- May 2, 2017
- 3,904 (2.66/day)
- Norwegian, currently in Lund, Sweden
|Processor||AMD Ryzen 5 1600X|
|Cooling||Custom CPU+GPU water loop|
|Memory||16GB G.Skill TridentZ DDR4-3200 C16|
|Video Card(s)||AMD R9 Fury X|
|Storage||500GB 960 Evo (OS ++), 500GB 850 Evo (Games)|
|Power Supply||EVGA Supernova G2 750W|
|Keyboard||Lenovo Compact Keyboard with Trackpoint|
|Software||Windows 10 Pro|
Because there isn't a single motherboard on the market that would be able to power such an APU. And how on earth would they fit a huge APU + HBM2 or GDDR6 onto the AM4 package? Or are you asking for a socket SP3 APU? 'Cause those boards don't have video outputs.Another glorified VEGA APU generation again
another mediocre APU
where's the RDNA APU
and most importantly, AMD already making PS5 & Serie X APUs
why not compete in the PC Market with similar offering
and APU with powerful GPU, maybe 28 - 40 CUs variants?, 6-8 cores with integrated HBM2 or GDDR6
HBM2 would make more sense because it can fit in the CPU PCB, yet more expensive
i'm for one i would gladly pay 700 $ for an APU as powerful as the Series X in the PC, AMD is leaving a huge market
I completely agree that it's beyond time for AMD to get RDNA(2) APUs out there. But these are still really good APUs, and graphics performance might not be better than Renoir, but it's still fine. I would absolutely love a ~20CU RDNA2 APU (which should be feasible in AM4 and with current motherboards), but it's highly doubtful that will happen due to the sheer die size needed. MCM APUs are our best hope for that kind of performance.
I seem to remember differently, as I seem to remember a conspicuous absence of any meniton of retail availability for Renoir.But Dr.Cuttress did say that AMD promises to bring Cezanne to retail. Buuuuuuuuut I could have sworn AMD said the same thing about Renoir so.........
Uh, what? Why? In what world? The iGPU in this will blow the pants off any GT 710, and will likely beat the 1030 in most scenarios too.Big iGPU in APU is waste, as most of these will paired with GT 710/ GT 1030 anyaway.
IC in an APU wouldn't be L3, as L3 is CPU-only. IC would either be a system-wide LLC (whether that acts as L3 or L4 for the CPU depends on whether the CPU has a discrete L3 + whether the IC is inclusive or is bypassed for CPU-only operations) or a GPU-only LLC/Lwhatever (L1/L2/L3/etc. distinctions for GPUs are a bit problematic in transferring CPU terminology 1:1 to a very different architecture).There's no evidence that "infinity cache" (fancy word for L3?) will work for an APU, as yet, not to mention LPDDR5 or LPDDR5x won't come close to the bandwidth a single HBM stack can provide.
As for whether it would work, of course it would. The much more relevant question is whether it would be feasible in a sufficient size without ballooning APU die size. Large cahces take up a lot of space.
Looks reasonable to me. RDNA2 APUs before DDR5/LPDDR5 is likely to be a bit of a whiff. Makes sense in the low-power segments where absolute performance matters less and efficiency matters more, as with Van Gogh. Of course they could no doubt make a faster (LP)DDR4 APU with RDNA2 than Vega, but how much faster would be the question. I'll be happy to wait another year to get a nice RDNA2 APU laptop, and thankfully I went grey-market Renoir for my HTPC