Intel Acquires The Team Behind ArrayFire GPU Acceleration / Parallel Computing Software

[phoronix]

Phoronix: Intel Acquires The Team Behind ArrayFire GPU Acceleration / Parallel Computing Software

In addition to Intel acquiring Linutronix as the company known for their work on the real-time (RT) kernel patches and other contributions and then back in June acquiring Codeplay Software, Intel has today made another notable software talent acquisition... Intel announced this afternoon that the team behind ArrayFire has joined the company to further their ambitious software endeavors...

https://www.phoronix.com/news/Intel-ArrayFire-oneAPI-Hire

[andrei_me]

OneAPI execution seems like what AMD wanted to do with HSA but couldn't deliver

[Ironmask]

God I hate the term "GPGPU", not only is it an oxymoron but it also sounds like the speaker is stuttering. Just call it a Parallel Processing Unit.

[coder]

God I hate the term "GPGPU", not only is it an oxymoron but it also sounds like the speaker is stuttering. Just call it a Parallel Processing Unit.

The term originated in the early 2000's and referred to running General Purpose workloads on GPUs. So, not an oxymoron and not referring strictly to the hardware, but rather to the hardware/software combination.

Here's a site I used to follow, back then: https://web.archive.org/web/20031011...www.gpgpu.org/

[coder]

AMD should've bought a couple of these software houses. Then, their driver team could've focused on core ROCm features + hardware support, and let the others focus on HiP and porting CUDA code to HiP/ROCm.

Probably AMD's biggest failing, in their attempts to borrow Nvidia's compute strategy, was the lack of investment in software on anything like the scale Nvidia has done.

[WannaBeOCer]

AMD should've bought a couple of these software houses. Then, their driver team could've focused on core ROCm features + hardware support, and let the others focus on HiP and porting CUDA code to ROCm.

Probably AMD's biggest failing in their attempts to borrow Nvidia's compute strategy was the lack of investment in software, on anything like the scale Nvidia has done.

Seemed like one of the main reasons Raja Koduri left. He had many ideas but AMD was focusing on gaming instead. Now at Intel we can see his ideas come to fruition.

Raja Koduri: “AMD has no software ecosystem that's meaningful without Intel“

[Quackdoc]

at this rate I could see rocm simply becoming a backend for oneapi in most cases

[qarium]

Probably AMD's biggest failing, in their attempts to borrow Nvidia's compute strategy, was the lack of investment in software on anything like the scale Nvidia has done.

according to bridgman back in 2017 they did not have the money to do so. they now have the money and because of this we see vega64 support in blender 3.3.

[Ironmask]

running General Purpose workloads on GPUs. So, not an oxymoron and not referring strictly to the hardware, but rather to the hardware/software combination.

Well the 8086 architecture was originally made to control traffic lights.
Should we call all our software "general purpose traffic light processors"?

[coder]

according to bridgman back in 2017 they did not have the money to do so.

I understand that AMD's finances were still a little rough, back then. However, if they were serious about playing in the GPU compute and AI markets, then they should've appreciated the need to make these kinds of investments.