[rocroller] Build Basic-Block Dependence DAGs by pjots · Pull Request #4785 · ROCm/rocm-libraries

pjots · 2026-02-20T23:36:39Z

Motivation

Build basic block dependence directed acyclic graph (DAG) using the HyperGraph data structure. It can be used to query if two nodes have a data dependency between them.

Technical Details

It supports flow(or true), anti, and output dependence.

Test Plan

Use this dependence DAG to query if two Multiply nodes have output dependence between them.

Test Result

Submission Checklist

Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

…sandhu/BasicBlockDependenceDAGs

Copilot

Pull request overview

Builds a basic-block scoped data-dependence DAG (flow/anti/output) using the existing graph infrastructure and uses it to preserve only dependency-relevant sequencing during implicit-scheduling removal.

Changes:

Added NodeScheduling::constructDataDependenceDAG() to derive a dependence DAG from RW trace records.
Updated RemoveImplicitScheduling to keep/add sequencing edges based on dependence edges instead of the prior colouring-based “edgesToKeep” logic.
Minor doc/comment cleanup and noted grouping by immediate body-parent.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
shared/rocroller/lib/source/KernelGraph/Transformations/RemoveImplicitScheduling.cpp	Switches preserved sequencing to be driven by the constructed data-dependence DAG.
shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp	Implements dependence DAG construction based on per-coordinate RW tracing.
shared/rocroller/lib/include/rocRoller/KernelGraph/NodeSchedulingUtils.hpp	Exposes the new dependence DAG builder API and tweaks comment formatting.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp

shared/rocroller/lib/include/rocRoller/KernelGraph/NodeSchedulingUtils.hpp

shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp

math-ci-webhook · 2026-02-21T00:57:16Z

Performance Report for gfx942

Results

Details

Comparison Summary

Compared: 115
Significant diffs: 20
Insignificant diffs: 95
Not compared (reference-only): 0
Not compared (candidate-only): 0
Not compared (total): 0

@@            Significant (p-val <0.05) Performance Diffs            @@
====================================================================================================
+   0.05% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 3072, N: 4096, K: 4096, alpha: 2, beta: 0.5, types: {'type_A': 'float', 'type_B': 'float', 'type_C': 'float', 'type_D': 'float', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 32, wave_n: 32, wave_k: 2, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: True, prefetchInFlight: 2, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: None, numWGs: 0, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=3072, N=4096, K=4096, alpha=2, beta=0.5, types={'type_A': 'float', 'type_B': 'float', 'type_C': 'float', 'type_D': 'float', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=32, wave_n=32, wave_k=2, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=True, prefetchInFlight=2, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='None', numWGs=0, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   0.04% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8192, alpha: 2, beta: 0, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: True, prefetchInFlight: 2, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: None, numWGs: 0, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8192, alpha=2, beta=0, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=True, prefetchInFlight=2, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='None', numWGs=0, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.02% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8192, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: True, prefetchInFlight: 2, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: None, numWGs: 0, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8192, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=True, prefetchInFlight=2, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='None', numWGs=0, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   0.38% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 128, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=128, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.70% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 128, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=128, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.59% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 64, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=64, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   1.81% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 64, mac_k: 32, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=64, mac_k=32, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   2.34% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 64, mac_k: 32, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=64, mac_k=32, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   1.73% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 64, mac_k: 32, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTileDPFirst, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=64, mac_k=32, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTileDPFirst', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.99% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 128, mac_n: 64, mac_k: 64, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=128, mac_n=64, mac_k=64, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   1.07% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 128, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=128, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.21% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 256, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=256, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.09% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 256, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=256, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.32% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 256, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTileDPFirst, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=256, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTileDPFirst', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.31% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 256, mac_k: 32, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTileDPFirst, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=256, mac_k=32, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTileDPFirst', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   0.31% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   2.90% | p=5.6994e-05 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 16, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=16, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   0.26% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 32, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: TwoTile, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=32, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='TwoTile', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
+   0.60% | p=1.7451e-03 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8448, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 32, wave_n: 32, wave_k: 8, wave_b: 1, workgroup_size_x: 128, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 0, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: Standard, numWGs: 304, architecture: {'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops: True, version: 393404632b9) | GEMMProblem(M=7680, N=8448, K=8448, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'N', 'trans_B': 'T', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=32, wave_n=32, wave_k=8, wave_b=1, workgroup_size_x=128, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=0, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='Standard', numWGs=304, architecture={'ArchString': 'gfx942', 'Xnack': False, 'Sramecc': True}, tailLoops=True, version='')
-   0.70% | p=5.6994e-05 
	| CodeGen(instCount: 40000, instructions: simple_mi)| CodeGen() | CodeGen(instCount: 40000, instructions: simple_mi)

