Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

>GPU was never the botteneck >it was memory layout

ah right so the GPU was the bottleneck then



No because he was able to achieve the speedup without changing the GPU.


A more technically correct way to express this feeling is:

"The computational power of the cores on the GPU was never the issue-- however the code that I wrote resulted in a memory bandwidth bottleneck that starved the GPU cores of data to work on, which is firmly within my responsibilities as a programmer -- to fully understand the bandwidth and latency characteristics of the device(s) i'm running on"


I mean they didn't write the code


And that's the reason why they misspoke




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: