Re: CRC32C Parallel Computation Optimization on ARM

On Thu, Oct 26, 2023 at 07:28:35AM +0000, Xiang Gao wrote:
> On Wed, 25 Oct, 2023 at 10:43:25 -0500, Nathan Bossart wrote:
>>+# Use ARM VMULL if available and ARM CRC32C intrinsic is avaliable too.
>>+if test x"$USE_ARMV8_VMULL" = x"" && (test x"$USE_ARMV8_CRC32C" = x"1" || test x"$USE_ARMV8_CRC32C_WITH_RUNTIME_CHECK" = x"1"); then
>>+ if test x"$pgac_armv8_vmull_intrinsics" = x"yes"; then
>>+ USE_ARMV8_VMULL=1
>>+ fi
>>+fi
>
>>Hm. I wonder if we need to switch to a runtime check in some cases. For
>>example, what happens if the ARMv8 intrinsics used today are found with the
>>default compiler flags, but vmull_p64() is only available if
>>-march=armv8-a+crypto is added? It looks like the precedent is to use a
>>runtime check if we need extra CFLAGS to produce code that uses the
>>intrinsics.
>
> We consider that a runtime check needs to be done in any scenario.
> Here we only confirm that the compilation can be successful.
> A runtime check will be done when choosing which algorithm.
> You can think of us as merging USE_ARMV8_VMULL and USE_ARMV8_VMULL_WITH_RUNTIME_CHECK into USE_ARMV8_VMULL.

Oh. Looking again, I see that we are using a runtime check for ARM in all
cases with this patch. If so, maybe we should just remove
USE_ARV8_CRC32C_WITH_RUNTIME_CHECK in a prerequisite patch (and have
USE_ARMV8_CRC32C always do the runtime check). I suspect there are other
opportunities to simplify things, too.

--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com