Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Clang][SME2] Add multi-vector add/sub builtins #69725

Merged
merged 7 commits into from
Nov 7, 2023
Merged

[Clang][SME2] Add multi-vector add/sub builtins #69725

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
8c9c9bd
[Clang][SME2] Add multi-vector add/sub builtins
kmclaughlin-arm Oct 11, 2023
2c64efe
- Rename GetAArch64SMEProcessedOperands to GetAArch64SVEProcessedOper…
kmclaughlin-arm Oct 23, 2023
3c770ac
- Removed the add of an immediate offset from the tests
kmclaughlin-arm Oct 26, 2023
08fefc9
- Removed incorrect tests added for multi-vector sub
kmclaughlin-arm Oct 26, 2023
4cf4256
- Run update_cc_test_checks script on add/sub tests
kmclaughlin-arm Nov 2, 2023
2bf7de4
- Add __arm_streaming and __arm_shared_za attributes to the tests
kmclaughlin-arm Nov 6, 2023
65b7b66
- Moved the multi-single vector adds to a separate test file
kmclaughlin-arm Nov 6, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
35 changes: 35 additions & 0 deletions clang/include/clang/Basic/arm_sme.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -263,3 +263,38 @@ multiclass ZAFPOuterProd<string n_suffix> {

defm SVMOPA : ZAFPOuterProd<"mopa">;
defm SVMOPS : ZAFPOuterProd<"mops">;

////////////////////////////////////////////////////////////////////////////////
// SME2 - ADD, SUB

multiclass ZAAddSub<string n_suffix> {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit:

Suggested change
multiclass ZAAddSub<string n_suffix> {
multiclass AddSubZA<string n_suffix> {

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi @sdesmalen-arm, I'm happy to change this, though the reason I chose this name was to match other multiclasses in the file such as ZAStore, ZAAdd, ZAFPOuterProd, etc.

let TargetGuard = "sme2" in {
def NAME # _WRITE_SINGLE_ZA32_VG1X2_I32 : Inst<"sv" # n_suffix # "_write[_single]_za32[_{d}]_vg1x2", "vm2d", "iUi", MergeNone, "aarch64_sme_" # n_suffix # "_write_single_za_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _WRITE_SINGLE_ZA32_VG1X4_I32 : Inst<"sv" # n_suffix # "_write[_single]_za32[_{d}]_vg1x4", "vm4d", "iUi", MergeNone, "aarch64_sme_" # n_suffix # "_write_single_za_vg1x4", [IsStreaming, IsSharedZA], []>;

def NAME # _WRITE_ZA32_VG1X2_I32 : Inst<"sv" # n_suffix # "_write_za32[_{d}]_vg1x2", "vm22", "iUi", MergeNone, "aarch64_sme_" # n_suffix # "_write_za_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _WRITE_ZA32_VG1X4_I32 : Inst<"sv" # n_suffix # "_write_za32[_{d}]_vg1x4", "vm44", "iUi", MergeNone, "aarch64_sme_" # n_suffix # "_write_za_vg1x4", [IsStreaming, IsSharedZA], []>;

def NAME # _ZA32_VG1x2_I32 : Inst<"sv" # n_suffix # "_za32[_{d}]_vg1x2", "vm2", "iUif", MergeNone, "aarch64_sme_" # n_suffix # "_za32_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _ZA32_VG1X4_I32 : Inst<"sv" # n_suffix # "_za32[_{d}]_vg1x4", "vm4", "iUif", MergeNone, "aarch64_sme_" # n_suffix # "_za32_vg1x4", [IsStreaming, IsSharedZA], []>;

let TargetGuard = "sme-i16i64" in {
def NAME # _WRITE_SINGLE_ZA64_VG1X2_I64 : Inst<"sv" # n_suffix # "_write[_single]_za64[_{d}]_vg1x2", "vm2d", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_write_single_za_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _WRITE_SINGLE_ZA64_VG1X4_I64 : Inst<"sv" # n_suffix # "_write[_single]_za64[_{d}]_vg1x4", "vm4d", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_write_single_za_vg1x4", [IsStreaming, IsSharedZA], []>;

