HW3: RC2 - A prephase added for v1 and v2

homework_3/.gitignore

@@ -20,4 +20,8 @@ various/
 .vs/
 .vscode/
 
+# nvidia
+*.ncu-proj
+
+
 

homework_3/analyse/RC1-7a6f7f5/Pending-PIDs-ampere

@@ -0,0 +1,72 @@
+Submitting: hpc/BitncV0Q20.sh
+Submitted batch job 1914643
+Submitting: hpc/BitncV0Q21.sh
+Submitted batch job 1914644
+Submitting: hpc/BitncV0Q22.sh
+Submitted batch job 1914645
+Submitting: hpc/BitncV0Q23.sh
+Submitted batch job 1914646
+Submitting: hpc/BitncV0Q24.sh
+Submitted batch job 1914647
+Submitting: hpc/BitncV0Q25.sh
+Submitted batch job 1914648
+Submitting: hpc/BitncV0Q26.sh
+Submitted batch job 1914649
+Submitting: hpc/BitncV0Q27.sh
+Submitted batch job 1914650
+Submitting: hpc/BitncV0Q28.sh
+Submitted batch job 1914651
+Submitting: hpc/BitncV0Q29.sh
+Submitted batch job 1914652
+Submitting: hpc/BitncV0Q30.sh
+Submitted batch job 1914653
+
+
+Submitting: hpc/BitncV1Q20.sh
+Submitted batch job 1914654
+Submitting: hpc/BitncV1Q21.sh
+Submitted batch job 1914655
+Submitting: hpc/BitncV1Q22.sh
+Submitted batch job 1914656
+Submitting: hpc/BitncV1Q23.sh
+Submitted batch job 1914657
+Submitting: hpc/BitncV1Q24.sh
+Submitted batch job 1914658
+Submitting: hpc/BitncV1Q25.sh
+Submitted batch job 1914659
+Submitting: hpc/BitncV1Q26.sh
+Submitted batch job 1914660
+Submitting: hpc/BitncV1Q27.sh
+Submitted batch job 1914661
+Submitting: hpc/BitncV1Q28.sh
+Submitted batch job 1914662
+Submitting: hpc/BitncV1Q29.sh
+Submitted batch job 1914663
+Submitting: hpc/BitncV1Q30.sh
+Submitted batch job 1914664
+
+
+Submitting: hpc/BitncV2Q20.sh
+Submitted batch job 1914665
+Submitting: hpc/BitncV2Q21.sh
+Submitted batch job 1914666
+Submitting: hpc/BitncV2Q22.sh
+Submitted batch job 1914667
+Submitting: hpc/BitncV2Q23.sh
+Submitted batch job 1914668
+Submitting: hpc/BitncV2Q24.sh
+Submitted batch job 1914669
+Submitting: hpc/BitncV2Q25.sh
+Submitted batch job 1914670
+Submitting: hpc/BitncV2Q26.sh
+Submitted batch job 1914671
+Submitting: hpc/BitncV2Q27.sh
+Submitted batch job 1914672
+Submitting: hpc/BitncV2Q28.sh
+Submitted batch job 1914673
+Submitting: hpc/BitncV2Q29.sh
+Submitted batch job 1914674
+Submitting: hpc/BitncV2Q30.sh
+Submitted batch job 1914675
+
+

