Vega 12 Support Is Now Available For RadeonSI Gallium3D

[coder]

Must be the vega mobile discrete chip, rumored to have 24 CUs. I would also bet that this is the exacly same chip found in the Intel Kaby Lake G, the only difference being Intel's EMIB vs. AMD's inteposer for the HBM.

https://videocardz.com/74712/amd-fin...le-radeon-vega

I would love to have that on a 75 W low-profile PCIe card.

[Figueiredo]

Must be the vega mobile discrete chip, rumored to have 24 CUs. I would also bet that this is the exacly same chip found in the Intel Kaby Lake G, the only difference being Intel's EMIB vs. AMD's inteposer for the HBM. Can't see the sense in AMD developing a whole different chip with such similar features for such a niche product. Probably a contract thing due to Intel wanting to market their design as something very exclusive...

https://videocardz.com/74712/amd-finally-announces-discrete-mobile-radeon-vega

[Aleksei]

You've apparently been in a deep, dark hole.

This sounds wrong, somehow.

This exploded our collective heads, a few months back.
https://www.anandtech.com/show/12207...-overclockable
https://www.anandtech.com/show/12220...phics-launched

More details:
https://ark.intel.com/products/13040...up-to-4_20-GHz
https://ark.intel.com/products/13040...up-to-4_10-GHz
https://ark.intel.com/products/13040...up-to-4_10-GHz
https://ark.intel.com/products/13041...up-to-4_10-GHz
https://ark.intel.com/products/13040...up-to-3_80-GHz

Thanks! I had to check that anandtech link twice to make sure it's 1st January, not 1st April.

[valici]

Yeah, I've already been down that rabbit hole and was fearing to confirm whether Vega 11 is really in the APU with 11 CUs or something else. At least, that's what comes up if you search for Vega 11.

As my previous question implied, questionable wisdom behind using both of those naming schemes - especially on the same family.

I really like the commercial naming scheme. You know right away from the name what kind of performance you should expect from the card and what you pay for.
Like you know Vega 56 should be just under Vega 64.
Just that chip codename should not overlap with the commercial name. Like Vega should've started from Vega100 or so... because there is no way a Vega with 100 CU can exist.

[coder]

if Vega10 is launched, Vega12 is close, than Vega11 should be also because it was started to be developed before Vega12.

Yeah, I've already been down that rabbit hole and was fearing to confirm whether Vega 11 is really in the APU with 11 CUs or something else. At least, that's what comes up if you search for Vega 11.

As my previous question implied, questionable wisdom behind using both of those naming schemes - especially on the same family.

[valici]

They are going in the order we started working on them.

That's the whole idea of having code names, so we can reference products before launch without exposing all of the details

Raja also confirmed the numbering used. But now the logical conclusion to this is if Vega10 is launched, Vega12 is close, than Vega11 should be also because it was started to be developed before Vega12.

[pal666]

Really?

not it shipping yet

[pal666]

I hope they're not going to mix nomenclature. We had Polaris 10, 11, and 12, for the 36, 16, and 8 CU chips.

Then, we get Vega 56 and 64, for 56 CUs and 64 CUs.

no, we get vega 10, 11 and 12. vega 64's counterpart is 580, not polaris

[coder]

Really?

You've apparently been in a deep, dark hole.

This exploded our collective heads, a few months back.
https://www.anandtech.com/show/12207...-overclockable
https://www.anandtech.com/show/12220...phics-launched

More details:
https://ark.intel.com/products/13040...up-to-4_20-GHz
https://ark.intel.com/products/13040...up-to-4_10-GHz
https://ark.intel.com/products/13040...up-to-4_10-GHz
https://ark.intel.com/products/13041...up-to-4_10-GHz

https://ark.intel.com/products/130408/Intel-Core-i5-8305G-Processor-with-Radeon-RX-Vega-M-GL-graphics-6M-Cache-up-to-3_80-GHz

[Aleksei]

the Vega GPU found on select Intel CPUs.

Really?