New version of this open source high-speed linker

New version of this open source high-speed linker

“Samsung says the PM9E1 is designed to boost on-device AI performance, which laptop makers are beginning to implement with the new incoming Copilot+ PCs.”

Until then … bug reports and fixes !~!


This week I and many other major Plasma contributors are are at Akademy, planning the future and having many fruitful in-person discussions! As a result, probably next week’s post will be a bit light. This week though, lots of pre-Akademy changes got merged, most of them bug-fixes and performance improvements scheduled for Plasma 6.2. We […]
This week in Plasma: Stabilization for 6.2
It was released on October 11, 1999!
NVIDIA’s GeForce 256 had 32MB of VRAM on a 128-bit memory bus with up to 4.6GB/sec of memory bandwidth, while the flagship GeForce RTX 4090 has 24GB of VRAM on a 384-bit memory bus with up to 1008GB/sec (1TB/sec+).
The GeForce 256 had a 50W TDP, while the GeForce RTX 4090 has a 450W+ TDP.
This means that the GPU transistor count between the GeForce 256 and GeForce RTX 4090 is a difference of 3300x while the VRAM capacity has blown up by 768,000x.
