[mlir][arith] Add division integration tests

[pingshiyu]

Division operations regression tests, from the original larger PR that has been broken down: #92272

[llvmbot]

@llvm/pr-subscribers-mlir

Author: Jacob Yu (pingshiyu)

Changes

Division operations regression tests, from the original larger PR that has been broken down: #92272

---

Full diff: https://github.com/llvm/llvm-project/pull/98181.diff

1 Files Affected:

• (added) mlir/test/Integration/Dialect/Arith/CPU/divide.mlir (+109)

diff --git a/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir b/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir
new file mode 100644
index 0000000000000..a08f7975f5968
--- /dev/null
+++ b/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir
@@ -0,0 +1,109 @@
+// Tests division operations and their variants (e.g. ceil/floordiv, rem etc)
+
+// RUN: mlir-opt %s --arith-expand --test-lower-to-llvm | \
+// RUN:   mlir-cpu-runner -e entry -entry-point-result=void \
+// RUN:                   --shared-libs=%mlir_c_runner_utils | \
+// RUN:   FileCheck %s --match-full-lines
+
+func.func @divsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@divsi_i8\n"
+    %0 = arith.divsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @divsi_i1(%v1 : i1, %v2 : i1) {
+    vector.print str "@divsi_i1\n"
+    %0 = arith.divsi %v1, %v2 : i1
+    vector.print %0 : i1
+    return
+
+}
+
+func.func @remsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@remsi_i8\n"
+    %0 = arith.remsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @ceildivsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@ceildivsi_i8\n"
+    %0 = arith.ceildivsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @divsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %c68 = arith.constant 68 : i8
+    %cn97 = arith.constant -97 : i8
+
+    // divsi should round towards zero (rather than -infinity)
+    // divsi -97 68 = -1
+    // CHECK-LABEL: @divsi_i8
+    // CHECK-NEXT:  -1
+    func.call @divsi_i8(%cn97, %c68) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // Test i1
+    // ------------------------------------------------
+    %false = arith.constant false
+    %true = arith.constant true
+
+    // CHECK-LABEL: @divsi_i1
+    // CHECK-NEXT:  1
+    func.call @divsi_i1(%true, %true) : (i1, i1) -> ()
+
+    // ------------------------------------------------
+    // TODO: i16, i32 etc
+    // ------------------------------------------------   
+    return 
+}
+
+func.func @remsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %cn1 = arith.constant -1 : i8
+    %i8_min_p1 = arith.constant -127 : i8
+    
+    // remsi minIntPlus1 -1 = remsi -2^(w-1) -1 = 0
+    // CHECK-LABEL: @remsi_i8
+    // CHECK-NEXT:  0
+    func.call @remsi_i8(%i8_min_p1, %cn1) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // TODO: i1, i16 etc
+    // ------------------------------------------------ 
+    return
+}
+
+func.func @ceildivsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %c7 = arith.constant 7 : i8
+    %i8_min = arith.constant -128 : i8
+
+    // ceildivsi should keep signs
+    // forall w, y. (w > 0, y > 0) => -2^w `ceildiv` y : i_w < 0
+    // CHECK-LABEL: @ceildivsi_i8
+    // CHECK-NEXT:  -18
+    func.call @ceildivsi_i8(%i8_min, %c7) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // TODO: i1, i16 etc
+    // ------------------------------------------------ 
+
+    return
+}
+
+func.func @entry() {
+    func.call @divsi() : () -> ()
+    func.call @remsi() : () -> ()
+    func.call @ceildivsi() : () -> ()
+    return
+}

[llvmbot]

@llvm/pr-subscribers-mlir-arith

Author: Jacob Yu (pingshiyu)

Changes

Division operations regression tests, from the original larger PR that has been broken down: #92272

