Patrick Schur

@_rogame @RajaJagadeesan Both are mobile reference platforms. Mayan is for FP7 and Lilac for FP7r2. The desktop reference platform is called Splinter.

@RajaJagadeesan My guess would be U series vs H series. Nothing confirmed atm

AMD Lilac-RMB

@ExecuFix Pssst... 6 nm and RDNA2 🤫

AMD Mendocino | MDN-A0 Zen 2 | socket FT6 CPUID 0x8A0F00 🤔

@wild_cracks Where is Turin and Bergamo coming from and what about Trento and Badami? 🙃

@patrickschur_ Oh, so this is the Zen 5 Epyc IOD? Naples = Zeppelin Rome = Starship Milan = Genesis Peak Genoa = Rolling Stones Turin = Storm Peak Or maybe this is Bergamo?

AMD Storm Peak 🧐

@Arcturus_PRD CPU

@patrickschur_ CPU or APU?

@VideoCardz Yep 😋

@patrickschur_ Peak, sounds like Threadrippa

@davideneco25320 Tell me something new... 🙃

AMD : PCIE 5.0 pour serveur, pas pour le mainstream Intel : PCIE 5.0 pour le mainstream mais uniquement les GPU

@OrinLupe I was thinking about something like zoc (Zen OC) and noc (Navi OC). 😅

@patrickschur_ Something simple like Ryzen Overclocker. and when you want to shorten the name call it ROC (Ryzen OC). It all makes sense

Just to let you know, I'm currently working on an open source overclocking and monitoring tool for Linux and with support for all Zen based CPUs. 😋

@ExecuFix @Nexusunny What do you mean? It's a normal TDP value as it's defined by AMD.

@Nexusunny @patrickschur_ That 170W isn't a normal TDP value. 120W is the max for the normal SKUs

The exact TDP numbers for Raphael are 65, 95, 105, 120 and 170 W.

@aschilling It's the TDP. PPT can go beyond the TDP. PPT limit = TDP * 135 %

@patrickschur_ Is this what AMD defines as TDP or could there be any changes and your numbers are what PPT (Package Power Tracking) is today?

@_rogame I'm pretty sure that we will not see an odd number of cores, but it's definitely possible that some CPUs will only have one CCD. But don't quote me on that. 😀

@patrickschur_ Can we expect some odd number of cores/ccd ?

Raphael 🥰 - up to 16 Cores - up to 170 W - 5 nm

@Wolf9466 @OrionMatsuura Help is always appreciated.

@patrickschur_ @OrionMatsuura Want some help? >:3

Also thanks to @tekwendell for supporting me and giving me remote access to his Epyc servers. 🥳

@OrionMatsuura That's also something I'm looking at. 👀 But it's too early to talk about it. I'll concentrate on CPUs first.

@patrickschur_ Hell yes. now if only a tool for Radeon GPU Overclocking was a thing too

@peresuslog @MauriQHD @davideneco25320 twitter.com/patrickschur_/…

@patrickschur_ @MauriQHD @davideneco25320 What's the difference between FP7 and FP7r2?

AMD Ryzen Embedded - V3000 Zen 3 (6 nm) - FP7r2 - up to 8 Cores / 16 Threads - 20x PCIe 4.0 lanes (8x dGPU) - 4x DDR5-4800 (ECC) - two 10G ethernet PHYs - 2x USB 4.0 - 15-30 W and 35-54 W models - up to 12 CUs (RDNA2)

@Locuza_ @peresuslog @MauriQHD @davideneco25320 According to my information only 4x 32b are possible.

@patrickschur_ @peresuslog @MauriQHD @davideneco25320 Can/will AMD expose all 128-Bit under FP7r2, meaning an Octa-Channel (8x16b) LPDDR5 memory system?