.

There are add-in cards with PCIe switch chips that will let you put a large number of drives into a single PCIe slot.

> Alas, most mobos only seem to hand out lanes in multiples of 4.

AFAIK, the cpu lanes can't be broken up beyond x4; it's a limitation of the pci-e root complex. The Promontory 21 chipset that is mainstream for AM5 does two more x4 and four choose sata or pci-e x1. I don't think you can bifurcate those x4s, but you might be able to aggregate two or four of the x1s. And you can daisy chain a second Prom21 chipset to net one more x4 and another 4 x1.

Of course, it's pretty typical for a motherboard to use some of those lanes for onboard network and what nots. Nobody sells a bare minimum board with an x16 slot, two cpu based x4 slots, two chipset x4 slots, and four chipset x1 slots and no onboard perhipherals, only the USB from the cpu and chipset. Or if they do, it's not sold in US stores anyway.

If pci-e switches weren't so expensive, you might see boards with more slots behind a switch (which the chipsets kind of are, but...)

Bifurcation is done on the CPU.

Asus 5.0 Hyper M.2 with yield 4x4x4x4 in a board that is setup for bifurcation of PCIE slot 1 like that.

I have run 4 GPUs this way and it works very well.