homework_3/analyse/RC1-7a6f7f5/Pending-PIDs-gpu

@@ -0,0 +1,70 @@
+Submitting: hpc/BitncV0Q20.sh
+Submitted batch job 1914677
+Submitting: hpc/BitncV0Q21.sh
+Submitted batch job 1914678
+Submitting: hpc/BitncV0Q22.sh
+Submitted batch job 1914679
+Submitting: hpc/BitncV0Q23.sh
+Submitted batch job 1914680
+Submitting: hpc/BitncV0Q24.sh
+Submitted batch job 1914681
+Submitting: hpc/BitncV0Q25.sh
+Submitted batch job 1914682
+Submitting: hpc/BitncV0Q26.sh
+Submitted batch job 1914683
+Submitting: hpc/BitncV0Q27.sh
+Submitted batch job 1914684
+Submitting: hpc/BitncV0Q28.sh
+Submitted batch job 1914685
+Submitting: hpc/BitncV0Q29.sh
+Submitted batch job 1914686
+Submitting: hpc/BitncV0Q30.sh
+Submitted batch job 1914687
+
+
+Submitting: hpc/BitncV1Q20.sh
+Submitted batch job 1914688
+Submitting: hpc/BitncV1Q21.sh
+Submitted batch job 1914689
+Submitting: hpc/BitncV1Q22.sh
+Submitted batch job 1914690
+Submitting: hpc/BitncV1Q23.sh
+Submitted batch job 1914691
+Submitting: hpc/BitncV1Q24.sh
+Submitted batch job 1914692
+Submitting: hpc/BitncV1Q25.sh
+Submitted batch job 1914693
+Submitting: hpc/BitncV1Q26.sh
+Submitted batch job 1914694
+Submitting: hpc/BitncV1Q27.sh
+Submitted batch job 1914695
+Submitting: hpc/BitncV1Q28.sh
+Submitted batch job 1914696
+Submitting: hpc/BitncV1Q29.sh
+Submitted batch job 1914697
+Submitting: hpc/BitncV1Q30.sh
+Submitted batch job 1914698
+
+
+Submitting: hpc/BitncV2Q20.sh
+Submitted batch job 1914699
+Submitting: hpc/BitncV2Q21.sh
+Submitted batch job 1914700
+Submitting: hpc/BitncV2Q22.sh
+Submitted batch job 1914701
+Submitting: hpc/BitncV2Q23.sh
+Submitted batch job 1914702
+Submitting: hpc/BitncV2Q24.sh
+Submitted batch job 1914703
+Submitting: hpc/BitncV2Q25.sh
+Submitted batch job 1914704
+Submitting: hpc/BitncV2Q26.sh
+Submitted batch job 1914705
+Submitting: hpc/BitncV2Q27.sh
+Submitted batch job 1914706
+Submitting: hpc/BitncV2Q28.sh
+Submitted batch job 1914707
+Submitting: hpc/BitncV2Q29.sh
+Submitted batch job 1914708
+Submitting: hpc/BitncV2Q30.sh
+Submitted batch job 1914709