Links

math-ci-webhook · 2026-02-21T00:57:21Z

Resource Report for gfx942

Results

Details

✔️ No Resource Usage Changes ✔️

Links

math-ci-jobs · 2026-02-21T01:00:47Z

Code Coverage Report for gfx942

Summary

Type	Total	Missed	Master Missed	Missed Change	Coverage	Master Coverage	Coverage Change
Lines	66321	12852	12831	21	80.62%	80.63%	-.01%
Functions	6067	1172	1171	1	80.68%	80.67%	.01%
Regions	41122	11580	11569	11	71.84%	71.82%	.02%
Branches	22551	6431	6422	9	71.48%	71.47%	.01%

This PR adds/edits 26 newly uncovered lines.

Artifacts

Commit Hashes

math-ci-jobs · 2026-02-21T01:56:42Z

Generated Documentation

Link to view Generated Docs.

math-ci-jobs · 2026-02-21T02:18:24Z

CodeQL report

Results Summary

Full table of results

Tool	Severity	Code	Location	Line

Links

HTML
Sarif (for download and usage in conjunction with SARIF viewers)

math-ci-webhook · 2026-02-21T06:40:05Z

Performance Report for gfx12

Results

Details

Comparison Summary

Compared: 87
Significant diffs: 4
Insignificant diffs: 83
Not compared (reference-only): 0
Not compared (candidate-only): 0
Not compared (total): 0

@@            Significant (p-val <0.05) Performance Diffs            @@
====================================================================================================
+   0.18% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8192, alpha: 2, beta: 0.5, types: {'type_A': 'fp8', 'type_B': 'bf8', 'type_C': 'float', 'type_D': 'float', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 16, wave_n: 16, wave_k: 16, wave_b: 1, workgroup_size_x: 64, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Priority, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 2, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: None, numWGs: 0, architecture: {'ArchString': 'gfx1201', 'Xnack': False, 'Sramecc': False}, tailLoops: True, version: ea602e0dfc8) | GEMMProblem(M=7680, N=8448, K=8192, alpha=2, beta=0.5, types={'type_A': 'fp8', 'type_B': 'bf8', 'type_C': 'float', 'type_D': 'float', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=16, wave_n=16, wave_k=16, wave_b=1, workgroup_size_x=64, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Priority', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=2, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='None', numWGs=0, architecture={'ArchString': 'gfx1201', 'Xnack': False, 'Sramecc': False}, tailLoops=True, version='')
+   0.10% | p=2.5347e-02 
	| 3. FloatsGEMM(M: 7680, N: 8448, K: 8192, alpha: 2, beta: 0.5, types: {'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A: 1, scaleValue_B: 1, initMode_A: DataInitMode(Bounded), initMode_B: DataInitMode(Bounded), initMode_C: DataInitMode(Bounded), workgroupMappingDim: -1, workgroupMappingValue: -1)| GEMM(mac_m: 64, mac_n: 64, mac_k: 64, wave_m: 16, wave_n: 16, wave_k: 16, wave_b: 1, workgroup_size_x: 64, workgroup_size_y: 2, workgroupRemapXCC: False, workgroupRemapXCCValue: -1, load_A: BufferToLDSViaVGPR, load_B: BufferToLDSViaVGPR, store: VGPRToGlobalMemoryViaLDSWithBuffer, betaInFma: True, padLDS_A: [0, 0], padLDS_B: [0, 0], scheduler: Sequential, schedulerCost: LinearWeighted, prefetch: False, prefetchInFlight: 2, prefetchLDSFactor: 0, prefetchMixMemOps: False, loadScale_A: BufferToVGPR, loadScale_B: BufferToVGPR, swizzleScale: False, swizzleTileSize: {'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale: False, pretileScale: False, streamK: None, numWGs: 0, architecture: {'ArchString': 'gfx1201', 'Xnack': False, 'Sramecc': False}, tailLoops: True, version: ea602e0dfc8) | GEMMProblem(M=7680, N=8448, K=8192, alpha=2, beta=0.5, types={'type_A': 'half', 'type_B': 'half', 'type_C': 'half', 'type_D': 'half', 'type_acc': 'float', 'trans_A': 'T', 'trans_B': 'N', 'scale_A': 'None', 'scaleType_A': 'None', 'scale_B': 'None', 'scaleType_B': 'None', 'scaleBlockSize': -1, 'scaleSkipPermlane': 'None', 'scaleShuffleTileA': [], 'scaleShuffleTileB': [], 'scalePreTileA': [], 'scalePreTileB': [], 'pretileB': []}, scaleValue_A=1, scaleValue_B=1, initMode_A='DataInitMode(Bounded)', initMode_B='DataInitMode(Bounded)', initMode_C='DataInitMode(Bounded)', workgroupMappingDim=-1, workgroupMappingValue=-1)GEMMSolution(mac_m=64, mac_n=64, mac_k=64, wave_m=16, wave_n=16, wave_k=16, wave_b=1, workgroup_size_x=64, workgroup_size_y=2, workgroupRemapXCC=False, workgroupRemapXCCValue=-1, load_A='BufferToLDSViaVGPR', load_B='BufferToLDSViaVGPR', store='VGPRToGlobalMemoryViaLDSWithBuffer', betaInFma=True, padLDS_A=[0, 0], padLDS_B=[0, 0], scheduler='Sequential', schedulerCost='LinearWeighted', prefetch=False, prefetchInFlight=2, prefetchLDSFactor=0, prefetchMixMemOps=False, loadScale_A='BufferToVGPR', loadScale_B='BufferToVGPR', swizzleScale=False, swizzleTileSize={'m': 0, 'k': 0, 'n': 0, 'l': 0}, prefetchScale=False, pretileScale=False, streamK='None', numWGs=0, architecture={'ArchString': 'gfx1201', 'Xnack': False, 'Sramecc': False}, tailLoops=True, version='')
+   2.92% | p=5.6994e-05 
	| CodeGen(instCount: 40000, instructions: complex_mi_with_coop)| CodeGen() | CodeGen(instCount: 40000, instructions: complex_mi_with_coop)
-   3.98% | p=5.6994e-05 
	| CodeGen(instCount: 40000, instructions: simple_mi)| CodeGen() | CodeGen(instCount: 40000, instructions: simple_mi)

