Repository navigation
[Mono] AdvSimd.Arm64 [Load/Store]Vector64 and [Load/Store]Vector128 for 2,3 and 4 variants with LLVM #93081
Description
Refer to #84510 for the full list of API's. Re-enable tests being disabled tests.
Part of the work of this issue is to handle tuple type input arguments in those API's. To implement it, see the following example for inspiration
runtime/src/mono/mono/mini/simd-intrinsics.c
Line 4183 in 4325acc
AdvSimd and AdvSimd.Arm64 intrinsics are currently only supported by LLVM. So these new intrinsics will follow this trend.
This is the list of APIs for AdvSimd:
- LoadVector64x2 - @fanyang-mono
- LoadVector64x3 - @fanyang-mono
- LoadVector64x4 - @fanyang-mono
- LoadVector64x2AndUnzip - @fanyang-mono
- LoadVector64x3AndUnzip - @fanyang-mono
- LoadVector64x4AndUnzip - @fanyang-mono
- LoadAndInsertScalar - @fanyang-mono
- LoadAndReplicateToVector64x2 - @fanyang-mono
- LoadAndReplicateToVector64x3 - @fanyang-mono
- LoadAndReplicateToVector64x4 - @fanyang-mono
- StoreVector64x2 - @jkurdek
- StoreVector64x3 - @jkurdek
- StoreVector64x4 - @jkurdek
- StoreVector64x2AndZip - @jkurdek
- StoreVector64x3AndZip - @jkurdek
- StoreVector64x4AndZip - @jkurdek
- StoreSelectedScalar - @fanyang-mono
This is the list of APIs for AdvSimd.Arm64:
- LoadVector128x2 - @fanyang-mono
- LoadVector128x3 - @fanyang-mono
- LoadVector128x4 - @fanyang-mono
- LoadVector128x2AndUnzip - @fanyang-mono
- LoadVector128x3AndUnzip - @fanyang-mono
- LoadVector128x4AndUnzip - @fanyang-mono
- LoadAndInsertScalar - @fanyang-mono
- LoadAndReplicateToVector128x2 - @fanyang-mono
- LoadAndReplicateToVector128x3 - @fanyang-mono
- LoadAndReplicateToVector128x4 - @fanyang-mono
- StoreVector128x2 - @jkurdek
- StoreVector128x3 - @jkurdek
- StoreVector128x4 - @jkurdek
- StoreVector128x2AndZip - @jkurdek
- StoreVector128x3AndZip - @jkurdek
- StoreVector128x4AndZip - @jkurdek
- StoreSelectedScalar - @fanyang-mono
Activity
[-][Mono] Arm64 AdvSimd - Implement LoadVector64x* and LoadVector128x* APIs[/-][+][Mono] AdvSimd.Arm64 [Load/Store]Vector64 and [Load/Store]Vector128 for 2,3 and 4 variants[/+][-][Mono] AdvSimd.Arm64 [Load/Store]Vector64 and [Load/Store]Vector128 for 2,3 and 4 variants[/-][+][Mono] AdvSimd.Arm64 [Load/Store]Vector64 and [Load/Store]Vector128 for 2,3 and 4 variants with LLVM[/+]kunalspathak commentedon Feb 8, 2024
@fanyang-mono - this should help you!
fanyang-mono commentedon Feb 21, 2024
Closing this issue, as all the tasks listed here have been completed.