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q20.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1048576 (Q=20)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 7118 [usec]
+[Timing] Mem-xch : 3881 [usec]
+[Timing] Sorting : 3233 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q21.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     2097152 (Q=21)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 7597 [usec]
+[Timing] Mem-xch : 3359 [usec]
+[Timing] Sorting : 4237 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q22.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     4194304 (Q=22)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 10 [msec]
+[Timing] Mem-xch : 4320 [usec]
+[Timing] Sorting : 5982 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q23.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     8388608 (Q=23)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 29 [msec]
+[Timing] Mem-xch : 14 [msec]
+[Timing] Sorting : 14 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q24.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     16777216 (Q=24)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 43 [msec]
+[Timing] Mem-xch : 13 [msec]
+[Timing] Sorting : 29 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q25.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     33554432 (Q=25)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 89 [msec]
+[Timing] Mem-xch : 29 [msec]
+[Timing] Sorting : 59 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q26.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     67108864 (Q=26)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 184 [msec]
+[Timing] Mem-xch : 63 [msec]
+[Timing] Sorting : 121 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q27.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     134217728 (Q=27)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 414 [msec]
+[Timing] Mem-xch : 157 [msec]
+[Timing] Sorting : 255 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q28.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     268435456 (Q=28)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 909 [msec]
+[Timing] Mem-xch : 363 [msec]
+[Timing] Sorting : 548 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q29.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     536870912 (Q=29)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 2005 [msec]
+[Timing] Mem-xch : 840 [msec]
+[Timing] Sorting : 1163 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV0Q30.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1073741824 (Q=30)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 3593 [msec]
+[Timing] Mem-xch : 1137 [msec]
+[Timing] Sorting : 2456 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q20.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1048576 (Q=20)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 5607 [usec]
+[Timing] Mem-xch : 4043 [usec]
+[Timing] Sorting : 1562 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q21.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     2097152 (Q=21)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 4605 [usec]
+[Timing] Mem-xch : 2073 [usec]
+[Timing] Sorting : 2367 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q22.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     4194304 (Q=22)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 11 [msec]
+[Timing] Mem-xch : 7261 [usec]
+[Timing] Sorting : 3887 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q23.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     8388608 (Q=23)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 16 [msec]
+[Timing] Mem-xch : 8281 [usec]
+[Timing] Sorting : 8624 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q24.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     16777216 (Q=24)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 33 [msec]
+[Timing] Mem-xch : 15 [msec]
+[Timing] Sorting : 18 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q25.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     33554432 (Q=25)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 65 [msec]
+[Timing] Mem-xch : 27 [msec]
+[Timing] Sorting : 38 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q26.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     67108864 (Q=26)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 136 [msec]
+[Timing] Mem-xch : 63 [msec]
+[Timing] Sorting : 72 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q27.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     134217728 (Q=27)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 281 [msec]
+[Timing] Mem-xch : 125 [msec]
+[Timing] Sorting : 156 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q28.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     268435456 (Q=28)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 731 [msec]
+[Timing] Mem-xch : 366 [msec]
+[Timing] Sorting : 362 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q29.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     536870912 (Q=29)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 1378 [msec]
+[Timing] Mem-xch : 632 [msec]
+[Timing] Sorting : 753 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV1Q30.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1073741824 (Q=30)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 3177 [msec]
+[Timing] Mem-xch : 1564 [msec]
+[Timing] Sorting : 1580 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q20.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1048576 (Q=20)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 3147 [usec]
+[Timing] Mem-xch : 1491 [usec]
+[Timing] Sorting : 1646 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q21.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     2097152 (Q=21)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 4908 [usec]
+[Timing] Mem-xch : 2369 [usec]
+[Timing] Sorting : 2545 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q22.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     4194304 (Q=22)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 8561 [usec]
+[Timing] Mem-xch : 4249 [usec]
+[Timing] Sorting : 4299 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q23.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     8388608 (Q=23)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 17 [msec]
+[Timing] Mem-xch : 8507 [usec]
+[Timing] Sorting : 9197 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q24.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     16777216 (Q=24)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 34 [msec]
+[Timing] Mem-xch : 14 [msec]
+[Timing] Sorting : 19 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q25.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     33554432 (Q=25)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 69 [msec]
+[Timing] Mem-xch : 28 [msec]
+[Timing] Sorting : 41 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q26.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     67108864 (Q=26)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 149 [msec]
+[Timing] Mem-xch : 71 [msec]
+[Timing] Sorting : 87 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q27.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     134217728 (Q=27)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 323 [msec]
+[Timing] Mem-xch : 151 [msec]
+[Timing] Sorting : 166 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q28.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     268435456 (Q=28)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 754 [msec]
+[Timing] Mem-xch : 367 [msec]
+[Timing] Sorting : 384 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q29.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     536870912 (Q=29)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 1425 [msec]
+[Timing] Mem-xch : 639 [msec]
+[Timing] Sorting : 796 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/RC1-7a6f7f5/ampere/slurm-BitncV2Q30.out

@@ -0,0 +1,22 @@
+[Log]: Array size:     1073741824 (Q=30)
+[Log]: Repeated sorts: 7
+[Log]: GPU:            NVIDIA A100-SXM4-40GB
+[Log]: Block size:     512
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Log]: Initialize array ...  Done.
+[Log]: Start sorting ...     Done.
+[Timing] Total   : 3231 [msec]
+[Timing] Mem-xch : 1532 [msec]
+[Timing] Sorting : 1676 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/Pending-PIDs

@@ -0,0 +1,45 @@
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q20.sh 
+Submitted batch job 1914456
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q21.sh 
+Submitted batch job 1914457
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q22.sh 
+Submitted batch job 1914458
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q23.sh 
+Submitted batch job 1914459
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q24.sh 
+Submitted batch job 1914460
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q25.sh 
+Submitted batch job 1914461
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q26.sh 
+Submitted batch job 1914462
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q27.sh 
+Submitted batch job 1914463
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q28.sh 
+Submitted batch job 1914464
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q29.sh 
+Submitted batch job 1914465
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV0Q30.sh 
+Submitted batch job 1914466
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q20.sh 
+Submitted batch job 1914467
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q21.sh 
+Submitted batch job 1914468
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q22.sh 
+Submitted batch job 1914469
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q23.sh 
+Submitted batch job 1914470
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q24.sh 
+Submitted batch job 1914471
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q25.sh 
+Submitted batch job 1914472
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q26.sh 
+Submitted batch job 1914473
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q27.sh 
+Submitted batch job 1914474
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q28.sh 
+Submitted batch job 1914475
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q29.sh 
+Submitted batch job 1914476
+[cchoutou@aristotle6 homework_3]$ sbatch --qos=small -p ampere hpc/BitncV1Q30.sh 
+Submitted batch job 1914477
+

