[core] Fix some more memtoreg issues. by schweitzpgi · Pull Request #5293 · NVIDIA/cuda-quantum

schweitzpgi · 2026-08-26T23:30:07Z

regressions in handling classical values discovered and fixed
add a canonical ordering to arguments appending
add missing guards that resulted in crashes
add new regression tests

schweitzpgi · 2026-08-26T23:32:39Z

Fixes #5286

sacpis

Thanks @schweitzpgi. Left few comments.

github-actions · 2026-08-27T00:08:47Z

CI Summary (`push`) — ✅ passed

Run #33116907599 · ✅ 7 · ⏩ 7 · ❌ 0 · ⛔ 0

Top-level jobs (14)

Job	Result
`binaries`	⏩ skipped
`build_and_test`	✅ success
`changes`	✅ success
`config_devdeps`	✅ success
`config_source_build`	⏩ skipped
`config_wheeldeps`	✅ success
`devdeps`	✅ success
`docker_image`	⏩ skipped
`gen_code_coverage`	⏩ skipped
`metadata`	✅ success
`python_metapackages`	⏩ skipped
`python_wheels`	⏩ skipped
`source_build`	⏩ skipped
`wheeldeps`	✅ success

⏩ Skipped jobs (7) — intentionally skipped on PR builds; run on merge_group / workflow_dispatch

Job
`binaries`
`config_source_build`
`docker_image`
`gen_code_coverage`
`python_metapackages`
`python_wheels`
`source_build`

All sub-jobs (43) — every matrix leg, with links

Job	Status	Link
Build and test (amd64, gcc12, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, gcc12, openmpi) / Dev environment (Python)	✅ success	view
Build and test (amd64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, llvm, openmpi) / Dev environment (Python)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Python)	✅ success	view
CI Summary	❔ in_progress	view
Check for stable CUDA-Q changes	✅ success	view
Configure build (devdeps)	✅ success	view
Configure build (source_build)	⏩ skipped	view
Configure build (wheeldeps)	✅ success	view
Create CUDA Quantum installer	⏩ skipped	view
Create Docker images	⏩ skipped	view
Create Python metapackages	⏩ skipped	view
Create Python wheels	⏩ skipped	view
Gen code coverage	⏩ skipped	view
Load dependencies (amd64, gcc12) / Caching	✅ success	view
Load dependencies (amd64, gcc12) / Finalize	✅ success	view
Load dependencies (amd64, gcc12) / Metadata	✅ success	view
Load dependencies (amd64, llvm) / Caching	✅ success	view
Load dependencies (amd64, llvm) / Finalize	✅ success	view
Load dependencies (amd64, llvm) / Metadata	✅ success	view
Load dependencies (arm64, gcc12) / Caching	✅ success	view
Load dependencies (arm64, gcc12) / Finalize	✅ success	view
Load dependencies (arm64, gcc12) / Metadata	✅ success	view
Load dependencies (arm64, llvm) / Caching	✅ success	view
Load dependencies (arm64, llvm) / Finalize	✅ success	view
Load dependencies (arm64, llvm) / Metadata	✅ success	view
Load source build cache	⏩ skipped	view
Load wheel dependencies (amd64, 12.6) / Caching	✅ success	view
Load wheel dependencies (amd64, 12.6) / Finalize	✅ success	view
Load wheel dependencies (amd64, 12.6) / Metadata	✅ success	view
Load wheel dependencies (amd64, 13.0) / Caching	✅ success	view
Load wheel dependencies (amd64, 13.0) / Finalize	✅ success	view
Load wheel dependencies (amd64, 13.0) / Metadata	✅ success	view
Load wheel dependencies (arm64, 12.6) / Caching	✅ success	view
Load wheel dependencies (arm64, 12.6) / Finalize	✅ success	view
Load wheel dependencies (arm64, 12.6) / Metadata	✅ success	view
Load wheel dependencies (arm64, 13.0) / Caching	✅ success	view
Load wheel dependencies (arm64, 13.0) / Finalize	✅ success	view
Load wheel dependencies (arm64, 13.0) / Metadata	✅ success	view
Prepare cache clean-up	❔ in_progress	view
Retrieve PR info	✅ success	view

✅ Required checks (6/6) — declared in .github/required-checks.yml for push

Required check	Status	Link
Build and test (amd64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, llvm, openmpi) / Dev environment (Python)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Python)	✅ success	view
Build and test (amd64, gcc12, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, gcc12, openmpi) / Dev environment (Python)	✅ success	view

khalatepradnya

Thanks, @schweitzpgi for the updates!
I tested the failing test on my branch with this change, and it is working.

nvq++ --target quantinuum --emulate  ../targettests/execution/estimate_resources_break.cpp && ./a.out
Total # of gates: 32, total # of qubits: 1, circuit depth: 32, multi-Q gate count: 0, multi-Q depth: 0
{ 
  reset :  10
  mz :  10
  x :  11
  h :  1
}

khalatepradnya

Is it possible to enable the test in issue #5286?

