AMD says its new laptop chips can beat Apple

… like the Mac book pro?

I started using Apple when the M1 was launched, the battery duration and the no moving parts (e.g. fan) in the Air plays in a different league. I continue to have a love "relationship" with my "old" XPSs but I can only use them plugged or buying a new battery every year...

BTW the Thunderbolt support on the 2018 XPS continues to be superior to Apple now (e.g. more monitors) but the battery duration is a lifechanger.

Unfortunately AMD is still quite a bit off in terms IPC.

The FPGA market has climbed ever more up-market, building only ever bigger better but at exponentially vaster stratospheric price points. There's a couple mid-tier Lattice parts that have emerged, and Gowin has some real disruption from below just starting, but big playerskke Xilinx/AMD and Altera/Intel (since re-spun off) have made FPGAs ever more unattractive & unaffordable year by year.

It seems very unlikely that the FPGA people actually want their parts accessible useful or commoditied. There is a very strong demand for FPGAs that justifies their multi-thousand dollar price points for ever pretty modest sized chips (which can be scaled up to huge house-priced sizes too); nothing else gives you what fpgas do. There's no substitute. It would be too risky for these folks to actually allow reconfigurable computing to be in reach.

Some of the SmartNIC offerings that hyperscalers and hyperscalers alone use have nice but kind of specialized FPGAs onboard. Marvell/Nvidias's Bluefield rocks. But again it's a $500+ network card with capabilities most folks won't be able to use well. It's especially damned a proposition because there's no solutions available for use & the necessary AMD Vivado software needed to program the fpga is thousands of dollars and that's the starting cost before you start signing contracts to get access to the dram controllers or to Ethernet phy.

Every now and then there's something that seems like a ray of hope in the gloom - for ex AMD's nanotube ebpf -> fpga converter (https://www.phoronix.com/news/AMD-Xilinx-Nanotube-Compiler) - but you need that expensive closed tool chain to start and the efforts usually seem quickly abandoned after a code drop or two, no real effort to make it happen.

It's hellworld. It's so sad to see an industry so near to the rest of computing and the pax Romana of "PC compatible" interoperability & mutual growth, but so unable to cross the chasm & flourish, just in such a closed off isolated locked-in-its-head niche darkworld. I get it, there's no evidence the volume isn't there so chasing whales is what you have to do.

FPGAs are such amazing & interesting systems, but it's been a terrible decade of them getting nowhere. Zynq is an ancient relic reminding us of hope lost. I feel like the best case possibility is that this NPU AI thing maybe breaths some life in; XDNA is supposedly a somewhat special small FPGA, and that maybe possibly over time could help bring FPGAs out of the dark & into the light of day. One could hope.

AMD already makes Arm SoCs for Xilinx markets.

Nvidia and Mediatek are likely working on Arm laptop SoCs. AMD could follow.

The appeal of the M1 MBA even today isn't that it's fast, it's that it's fast enough while also delivering great battery life, not throttling when unplugged, and being dead silent and cool to the touch. It's everything that an everyday laptop should be.