Possibly throwing a barebones video output into the chipset? Maybe, but that doesn't require a socket change. And it would be useful for troubleshooting.
I think if AMD did want to offer something like this they would offer a low power small chip that MB companies could install for added functionality and not something they would include in a general chipset.
That's not what these are made for. I am not saying AMD can't do so but I think people tend to confuse dream states with reality. The APU is at its heart a laptop CPU just as a desktop i7 isn't anything other than a laptop CPU.
AMD has enough expertise and experience with their custom division to make what you are asking for. It still isn't a reasonable retail solution which means "value" doesn't play a part. We know by the XBX that this type of CPU serves a purpose. It just isn't in a desktop socket.
to compete with hades canyon nuc ?No because there is no market for it.
No, not just barebone but a powerful one, and maybe calling it TR4+ socket ?
For a new Threadripper (compatible with Threadripper APU) socket could we get Thunderbolt III/IV/V instead of dedicated video outputs? This way it could be also be used for I/O if no display on the system is used or needed.No, not just barebone but a powerful one, and maybe calling it TR4+ socket ?
There are 2 dies epyc it's for embedded market thoughThere can be a market for a TR branded APU in a BGA setup for select vendors. There isn't for the desktop market. The one and only reason AMD would even pause to contemplate one is that there is a server market use for this configuration. That said it's a year almost and we still haven't even seen a 2 Die EPYC. So while the platform is probably malleable enough for AMD to offer a TR APU in the sense that every supporting aspect of TR (the stuff it borrows from the server platform) could exist with a market to sell to, a TR desktop APU would be one of the most pointless products to offer. It eskews everything "HEDT" and it's only realistic purpose would be in the "home made" server market. Besides the fact that AMD isn't selling their hardware with enough price segmentation to make that worthwhile it really screws with the branding of Threadripper in general.
So while 4 channel memory bandwidth on an APU would look fantastic this a realistic hope or goal.
But how mainstream (really) is a CPU with 16C/32T?The mainstream platform won't get more than 24 lanes of PCIe because the vast majority of mainstream users don't need more than that. Adding more lanes increases complexity and costs (more pins in your socket, more traces to route, more layers in your motherboard), making the product less competitive.
1: Mainstream needs APU more than it needs more PCIe Lanes. APU is the heart of mainstream. Socketed mainstream isn't going away,1.) I think eventually the mainstream socket will no longer support APUs*. (This to free up PCIe lanes making the platform more high performance from a storage** standpoint.)
Example: If the current AM4 did not support APUs then the dCPU would be able to use all 32 PCIe lanes.
2.) I think eventually the Threadripper socket will support APUs. (This because there is room under the heatspreader for a rather large iGPU).
Example: The current Threadripper processor packages have 4 CPU dies, but only two of them are active (The other two dies are inactive and could be replaced with a large GPU die).
*All future mainstream APUs will be BGA only.
Yeah, and most mainstream users aren't going to use all of those PCI lanes to begin with. It's also more than Intel offers on their mainstream part (Ryzen has 24, Intel 16 on 8700K), so I'm not even sure what the argument is supposed to be. Unless things have changed recently, I don't think most graphics cards don't gain much from going from 8x to 16x either, and I don't think I've heard of anyone running tri- or quad-GPU setups recently either.1: Mainstream needs APU more than it needs more PCIe Lanes. APU is the heart of mainstream. Socketed mainstream isn't going away.
That could be a possibility. There's rumors that AMD might go to a 6-core CCX design soon. I could easily see them doing that to build their plain CPUs, but keeping a 4-core CCX variant around which would mean that you could get that, a GPU, and some HBM on a package instead of going with two of the larger 6-core CCX modules.2: The next mainstream socket may be slightly larger physically to have room for HBM memory on board the mainstream socket.
|Thread starter||Similar threads||Forum||Replies||Date|
|J||Question Two Intel Desktop CPUs: 2014-era Core i7 vs 2020 (current) Core i3?!||CPUs and Overclocking||9|
|Article 'CPUOverload Project: Testing every x86 desktop processor since 2010' - Anandtech Bench||CPUs and Overclocking||9|
|Question Ryzen Power plan behavior||CPUs and Overclocking||10|
|Question Why don't we have a massive push for ARM chips in desktop computers and laptops?||CPUs and Overclocking||121|
|N||Question 64C/128T is the most we will see on Desktop||CPUs and Overclocking||28|
|Question Two Intel Desktop CPUs: 2014-era Core i7 vs 2020 (current) Core i3?!|
|Article 'CPUOverload Project: Testing every x86 desktop processor since 2010' - Anandtech Bench|
|Question Ryzen Power plan behavior|
|Question Why don't we have a massive push for ARM chips in desktop computers and laptops?|
|Question 64C/128T is the most we will see on Desktop|