Fix cuda backend perf when nsoc is large

[govindchari]

No description provided.

[github-actions]

Download benchmark artifacts

Benchmark Summary

• Baseline solved: 129 problems
• Diff branch solved: 129 problems

Runtime regressions (> 5.0%)

• AUG3DCQP: diff=0.0239s, baseline=0.0206s, Δ=+16.3%
• CVXQP2_S: diff=0.0004s, baseline=0.0004s, Δ=+6.5%
• HS52: diff=0.0000s, baseline=0.0000s, Δ=+33.3%
• UBH1: diff=0.0309s, baseline=0.0283s, Δ=+9.0%

Runtime improvements (> 5.0%)

• DUAL2: diff=0.0022s, baseline=0.0024s, Δ=-6.0%
• GENHS28: diff=0.0000s, baseline=0.0000s, Δ=-20.0%
• HS268: diff=0.0000s, baseline=0.0000s, Δ=-5.6%
• HS76: diff=0.0000s, baseline=0.0000s, Δ=-5.9%
• QRECIPE: diff=0.0011s, baseline=0.0012s, Δ=-5.1%
• QSC205: diff=0.0009s, baseline=0.0010s, Δ=-8.0%
• QSCFXM1: diff=0.0063s, baseline=0.0068s, Δ=-7.0%
• QSCRS8: diff=0.0076s, baseline=0.0081s, Δ=-7.0%

[github-actions]

Download benchmark artifacts

Benchmark Summary

• Baseline solved: 129 problems
• Diff branch solved: 129 problems

Runtime regressions (> 5.0%)

• AUG2D: diff=0.0097s, baseline=0.0089s, Δ=+9.0%
• AUG2DCQP: diff=0.1374s, baseline=0.1218s, Δ=+12.9%
• DUALC1: diff=0.0013s, baseline=0.0012s, Δ=+7.5%
• QGROW7: diff=0.0045s, baseline=0.0042s, Δ=+8.5%
• QPTEST: diff=0.0000s, baseline=0.0000s, Δ=+9.1%
• STADAT1: diff=0.0169s, baseline=0.0159s, Δ=+6.4%

Runtime improvements (> 5.0%)

• HS76: diff=0.0000s, baseline=0.0000s, Δ=-5.9%
• QBANDM: diff=0.0046s, baseline=0.0049s, Δ=-5.6%

[github-actions]

Download benchmark artifacts

Benchmark Summary

• Baseline solved: 129 problems
• Diff branch solved: 129 problems

Runtime regressions (> 5.0%)

• DUALC5: diff=0.0007s, baseline=0.0007s, Δ=+7.1%
• HS268: diff=0.0000s, baseline=0.0000s, Δ=+5.3%
• HS35MOD: diff=0.0000s, baseline=0.0000s, Δ=+5.9%
• QBORE3D: diff=0.0025s, baseline=0.0024s, Δ=+5.3%
• ZECEVIC2: diff=0.0000s, baseline=0.0000s, Δ=+7.1%

Runtime improvements (> 5.0%)

• AUG3DC: diff=0.0016s, baseline=0.0017s, Δ=-9.0%
• HS52: diff=0.0000s, baseline=0.0000s, Δ=-25.0%
• QBRANDY: diff=0.0035s, baseline=0.0037s, Δ=-5.8%