4 ms·
CUDA the language itself isn't necessarily the biggest obstacle for porting code, it's the usage of low level hardware features. You'd have to make a chip that'
by ladberg 3y ago
CUDA the language itself isn't necessarily the biggest obstacle for porting code, it's the usage of low level hardware features. You'd have to make a chip that's pretty dang close to Nvidia's before most hand-written CUDA code could run on it performantly.
- rbanffy 3y agoWhat is CUDA not abstracting away? Is it geometry of the SIMD pipelines or the functionalities they have that would be absent (and couldn’t be emulated) by other implementations? In many ways it feels like the same situation we had in the 70’s and 80’s personal computers, with many different software bases divided by different hardware implementations. In that regard, Unix was a boon in that it allowed one to write software that’d be compiled out of a single source into multiple architectures (rather than the PC-clone way, where non-identical implementations were driven to extinction).