cuda-quantum/targettests/execution/estimate_resources_break.cpp

Lines 11 to 12 in 76213fb

    
           // XFAIL: * 
        
           // FIXME: https://github.com/NVIDIA/cuda-quantum/issues/5286

schweitzpgi · 2026-08-27T18:57:06Z

Is it possible to enable the test in issue #5286?

cuda-quantum/targettests/execution/estimate_resources_break.cpp

Lines 11 to 12 in 76213fb

// XFAIL: *

// FIXME: https://github.com/NVIDIA/cuda-quantum/issues/5286

Let me check...

schweitzpgi · 2026-08-27T19:21:29Z

Is it possible to enable the test in issue #5286?

cuda-quantum/targettests/execution/estimate_resources_break.cpp

Lines 11 to 12 in 76213fb

// XFAIL: *

// FIXME: https://github.com/NVIDIA/cuda-quantum/issues/5286

Let me check...

It looks like this was disabled because of how long it took it CI. I won't mess with it in this PR. That's an unrelated issue.

- regressions in handling classical values discovered and fixed - add a canonical ordering to arguments appending - add missing guards that resulted in crashes - add new regression tests Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

atgeller

👍

Signed-off-by: Pradnya Khalate <pkhalate@nvidia.com>

khalatepradnya · 2026-08-27T20:17:30Z

Is it possible to enable the test in issue #5286?

cuda-quantum/targettests/execution/estimate_resources_break.cpp

Lines 11 to 12 in 76213fb

// XFAIL: *

// FIXME: https://github.com/NVIDIA/cuda-quantum/issues/5286

Let me check...

It looks like this was disabled because of how long it took it CI. I won't mess with it in this PR. That's an unrelated issue.

It passed in the latest CI (Ref: logs)

Unexpectedly Passed Tests (1):
  CUDAQ-Target :: execution/estimate_resources_break.cpp

I have pushed a commit to enable it.

atgeller

I've found a regression from main in a simple test case with a quantum and classical variable:

func.func @min(%c: i1, %k: i32) {
  %q = quake.alloca !quake.ref
  %m = cc.alloca i32
  cc.if (%c) {
    %v = cc.load %m : !cc.ptr<i32>
    cc.store %k, %m : !cc.ptr<i32>
    quake.x %q : (!quake.ref) -> ()
  }
  quake.y %q : (!quake.ref) -> ()
  %f = cc.load %m : !cc.ptr<i32>
  quake.log_output %f : (i32) -> () {compilerGenerated}
  quake.dealloc %q : !quake.ref
  return
}

With cudaq-opt --memtoreg fails with

test.qke:4:3: error: 'cc.if' op along control flow edge from Operation cc.continue to parent: successor operand type #0 '!quake.wire' should match successor input type #0 'i32'
  cc.if (%c) {
  ^
test.qke:4:3: note: see current operation: 
%2:2 = "cc.if"(%arg0, %0) ({
^bb0(%arg3: !quake.wire):
  %4 = "quake.x"(%arg3) <{operandSegmentSizes = array<i32: 0, 0, 1>}> : (!quake.wire) -> !quake.wire
  "cc.continue"(%4, %arg1) : (!quake.wire, i32) -> ()
}, {
^bb0(%arg2: !quake.wire):
  "cc.continue"(%arg2, %1) : (!quake.wire, i32) -> ()
}) : (i1, !quake.wire) -> (i32, !quake.wire)
test.qke:4:3: note: region branch point

This works on main.

Signed-off-by: Sachin Pisal <spisal@nvidia.com>

sacpis · 2026-08-27T21:11:29Z

I've found a regression from main in a simple test case with a quantum and classical variable:

func.func @min(%c: i1, %k: i32) {
  %q = quake.alloca !quake.ref
  %m = cc.alloca i32
  cc.if (%c) {
    %v = cc.load %m : !cc.ptr<i32>
    cc.store %k, %m : !cc.ptr<i32>
    quake.x %q : (!quake.ref) -> ()
  }
  quake.y %q : (!quake.ref) -> ()
  %f = cc.load %m : !cc.ptr<i32>
  quake.log_output %f : (i32) -> () {compilerGenerated}
  quake.dealloc %q : !quake.ref
  return
}

With cudaq-opt --memtoreg fails with

test.qke:4:3: error: 'cc.if' op along control flow edge from Operation cc.continue to parent: successor operand type #0 '!quake.wire' should match successor input type #0 'i32'
  cc.if (%c) {
  ^
test.qke:4:3: note: see current operation: 
%2:2 = "cc.if"(%arg0, %0) ({
^bb0(%arg3: !quake.wire):
  %4 = "quake.x"(%arg3) <{operandSegmentSizes = array<i32: 0, 0, 1>}> : (!quake.wire) -> !quake.wire
  "cc.continue"(%4, %arg1) : (!quake.wire, i32) -> ()
}, {
^bb0(%arg2: !quake.wire):
  "cc.continue"(%arg2, %1) : (!quake.wire, i32) -> ()
}) : (i1, !quake.wire) -> (i32, !quake.wire)
test.qke:4:3: note: region branch point

This works on main.