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q20-1914456.out

@@ -0,0 +1,2 @@
+[Timing] Total: 5920 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q21-1914457.out

@@ -0,0 +1,2 @@
+[Timing] Total: 6571 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q22-1914458.out

@@ -0,0 +1,2 @@
+[Timing] Total: 13 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q23-1914459.out

@@ -0,0 +1,2 @@
+[Timing] Total: 24 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q24-1914460.out

@@ -0,0 +1,2 @@
+[Timing] Total: 46 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q25-1914461.out

@@ -0,0 +1,2 @@
+[Timing] Total: 92 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q26-1914462.out

@@ -0,0 +1,2 @@
+[Timing] Total: 213 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q27-1914463.out

@@ -0,0 +1,2 @@
+[Timing] Total: 440 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q28-1914464.out

@@ -0,0 +1,2 @@
+[Timing] Total: 935 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q29-1914465.out

@@ -0,0 +1,2 @@
+[Timing] Total: 1847 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV0Q30-1914466.out

@@ -0,0 +1,2 @@
+[Timing] Total: 3798 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q20-1914467.out

@@ -0,0 +1,2 @@
+[Timing] Total: 2843 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q21-1914468.out

@@ -0,0 +1,2 @@
+[Timing] Total: 4979 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q22-1914469.out

@@ -0,0 +1,2 @@
+[Timing] Total: 9909 [usec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q23-1914470.out

@@ -0,0 +1,2 @@
+[Timing] Total: 20 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q24-1914471.out

@@ -0,0 +1,2 @@
+[Timing] Total: 35 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q25-1914472.out

@@ -0,0 +1,2 @@
+[Timing] Total: 70 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q26-1914473.out

@@ -0,0 +1,2 @@
+[Timing] Total: 170 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q27-1914474.out

@@ -0,0 +1,2 @@
+[Timing] Total: 346 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q28-1914475.out

@@ -0,0 +1,2 @@
+[Timing] Total: 735 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q29-1914476.out

@@ -0,0 +1,2 @@
+[Timing] Total: 1522 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/b31ca23/slurm-ampere-BitncV1Q30-1914477.out

@@ -0,0 +1,2 @@
+[Timing] Total: 2950 [msec]
+[Validation] Results validation ... [PASSED] 

homework_3/analyse/prof.sh

@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+
+#
+# prof.sh <exec> <report.file>
+#
+
+sudo /usr/local/cuda-11.4/bin/ncu \
+     --target-processes all       \
+     --metrics "$(echo -n \
+"smsp__inst_executed,"\
+"smsp__cycles_active.avg,"\
+"smsp__cycles_active.sum,"\
+"gpu__time_duration.sum,"\
+"smsp__average_warp_latency_issue_stalled_barrier,"\
+"smsp__warp_issue_stalled_barrier_per_warp_active,"\
+"l1tex__average_t_sectors_per_request_pipe_lsu_mem_global_op_ld,"\
+"l1tex__average_t_sectors_per_request_pipe_lsu_mem_global_op_st,"\
+"l1tex__data_pipe_lsu_wavefronts_mem_shared_cmd_read,"\
+"l1tex__data_pipe_lsu_wavefronts_mem_shared_cmd_write,"\
+"l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_ld.sum,"\
+"l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_st.sum "\
+     )" \
+     "$1" -q 20 -b 512 > "$2"

homework_3/hpc/submitJobs.sh

@@ -2,8 +2,8 @@
 
 # Submission parameters
 QOS="small"
-PARTITION="ampere"
+PARTITION="ampere"  # ampere 	gpu
-SCRIPT_DIR="hpc"  # Directory containing the job scripts
+SCRIPT_DIR="hpc"    # Directory containing the job scripts
 
