But will it be only small Phoenix 2, or also the full Phoenix? Because I really want all the power
UNLIMITED POOOOOWWWEEEERRRR!!!
There’s no entry level successor for the 200ge and 3000g, or am I mistaken?
The 4300g and 5300g were the smallest after those, OEM only and closer to the 2200g/3200g.
I’d love to see another super cheap banger, sooner or later.
In terms of tech, the 4300G and 5300G were the equivalents. The 3000G was 2 cores in the age where the mainstream APUs were 4 cores, but now they’re 8 cores, so the entry level being 4 cores makes sense.
AMD does have Mendocino for mobile (weak 4 core chip), but it can’t run on the desktop (it only supports LPDDR5).
Is the 12CU RDNA 3 GPU on Ryzen 7 8700g same GPU as 780M found in the laptops?
2c + 4c and 1c + 3c?
Is that a big.LITTLE approach for multitasking?
Kind of. The Zen 4c cores aren’t really “LITTLE”, in the sense that they provide the same performance at the same clock speed, and can only reach lower clocks. When all cores are running together clocks drop anyway, so it’s not a big difference from having only Zen 4 cores.
It’s not completely the same, and for high power desktop chips there will likely be a difference, but it’s still nowhere near what Intel is doing with its P and E cores, and what ARM is doing in mobile.
Neat, thanks
Another garbage namming?
I’m curious, is it possible to make APU with 3D cache? Is it probable AMD will make one at some point?
Intel kinda did it back in the day with 5th gen.
All APUs so far been Mobile chips ported to desktop socket. These chips doesn’t have 3Dcache as cuz they have to support them – must have connectors that enable them to be linked. So no, there is not such for now. Yes, they can do one in the future.
I hope AMD can create a mobile APU with a 3D V-cache in it. That will be on the next steam deck for sure
Technically yes, but AMD had used monolithic dies for their APUs for years and trying to add cache on top of a bigger piece of silicon than a CPU chiplet would need a redesign at least and it would jack up prices quite significantly and most likely we would only see a single SKU with it.
I’d guess a better alternative would be a chiplet based APU that’s similar to Navi 32/31 design (GPU die surrounded by cache chiplets). It would be quite malleable, you could add a Zen X/Zen XC chiplet (or a hybrid of both that you can add 3D V-Cache on top if you wish), a GPU chiplet and a cache chiplet to keep more data to the GPU so it’s less reliant on RAM bandwidth.
Probably AMD may do that in the near future with a RDNA 4/5 APU, the main problem is power consumption (Transferring data between chiplets consumes more power than a monolithic chip with all of its components in a single die) and that would be a significant downside for mobile devices, that’s why they haven’t made a chiplet based APU yet and it’s easier to just recycle/resell any laptop APU as a desktop one.
Technically yes
Techunically IMPOSSIBLE. Not on any existing APU chips or Zen4c chiplet. There’s no TSV contacts to connect the cache.
It’ll require a complete redesign of the chip. But it’s possible if AMD did that.
but AMD had used monolithic dies for their APUs
Irrelevant.
add cache on top of a bigger piece of silicon than a CPU chiplet would need a redesign at least
No. Size is completely irrelevant. Redesign is needed indeed, but it’s the APU that needed redesign, not the cache.
and it would jack up prices quite significantly
No. The filler silicon ALSO USED ON CURRENT X3D costs a few cents at most. Even 200mm² costs pennies.
Also there is HBM memory route. Shame that this technology isn’t cheaper.
Yes, for the looks of it Intel’s alternative to compete with 3D V-Cache is something akin to HBM but more efficient, I bet Intel would use it for their integrated graphics.
It is quite sad that HBM hasn’t decreased in price significantly for it to be embedded in consumer APUs, even 2/4 GB wouldn’t be that bad if HBCC (Uses system RAM as VRAM and HBM is used to cache data) could be used.
Why wouldn’t it be impossible? The 3d Cache sits on top of the ordinary cache and gets data and power connections through vias embedded in the sillicon of the ordinary L3 cache. The distinction that desktop CPUs are chiplet based vs the mobile chips that are monolythic is not valid, because each chiplet is monolythic by itself and can obviously have 3d cache. So, if it suits them AMD can make 3d cache variants of their mobile chips. What’s needed is for them to decide that it’s worth it, then embed the vias in All the silicon (as in the Zen 3 line) and then pull the trigger sometime down the line when they decide that the market conditions are right.
The valid distinction here is that 3d cache is often very useful in server applications, while in normal PCs and laptops it benefits mostly gaming. And, the same chiplets are first binned for server… We gamers happen to enjoy this technology by accident in a way. Having in mind how bandwidth starved are all the CUs in the APUs it will be a godsend for sure. Will AMD do it, considering that they do have the vias in the chiplets in the high end GPUs and still haven’t done it there… Well…
Not sure now if I’m mistaken on what an APU is, but my recently bought 7950x3d has 3d cache and an integrated gpu. Vega I think. Is this not an APU?
Do these have nerfed pcie again?
Phoenix has 20 Gen 4 lanes, vs the 28 Gen 5 lanes of Raphael/Dragon Range
Also AMD:
Announced the end of driver optimizations for the 8000G graphics unit 3 years later.
I hope the 8700G doesn’t jump from last gens $359 to $399. The reduced PC demand may play a hand here to keep pricing the same, but I can totally expect them to push TSMC wafer cost increases down the line as well.
I tried this morning F20A for X670E Master.
I rollback to previous version. It’s not stable at all (stock setting)I hope more stable release until end of january.
Something tells me a r5 7600 + RX 6600 will be cheaper and deliver a lot more FPS in games.
Please, I want to be wrong.
There’s zero chance for AMD to charge $400 for an APU.
Is this the strix point with 36+ gpu cores?
i’d love to see APUs being actually good.
custom sffpc will go on a next level
Be good if their recent bios 20a (x670 Master) didn’t bork the async blk overclocking making it useless for x3d chips.
Well done on the memory though, oc to 8000Mt/s is relatively easy with Hynix a-die
Edit: plenty of options for ocing the integral gpu which is good, no idea if they actually work yet.