def NAME # _WRITE_ZA64_VG1x2_I64 : Inst<"sv" # n_suffix # "_write_za64[_{d}]_vg1x2", "vm22", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_write_za_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _WRITE_ZA64_VG1x4_I64 : Inst<"sv" # n_suffix # "_write_za64[_{d}]_vg1x4", "vm44", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_write_za_vg1x4", [IsStreaming, IsSharedZA], []>;

def NAME # _ZA64_VG1X2_I64 : Inst<"sv" # n_suffix # "_za64[_{d}]_vg1x2", "vm2", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_za64_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _ZA64_VG1X4_I64 : Inst<"sv" # n_suffix # "_za64[_{d}]_vg1x4", "vm4", "lUl", MergeNone, "aarch64_sme_" # n_suffix # "_za64_vg1x4", [IsStreaming, IsSharedZA], []>;
}

let TargetGuard = "sme-f64f64" in {
def NAME # _ZA64_VG1X2_F64 : Inst<"sv" # n_suffix # "_za64[_{d}]_vg1x2", "vm2", "d", MergeNone, "aarch64_sme_" # n_suffix # "_za64_vg1x2", [IsStreaming, IsSharedZA], []>;
def NAME # _ZA64_VG1X4_F64 : Inst<"sv" # n_suffix # "_za64[_{d}]_vg1x4", "vm4", "d", MergeNone, "aarch64_sme_" # n_suffix # "_za64_vg1x4", [IsStreaming, IsSharedZA], []>;
}
}
}

defm SVADD : ZAAddSub<"add">;
defm SVSUB : ZAAddSub<"sub">;
12 changes: 12 additions & 0 deletions clang/include/clang/Basic/arm_sve.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1980,3 +1980,15 @@ def SVCNTP_COUNT : SInst<"svcntp_{d}", "n}i", "QcQsQiQl", MergeNone, "aarch64_sv

defm SVREVD : SInstZPZ<"svrevd", "csilUcUsUiUl", "aarch64_sve_revd">;
}

////////////////////////////////////////////////////////////////////////////////
// SME2

// SME intrinsics which operate only on vectors and do not require ZA should be added here,
// as they could possibly become SVE instructions in the future.

let TargetGuard = "sme2" in {
kmclaughlin-arm marked this conversation as resolved.
Show resolved Hide resolved
// == ADD (vectors) ==
def SVADD_SINGLE_X2 : SInst<"svadd[_single_{d}_x2]", "22d", "cUcsUsiUilUl", MergeNone, "aarch64_sve_add_single_x2", [IsStreaming], []>;
def SVADD_SINGLE_X4 : SInst<"svadd[_single_{d}_x4]", "44d", "cUcsUsiUilUl", MergeNone, "aarch64_sve_add_single_x4", [IsStreaming], []>;
}
2 changes: 1 addition & 1 deletion clang/include/clang/Basic/arm_sve_sme_incl.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -257,7 +257,7 @@ class ImmCheck<int arg, ImmCheckType kind, int eltSizeArg = -1> {
}

class Inst<string n, string p, string t, MergeType mt, string i,
list<FlagType> ft, list<ImmCheck> ch, MemEltType met> {
list<FlagType> ft, list<ImmCheck> ch, MemEltType met = MemEltTyDefault> {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am not sure about propose of this change "met = MemEltTyDefault"? The last parameter in multiclass ZAAddSub with Inst class usage is always []?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi @dtemirbulatov, this change is just to set the default MemEltType to MemEltTyDefault, so that we don't have to provide this in every builtin definition.

string Name = n;
string Prototype = p;
string Types = t;
Expand Down
1 change: 1 addition & 0 deletions clang/lib/CodeGen/CodeGenFunction.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4297,6 +4297,7 @@ class CodeGenFunction : public CodeGenTypeCache {
/// the wider vector. This avoids the error when allocating space in llvm
/// for struct of scalable vectors if a function returns struct.
llvm::Value *FormSVEBuiltinResult(llvm::Value *Call);

llvm::Value *EmitAArch64SVEBuiltinExpr(unsigned BuiltinID, const CallExpr *E);

llvm::Value *EmitSMELd1St1(const SVETypeFlags &TypeFlags,
Expand Down
Loading