 # Range of values for the -q parameter
 VERSIONS=("V0" "V1" "V2")
@@ -17,8 +17,9 @@ for version in "${VERSIONS[@]}"; do
     script_path="${SCRIPT_DIR}/${script_name}"
 
     if [[ -f "$script_path" ]]; then
+      echo "Submitting: $script_path"
       sbatch --qos="$QOS" -p "$PARTITION" "$script_path"
-      echo "Submitted: $script_path"
+      #sbatch -p "$PARTITION" "$script_path"
     else
       echo "Warning: File not found - $script_path"
     fi

homework_3/src/bitonicsort.hpp

@@ -15,6 +15,7 @@
 #include <cmath>
 #include <cstdint>
 #include <utility>
+#include <stdexcept>
 
 #include "utils.hpp"
 
@@ -159,9 +160,9 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMalloc(&dev_data, size * sizeof(value_t)) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not allocate memory\n");
+        throw std::runtime_error("[CUDA] - Can not allocate memory");
     if (cudaMemcpy(dev_data, data.data(), size * sizeof(value_t), cudaMemcpyHostToDevice) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory to device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory to device");
     Timer_memory.stop();
 
     size_t Nth = config.blockSize;
@@ -180,7 +181,7 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMemcpy(data.data(), dev_data, size * sizeof(value_t), cudaMemcpyDeviceToHost) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory from device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory from device");
     cudaFree(dev_data);
     Timer_memory.stop();
 }
@@ -247,6 +248,31 @@ __global__ void inBlockStep(ValueT* data, size_t n, size_t innerSteps, size_t st
     }
 }
 
+
+/*!
+ * This is unrolled part of the bitonic double loop for the init phase where the entire
+ * double loop can fit in one block with shared memory access.
+ *
+ * First each thread caches its corresponding data point from the current and the following data block.
+ * After that we execute the pre-phase on the local data and then we write back to global memory.
+ *
+ * @tparam ValueT   The underlying data type of the array items
+ * @param data      [ValueT*]   Pointer to data array
+ * @param n         [size_t]    The total size of the array
+ * @param stages    [size_t]    The number of stages to pre execute
+ * @param maxStages [size_t]    The maximum number of stages for the entire sort
+ */
+template <typename ValueT>
+__global__ void prephase(ValueT* data, size_t n, size_t stages, size_t maxStages) {
+    for (size_t stage = 1; (stage <= stages) && (stage <= maxStages); ++stage) {
+        for (size_t step = stage; step > 0; ) {
+            --step;
+            interBlockStep_(data, n, step, stage);
+            __syncthreads();
+        }
+    }
+}
+
 /*!
  * A CUDA version of the Bitonic sort algorithm.
  *
@@ -262,18 +288,22 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMalloc(&dev_data, size * sizeof(value_t)) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not allocate memory\n");
+        throw std::runtime_error("[CUDA] - Can not allocate memory");
     if (cudaMemcpy(dev_data, data.data(), size * sizeof(value_t), cudaMemcpyHostToDevice) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory to device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory to device");
     Timer_memory.stop();
 
     size_t Nth = config.blockSize;
     size_t Nbl = NBlocks(size);
 
     auto Stages          = static_cast<size_t>(log2(size));
-    auto InnerBlockSteps = static_cast<size_t>(log2(Nth));    //
+    auto InnerBlockSteps = static_cast<size_t>(log2(Nth));
+    size_t PrephaseStages= InnerBlockSteps + 1;
+
     Timer_sorting.start();
-    for (size_t stage = 1; stage <= Stages; ++stage) {
+    prephase<<<Nbl, Nth>>>(dev_data, size, PrephaseStages, Stages);
+    cudaDeviceSynchronize();
+    for (size_t stage = PrephaseStages + 1; stage <= Stages; ++stage) {
         size_t step = stage - 1;
         for ( ; step > InnerBlockSteps; --step) {
             interBlockStep<<<Nbl, Nth>>>(dev_data, size, step, stage);
@@ -286,7 +316,7 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMemcpy(data.data(), dev_data, size * sizeof(value_t), cudaMemcpyDeviceToHost) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory from device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory from device");
     cudaFree(dev_data);
     Timer_memory.stop();
 }
@@ -301,7 +331,9 @@ void bitonicSort(DataT& data) {
  * @note
  *  Each block thread collection can exchange twice the size of data points.
  */
-inline size_t effectiveBlockSize() { return SizeToThreadsRatio * config.blockSize; }
+inline constexpr size_t effectiveBlockSize(size_t blockSize) {
+    return SizeToThreadsRatio * blockSize;
+}
 
