By: --- (---.delete@this.redheron.com), August 8, 2022 2:54 pm
Room: Moderated Discussions
I found a lot interesting in this update:
https://community.arm.com/arm-community-blogs/b/tools-software-ides-blog/posts/arm-compiler-for-linux-and-arm-performance-libraries-22-0
On the plus side, it's clear that there still remains some low-hanging fruit for the ARM ecosystem when compared to the x86 ecosystem, even in fairly basic things like special functions and BLAS.
On the negative side, the SVE results seem disappointing. One can spin this in a few different ways (isolated loops won't show the size improvements from simpler loops without head and tails, as opposed to real large apps using large shared libraries; these are already vector-dense loops, whereas SVE should help more code that's less trivially vectorizable), and clearly the SVE optimization has only just begun.
Still, somewhat disappointing results. (Except of course that himeno number. Given how much Phoronix pushes himeno, I look forward to seeing Michael try to justify this!)
Perhaps hoping for better loops in generic code requires SVE2 and we can't hope for much of that with just SVE?
https://community.arm.com/arm-community-blogs/b/tools-software-ides-blog/posts/arm-compiler-for-linux-and-arm-performance-libraries-22-0
On the plus side, it's clear that there still remains some low-hanging fruit for the ARM ecosystem when compared to the x86 ecosystem, even in fairly basic things like special functions and BLAS.
On the negative side, the SVE results seem disappointing. One can spin this in a few different ways (isolated loops won't show the size improvements from simpler loops without head and tails, as opposed to real large apps using large shared libraries; these are already vector-dense loops, whereas SVE should help more code that's less trivially vectorizable), and clearly the SVE optimization has only just begun.
Still, somewhat disappointing results. (Except of course that himeno number. Given how much Phoronix pushes himeno, I look forward to seeing Michael try to justify this!)
Perhaps hoping for better loops in generic code requires SVE2 and we can't hope for much of that with just SVE?
| Topic | Posted By | Date |
|---|---|---|
| Interesting ARM compiler data | --- | 2022/08/08 02:54 PM |
| Interesting ARM compiler data | noko | 2022/08/08 09:30 PM |
| V1 bottleneck | Jan Wassenberg | 2022/08/09 12:38 AM |
| Interesting ARM compiler data | --- | 2022/08/09 10:15 AM |
| Interesting ARM compiler data | noko | 2022/08/09 11:34 AM |
| Interesting ARM compiler data | Jörn Engel | 2022/08/09 01:45 PM |
| Interesting ARM compiler data | --- | 2022/08/09 01:49 PM |