Pushed a fix with 023be4e. Thanks @atgeller.

sacpis · 2026-08-27T21:14:59Z

I've found a regression from main in a simple test case with a quantum and classical variable:

func.func @min(%c: i1, %k: i32) {
  %q = quake.alloca !quake.ref
  %m = cc.alloca i32
  cc.if (%c) {
    %v = cc.load %m : !cc.ptr<i32>
    cc.store %k, %m : !cc.ptr<i32>
    quake.x %q : (!quake.ref) -> ()
  }
  quake.y %q : (!quake.ref) -> ()
  %f = cc.load %m : !cc.ptr<i32>
  quake.log_output %f : (i32) -> () {compilerGenerated}
  quake.dealloc %q : !quake.ref
  return
}

With cudaq-opt --memtoreg fails with

test.qke:4:3: error: 'cc.if' op along control flow edge from Operation cc.continue to parent: successor operand type #0 '!quake.wire' should match successor input type #0 'i32'
  cc.if (%c) {
  ^
test.qke:4:3: note: see current operation: 
%2:2 = "cc.if"(%arg0, %0) ({
^bb0(%arg3: !quake.wire):
  %4 = "quake.x"(%arg3) <{operandSegmentSizes = array<i32: 0, 0, 1>}> : (!quake.wire) -> !quake.wire
  "cc.continue"(%4, %arg1) : (!quake.wire, i32) -> ()
}, {
^bb0(%arg2: !quake.wire):
  "cc.continue"(%arg2, %1) : (!quake.wire, i32) -> ()
}) : (i1, !quake.wire) -> (i32, !quake.wire)
test.qke:4:3: note: region branch point

This works on main.

Pushed a fix with 023be4e. Thanks @atgeller.

BTW, the root cause here was that canonicalizeArgumentOrder permuted region block arguments but left liveOutSet (which fixes the parent's appended result order) untouched. Two sorts patched over the gap and could not for cc.if, whose LinearTypeArgs trait keeps classical values out of block arguments entirely, yielding results (i32, wire) against cc.continue operands (wire, i32).

To fix this issue, we did reorder liveOutSet (and rebuild liveInArgs) to the canonical order, so results, terminator operands, and block arguments share one order. Also, dropped both sorts and the exitSiblingSuccessors map.

atgeller

Thanks for the fix @sacpis! Working now 😄

github-actions · 2026-08-27T22:38:22Z

CI Summary (`merge_group`) — ✅ passed

Run #33120401868 · ✅ 5 · ⏩ 9 · ❌ 0 · ⛔ 0

⚡ Fast merge queue — only the llvm dev-environment build+test legs ran.
Wheels, metapackages, installers, Docker images, and the gcc12 legs were skipped.
Set the repository variable MERGE_QUEUE_FAST to false to restore the full matrix.

Top-level jobs (14)

Job	Result
`binaries`	⏩ skipped
`build_and_test`	✅ success
`changes`	✅ success
`config_devdeps`	✅ success
`config_source_build`	⏩ skipped
`config_wheeldeps`	⏩ skipped
`devdeps`	✅ success
`docker_image`	⏩ skipped
`gen_code_coverage`	⏩ skipped
`metadata`	✅ success
`python_metapackages`	⏩ skipped
`python_wheels`	⏩ skipped
`source_build`	⏩ skipped
`wheeldeps`	⏩ skipped

⏩ Skipped jobs (9) — intentionally skipped on PR builds; run on merge_group / workflow_dispatch

Job
`binaries`
`config_source_build`
`config_wheeldeps`
`docker_image`
`gen_code_coverage`
`python_metapackages`
`python_wheels`
`source_build`
`wheeldeps`

All sub-jobs (24) — every matrix leg, with links

Job	Status	Link
Build and test (amd64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, llvm, openmpi) / Dev environment (Python)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Python)	✅ success	view
CI Summary	❔ in_progress	view
Check for stable CUDA-Q changes	✅ success	view
Configure build (devdeps)	✅ success	view
Configure build (source_build)	⏩ skipped	view
Configure build (wheeldeps)	⏩ skipped	view
Create CUDA Quantum installer	⏩ skipped	view
Create Docker images	⏩ skipped	view
Create Python metapackages	⏩ skipped	view
Create Python wheels	⏩ skipped	view
Gen code coverage	⏩ skipped	view
Load dependencies (amd64, llvm) / Caching	✅ success	view
Load dependencies (amd64, llvm) / Finalize	✅ success	view
Load dependencies (amd64, llvm) / Metadata	✅ success	view
Load dependencies (arm64, llvm) / Caching	✅ success	view
Load dependencies (arm64, llvm) / Finalize	✅ success	view
Load dependencies (arm64, llvm) / Metadata	✅ success	view
Load source build cache	⏩ skipped	view
Load wheel dependencies	⏩ skipped	view
Prepare cache clean-up	✅ success	view
Retrieve PR info	✅ success	view