Links

math-ci-webhook · 2026-02-21T06:40:08Z

Resource Report for gfx12

Results

Details

✔️ No Resource Usage Changes ✔️

Links

ThanHenderson · 2026-02-23T16:55:36Z

shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp

+        // This assumes that the trace is ordered and records for the
+        // same control operation are consecutive.


Is this enforced elsewhere? We should consider adding a verification step here before this algorithm.

That is essentially how our ControlFlowRWTracer is.
It walks the control flow graph and records when (at which program point) coordinates are accessed/modified. Each record contains control id, coordinate id and the type of access (R/W/RW), and the return type of coordinatesReadWrite() is a vector of these records.

And this comment is to remind that this algorithm uses that property.

added the verification step via seen data structure.

math-ci-webhook · 2026-02-23T19:27:45Z

perfci run on commit `7782f17`

math-ci run

shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-02-23T22:07:36Z

perfci run on commit `38fe5be`

math-ci run

…W dep

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-02-24T01:07:01Z

perfci run on commit `e21943f`

math-ci run

codecov-commenter · 2026-02-24T02:21:01Z

Codecov Report

✅ All modified and coverable lines are covered by tests.

❌ Your project status has failed because the head coverage (77.21%) is below the target coverage (80.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #4785   +/-   ##
========================================
  Coverage    67.22%   67.22%           
========================================
  Files         1839     1839           
  Lines       283079   283079           
  Branches     39714    39714           
========================================
  Hits        190277   190277           
  Misses       76369    76369           
  Partials     16433    16433

Flag	Coverage Δ	*Carryforward flag
hipBLAS	`90.67% <ø> (ø)`	Carriedforward from d95fe63
hipBLASLt	`43.55% <ø> (ø)`
hipCUB	`82.38% <ø> (ø)`	Carriedforward from d95fe63
hipDNN	`85.16% <ø> (ø)`	Carriedforward from d95fe63
hipFFT	`55.59% <ø> (ø)`	Carriedforward from d95fe63
hipRAND	`76.12% <ø> (ø)`	Carriedforward from d95fe63
hipSOLVER	`68.81% <ø> (ø)`	Carriedforward from d95fe63
hipSPARSE	`84.70% <ø> (ø)`	Carriedforward from d95fe63
rocBLAS	`47.97% <ø> (ø)`	Carriedforward from d95fe63
rocFFT	`53.24% <ø> (ø)`	Carriedforward from d95fe63
rocRAND	`57.07% <ø> (ø)`	Carriedforward from d95fe63
rocSOLVER	`77.21% <ø> (ø)`	Carriedforward from d95fe63
rocSPARSE	`71.48% <ø> (ø)`	Carriedforward from d95fe63