---

Full diff: https://github.com/llvm/llvm-project/pull/98181.diff

1 Files Affected:

• (added) mlir/test/Integration/Dialect/Arith/CPU/divide.mlir (+109)

diff --git a/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir b/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir
new file mode 100644
index 0000000000000..a08f7975f5968
--- /dev/null
+++ b/mlir/test/Integration/Dialect/Arith/CPU/divide.mlir
@@ -0,0 +1,109 @@
+// Tests division operations and their variants (e.g. ceil/floordiv, rem etc)
+
+// RUN: mlir-opt %s --arith-expand --test-lower-to-llvm | \
+// RUN:   mlir-cpu-runner -e entry -entry-point-result=void \
+// RUN:                   --shared-libs=%mlir_c_runner_utils | \
+// RUN:   FileCheck %s --match-full-lines
+
+func.func @divsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@divsi_i8\n"
+    %0 = arith.divsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @divsi_i1(%v1 : i1, %v2 : i1) {
+    vector.print str "@divsi_i1\n"
+    %0 = arith.divsi %v1, %v2 : i1
+    vector.print %0 : i1
+    return
+
+}
+
+func.func @remsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@remsi_i8\n"
+    %0 = arith.remsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @ceildivsi_i8(%v1 : i8, %v2 : i8) {
+    vector.print str "@ceildivsi_i8\n"
+    %0 = arith.ceildivsi %v1, %v2 : i8
+    vector.print %0 : i8
+    return
+}
+
+func.func @divsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %c68 = arith.constant 68 : i8
+    %cn97 = arith.constant -97 : i8
+
+    // divsi should round towards zero (rather than -infinity)
+    // divsi -97 68 = -1
+    // CHECK-LABEL: @divsi_i8
+    // CHECK-NEXT:  -1
+    func.call @divsi_i8(%cn97, %c68) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // Test i1
+    // ------------------------------------------------
+    %false = arith.constant false
+    %true = arith.constant true
+
+    // CHECK-LABEL: @divsi_i1
+    // CHECK-NEXT:  1
+    func.call @divsi_i1(%true, %true) : (i1, i1) -> ()
+
+    // ------------------------------------------------
+    // TODO: i16, i32 etc
+    // ------------------------------------------------   
+    return 
+}
+
+func.func @remsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %cn1 = arith.constant -1 : i8
+    %i8_min_p1 = arith.constant -127 : i8
+    
+    // remsi minIntPlus1 -1 = remsi -2^(w-1) -1 = 0
+    // CHECK-LABEL: @remsi_i8
+    // CHECK-NEXT:  0
+    func.call @remsi_i8(%i8_min_p1, %cn1) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // TODO: i1, i16 etc
+    // ------------------------------------------------ 
+    return
+}
+
+func.func @ceildivsi() {
+    // ------------------------------------------------
+    // Test i8
+    // ------------------------------------------------
+    %c7 = arith.constant 7 : i8
+    %i8_min = arith.constant -128 : i8
+
+    // ceildivsi should keep signs
+    // forall w, y. (w > 0, y > 0) => -2^w `ceildiv` y : i_w < 0
+    // CHECK-LABEL: @ceildivsi_i8
+    // CHECK-NEXT:  -18
+    func.call @ceildivsi_i8(%i8_min, %c7) : (i8, i8) -> ()
+
+    // ------------------------------------------------
+    // TODO: i1, i16 etc
+    // ------------------------------------------------ 
+
+    return
+}
+
+func.func @entry() {
+    func.call @divsi() : () -> ()
+    func.call @remsi() : () -> ()
+    func.call @ceildivsi() : () -> ()
+    return
+}

[banach-space]

Would you be able to add some contrasting edge cases to the ones that you already have here?

[pingshiyu]

Thank you for the suggestion!

Would you be able to add some contrasting edge cases to the ones that you already have here?

I am not 100% I grokked what you meant here by "contrasting", but I've added some edge cases to the suite now :)

[pingshiyu]

Thanks for the comments @banach-space!

I've just pushed a commit addressing those :)

[pingshiyu]

@kuhar @banach-space all comments addressed :)

One of the tests will fail due to an existing bug: #106519

[banach-space]

Given the discussion here:

• [mlir][arith] Add more canonicalization and integration tests coverage #92272,

and the resulting update to the MLIR guidelines on e2e tests:

• Add a paragraph on integration tests mlir-www#203,

I suggest that we park this for now.

@pingshiyu, thank you so much for your contribution and for bearing with us throughout this process. Apologies for not resolving the earlier discussion sooner—it would have saved you a lot of effort. I hope this doesn’t discourage you from contributing to MLIR in the future, and we truly appreciate your work!