No, but LPDDR means soldered, there are no LPDDR dimms

There's LPCAMM2, but it's very recent. The Framework Pro laptop supports it, for example, although only on the Intel variant.

MediaTek said MediaTek made the CPU: https://www.mediatek.com/press-room/mediatek-collaborates-wi...

Well, MediaTek actually said they made most of the SoC in fact. But the actual CPU cores themselves are all but certainly off-the-shelf Cortex parts, since MediaTek doesn't have a custom core design at all afaik.

NVIDIA hasn't done custom CPU cores for anything they've yet branded "Grace". The original Grace data center CPU (paired with the Hopper data center GPU) used ARM Neoverse V2 cores. The "GB10" chip shipped in DGX Spark and announced here for RTX Spark uses Cortex X925 and Cortex A725 CPU cores.

Physically, NVIDIA did the GPU chiplet and Mediatek did the other chiplet that has the CPU, DRAM controller, and IO.

I think that Nvidia made GPU and CPU, and Mediatek made other parts of SoC necessary for a notebook. Grace is Nvidia's own CPU ARM core

I believe Grace is an ARM designed core. Vera is the nVidia designed core.

They compare HGEMM implementations. At least CUBLAS has HGEMM functions.

HGEMM means half-precision (i.e. FP16) general matrix multiplication

just to make you even more paranoid - zen2/4 (and probably e-cores and zen5) can rename memory operands too (you start to do strange things when Intel limits you to 16 registers)!

so today it needs to go through SSD or maybe the network LOL. but for real conspiracy, we should use paper and pencil

what about 8-bit ANDN? SHL essentially uses CL, so it may be because 8-bit subregister of "constant register" isn't present on the bypass network

I once searched github for simd libraries, sorted by popularity, and added most popular of them to my list: https://github.com/stars/Bulat-Ziganshin/lists/simd

indeed, highway is the popularity leader, it implements lots of SIMD operations and supports lots of hardware including SVE/RVV with runtime SIMD width, although IMHO it has a bit longer learning curve

with m/t, the algorithm is memory-bound, so the performance should be determined strictly by the memory throughput

JOIN SQL operation is another usecase

it's how Google managed to convince young people all over the world to start reading those crappy A&DS books