*This pull request uses carry forward flags. Click here to find out more.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-03-05T14:00:32Z

perfci run on commit `22a735e`

math-ci run

…sandhu/BasicBlockDependenceDAGs

…OCm/rocm-libraries into users/psandhu/BasicBlockDependenceDAGs

shared/rocroller/lib/include/rocRoller/KernelGraph/ControlGraph/DataDependenceDAG.hpp

math-ci-webhook · 2026-03-06T20:47:50Z

perfci run on commit `78c9a9e`

math-ci run

caio96 · 2026-03-06T21:18:21Z

shared/rocroller/lib/source/KernelGraph/ControlGraph/DataDependenceDAG.cpp

+
+        namespace Detail
+        {
+            DataDependenceDAGDetail::DataDependenceDAGDetail(KernelGraph const& graph)


If it makes sense, the constructor could call constructDataDependenceDAG()

caio96 · 2026-03-06T21:33:19Z

shared/rocroller/lib/include/rocRoller/KernelGraph/ControlGraph/DataDependenceDAG_detail.hpp

+        int  getBodyParent(int control);
+        void addDependenceEdge(int sourceControl, int destControl);
+        void processReadWriteRecord(ControlFlowRWTracer::ReadWriteRecord const& record);
+        void constructDataDependenceDAG();
+        ControlGraph::ControlGraph getDataDependenceDAG();


Brief documentation for these functions would also be good

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-03-07T01:51:28Z

perfci run on commit `f872cdc`

math-ci run

math-ci-webhook · 2026-03-09T22:58:48Z

perfci run on commit `f38c6f0`

math-ci run

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-03-11T21:23:27Z

perfci run on commit `6e6902d`

math-ci run

…sandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-03-12T00:54:52Z

perfci run on commit `76cea6b`

math-ci run

math-ci-webhook · 2026-03-12T17:10:27Z

perfci run on commit `76cea6b`

math-ci run

…OCm/rocm-libraries into users/psandhu/BasicBlockDependenceDAGs

math-ci-webhook · 2026-03-12T20:22:35Z

perfci run on commit `7e9a01b`

math-ci run

…sandhu/BasicBlockDependenceDAGs

pjots added 2 commits February 20, 2026 18:29

construct dependenceDAG and update RemoveImplicitScheduling pass logic

d5c6b3f

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

7782f17

…sandhu/BasicBlockDependenceDAGs

pjots requested a review from a team as a code owner February 20, 2026 23:36

Copilot AI review requested due to automatic review settings February 20, 2026 23:36

github-actions bot added project: hipblaslt shared: rocroller labels Feb 20, 2026

Copilot started reviewing on behalf of pjots February 20, 2026 23:40 View session

assistant-librarian bot added the organization: ROCm label Feb 20, 2026

Copilot AI reviewed Feb 20, 2026

View reviewed changes

awhittle3 changed the title ~~Build Basic-Block Dependence DAGs~~ [rocroller] Build Basic-Block Dependence DAGs Feb 23, 2026

Merge branch 'develop' into users/psandhu/BasicBlockDependenceDAGs

5c3924f

ThanHenderson reviewed Feb 23, 2026

View reviewed changes

shared/rocroller/lib/source/KernelGraph/NodeSchedulingUtils.cpp Outdated Show resolved Hide resolved

pjots added 2 commits February 23, 2026 15:33

apply Nathan's suggested patch -- more modular and reusable code

38fe5be

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

4d30184

…sandhu/BasicBlockDependenceDAGs

pjots added 2 commits February 23, 2026 18:08

order might be undefined in kernelgraph for two Multiply nodes with W…

e21943f

…W dep

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

0dd47e2

…sandhu/BasicBlockDependenceDAGs

pjots added 2 commits February 25, 2026 22:11

WIP

b6df996

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

36aeca0

…sandhu/BasicBlockDependenceDAGs

pjots added 3 commits March 6, 2026 14:02

add a unit test

383333d

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

78c9a9e

…sandhu/BasicBlockDependenceDAGs

Merge branch 'users/psandhu/BasicBlockDependenceDAGs' of github.com:R…

966c558

…OCm/rocm-libraries into users/psandhu/BasicBlockDependenceDAGs

memmett reviewed Mar 6, 2026

View reviewed changes

shared/rocroller/lib/include/rocRoller/KernelGraph/ControlGraph/DataDependenceDAG.hpp Show resolved Hide resolved