✅ Required checks (4/4) — declared in .github/required-checks.yml for merge_group_fast

Required check	Status	Link
Build and test (amd64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (amd64, llvm, openmpi) / Dev environment (Python)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Debug)	✅ success	view
Build and test (arm64, llvm, openmpi) / Dev environment (Python)	✅ success	view

schweitzpgi requested review from 1tnguyen and khalatepradnya August 26, 2026 23:30

github-actions Bot added python-lang Anything related to the Python CUDA Quantum language implementation core compiler labels Aug 26, 2026

schweitzpgi force-pushed the ch-memtoreg branch from 6aa8ec8 to 8e937a5 Compare August 26, 2026 23:33

sacpis reviewed Aug 27, 2026

View reviewed changes

Comment thread cudaq/lib/Optimizer/Transforms/MemToReg.cpp Outdated

Comment thread cudaq/lib/Optimizer/Transforms/MemToReg.cpp

khalatepradnya reviewed Aug 27, 2026

View reviewed changes

schweitzpgi requested a review from khalatepradnya August 27, 2026 02:17

sacpis mentioned this pull request Aug 27, 2026

Incorrect value when a loop body copies the induction variable #5245

Open

4 tasks

schweitzpgi requested a review from atgeller August 27, 2026 15:01

schweitzpgi enabled auto-merge August 27, 2026 15:02

khalatepradnya reviewed Aug 27, 2026

View reviewed changes

schweitzpgi disabled auto-merge August 27, 2026 19:00

schweitzpgi added 4 commits August 27, 2026 12:22

[core] Fix some more memtoreg issues.

4a81b8b

- regressions in handling classical values discovered and fixed - add a canonical ordering to arguments appending - add missing guards that resulted in crashes - add new regression tests Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

Pick up some more straggling bugs along the way.

3e326d2

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

Fix python for-else regressions.

f3f01a4

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

Small fix for sigma node insertion inconsistency.

17bef63

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>

schweitzpgi force-pushed the ch-memtoreg branch from c877d41 to 17bef63 Compare August 27, 2026 19:22

schweitzpgi enabled auto-merge August 27, 2026 19:23

atgeller approved these changes Aug 27, 2026

View reviewed changes

Re-enable failing test

b3f85ff

Signed-off-by: Pradnya Khalate <pkhalate@nvidia.com>

atgeller requested changes Aug 27, 2026

View reviewed changes

fixing MemToReg result/terminator order mismatch

023be4e

Signed-off-by: Sachin Pisal <spisal@nvidia.com>

atgeller approved these changes Aug 27, 2026

View reviewed changes

schweitzpgi added this pull request to the merge queue Aug 27, 2026

Merged via the queue into NVIDIA:main with commit 1800140 Aug 27, 2026
75 checks passed

schweitzpgi deleted the ch-memtoreg branch August 27, 2026 22:47

github-actions Bot pushed a commit that referenced this pull request Aug 27, 2026

Cleaning up docs preview for PR #5293.

73fe827

	// XFAIL: *
	// FIXME: https://github.com/NVIDIA/cuda-quantum/issues/5286

Uh oh!

Conversation

schweitzpgi commented Aug 26, 2026

Uh oh!

schweitzpgi commented Aug 26, 2026

Uh oh!

sacpis left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented Aug 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CI Summary (push) — ✅ passed

Uh oh!

khalatepradnya left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

khalatepradnya left a comment

Choose a reason for hiding this comment

Uh oh!

schweitzpgi commented Aug 27, 2026

Uh oh!

schweitzpgi commented Aug 27, 2026

Uh oh!

atgeller left a comment

Choose a reason for hiding this comment

Uh oh!

khalatepradnya commented Aug 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

atgeller left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sacpis commented Aug 27, 2026

Uh oh!

sacpis commented Aug 27, 2026

Uh oh!

atgeller left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Aug 27, 2026

CI Summary (merge_group) — ✅ passed

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

github-actions Bot commented Aug 27, 2026 •

edited

Loading

CI Summary (`push`) — ✅ passed

khalatepradnya left a comment •

edited

Loading

khalatepradnya commented Aug 27, 2026 •

edited

Loading

atgeller left a comment •

edited

Loading

CI Summary (`merge_group`) — ✅ passed