caio96 reviewed Mar 6, 2026

View reviewed changes

pjots added 2 commits March 6, 2026 19:16

add a fun Nth fibonacci unit test

f872cdc

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

d3f184d

…sandhu/BasicBlockDependenceDAGs

updated hoist unit test

f38c6f0

pjots added 2 commits March 11, 2026 15:17

address Kerry's comments

6e6902d

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

e789bb6

…sandhu/BasicBlockDependenceDAGs

pjots added 2 commits March 11, 2026 18:01

adds documentation

76bebb8

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

76cea6b

…sandhu/BasicBlockDependenceDAGs

pjots force-pushed the users/psandhu/BasicBlockDependenceDAGs branch from 57302fa to 76cea6b Compare March 11, 2026 22:19

Merge branch 'develop' into users/psandhu/BasicBlockDependenceDAGs

ca52c33

pjots added 2 commits March 12, 2026 14:54

WIP add unit tests

35cdd49

Merge branch 'users/psandhu/BasicBlockDependenceDAGs' of github.com:R…

7e9a01b

…OCm/rocm-libraries into users/psandhu/BasicBlockDependenceDAGs

pjots added 6 commits March 16, 2026 18:05

added check for same basic block, unit testing WIP

c1eaf26

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

c80ba10

…sandhu/BasicBlockDependenceDAGs

return boolean for addDependenceEdge routine

76f2905

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

2542551

…sandhu/BasicBlockDependenceDAGs

remove belongToSameBasicBlock routine and add TODOs

cbdb67a

Merge branch 'develop' of github.com:ROCm/rocm-libraries into users/p…

955a4dd

…sandhu/BasicBlockDependenceDAGs

		// This assumes that the trace is ordered and records for the
		// same control operation are consecutive.

Conversation

pjots commented Feb 20, 2026

Motivation

Technical Details

Test Plan

Test Result

Submission Checklist

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

math-ci-webhook bot commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance Report for gfx942

Results

Comparison Summary

Uh oh!

math-ci-webhook bot commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Resource Report for gfx942

Results

Uh oh!

math-ci-jobs bot commented Feb 21, 2026 • edited by math-ci-status-gate bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Code Coverage Report for gfx942

Summary

Artifacts

Commit Hashes

Uh oh!

math-ci-jobs bot commented Feb 21, 2026

Generated Documentation

Uh oh!

math-ci-jobs bot commented Feb 21, 2026

CodeQL report

Results Summary

Links

Uh oh!

math-ci-webhook bot commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance Report for gfx12

Results

Comparison Summary

Uh oh!

math-ci-webhook bot commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Resource Report for gfx12

Results

Uh oh!

ThanHenderson Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

pjots Feb 24, 2026

Choose a reason for hiding this comment

Uh oh!

pjots Mar 11, 2026

Choose a reason for hiding this comment

Uh oh!

math-ci-webhook bot commented Feb 23, 2026

perfci run on commit 7782f17

Uh oh!

Uh oh!

math-ci-webhook bot commented Feb 23, 2026

perfci run on commit 38fe5be

Uh oh!

math-ci-webhook bot commented Feb 24, 2026

perfci run on commit e21943f

Uh oh!

codecov-commenter commented Feb 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

math-ci-webhook bot commented Mar 5, 2026

perfci run on commit 22a735e

Uh oh!

math-ci-webhook bot commented Feb 21, 2026 •

edited

Loading

math-ci-webhook bot commented Feb 21, 2026 •

edited

Loading

math-ci-jobs bot commented Feb 21, 2026 •

edited by math-ci-status-gate bot

Loading

math-ci-webhook bot commented Feb 21, 2026 •

edited

Loading

math-ci-webhook bot commented Feb 21, 2026 •

edited

Loading

perfci run on commit `7782f17`

perfci run on commit `38fe5be`

perfci run on commit `e21943f`

codecov-commenter commented Feb 24, 2026 •

edited

Loading

perfci run on commit `22a735e`

perfci run on commit `78c9a9e`

perfci run on commit `f872cdc`

perfci run on commit `f38c6f0`

perfci run on commit `6e6902d`

perfci run on commit `76cea6b`

perfci run on commit `76cea6b`

perfci run on commit `7e9a01b`