 
 
@@ -400,10 +432,70 @@ __global__ void inBlockStep(ValueT* data, size_t n, size_t innerSteps, size_t st
         __syncthreads();
     }
 
-    // Write back to global memory
+    // Write back to global memory (no sync here, there will be sync from host)
     data[gIdx0]              = shared_data[lIdx0];
     data[gIdx0 + blockDim.x] = shared_data[lIdx0 + blockDim.x];
+}
+
+/*!
+ * This is unrolled part of the bitonic double loop for the init phase where the entire
+ * double loop can fit in one block with shared memory access.
+ *
+ * First each thread caches its corresponding data point from the current and the following data block.
+ * After that we execute the pre-phase on the local data and then we write back to global memory.
+ *
+ * @tparam ValueT   The underlying data type of the array items
+ * @param data      [ValueT*]   Pointer to data array
+ * @param n         [size_t]    The total size of the array
+ * @param stages    [size_t]    The number of stages to pre execute
+ * @param maxStages [size_t]    The maximum number of stages for the entire sort
+ */
+template <typename ValueT>
+__global__ void prephase(ValueT* data, size_t n, size_t stages, size_t maxStages) {
+    extern __shared__ ValueT shared_data[];
+
+    /*
+     * Global and local(shared) memory indices (calculated once)
+     * Here we skip blocks every time (one for SizeToThreadsRatio = 2)
+     * And we cache the neighbor block address indexes in local (shared) memory
+     */
+    threadId_t gIdx0 = threadIdx.x + SizeToThreadsRatio * blockIdx.x * blockDim.x;
+    threadId_t lIdx0 = toLocal(gIdx0, blockDim.x);
+
+    if (gIdx0 + blockDim.x >= n)   // Boundary check
+        return;
+
+    // Fetch to local memory the entire effective block size (2 positions for each thread)
+    shared_data[lIdx0]              = data[gIdx0];
+    shared_data[lIdx0 + blockDim.x] = data[gIdx0 + blockDim.x];
     __syncthreads();
+    for (size_t stage = 1; (stage <= stages) && (stage <= maxStages); ++stage) {
+        for (size_t step = stage; step > 0; ) {
+            --step;
+
+            // Init thread global and local indices
+            threadId_t gIdx = gIdx0;
+            threadId_t lIdx = lIdx0;
+            // Find partner and keep-small configuration based on the global data positions
+            threadId_t pIdx  = partner(gIdx, step);
+            if (gIdx > pIdx) {
+                // Shift inside effective block
+                gIdx += blockDim.x; // global
+                pIdx += blockDim.x;
+                lIdx += blockDim.x;	// local
+            }
+            bool keep = keepSmall(gIdx, pIdx, stage);
+
+            // Exchange data on local(shared) copy
+            threadId_t lpIdx = toLocal(pIdx, blockDim.x);
+            exchange(shared_data, lIdx, lpIdx, keep);
+            __syncthreads();
+        }
+    }
+
+    // Write back to global memory (no sync here, there will be sync from host)
+    data[gIdx0]              = shared_data[lIdx0];
+    data[gIdx0 + blockDim.x] = shared_data[lIdx0 + blockDim.x];
 }
 
 /*!
@@ -421,19 +513,23 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMalloc(&dev_data, size * sizeof(value_t)) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not allocate memory\n");
+        throw std::runtime_error("[CUDA] - Can not allocate memory");
     if (cudaMemcpy(dev_data, data.data(), size * sizeof(value_t), cudaMemcpyHostToDevice) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory to device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory to device");
     Timer_memory.stop();
 
     size_t Nth = config.blockSize;
     size_t Nbl = NBlocks(size);
-    size_t kernelMemSize = effectiveBlockSize() * sizeof(value_t);
+    size_t kernelMemSize = effectiveBlockSize(config.blockSize) * sizeof(value_t);
 
     auto Stages          = static_cast<size_t>(log2(size));
     auto InnerBlockSteps = static_cast<size_t>(log2(Nth));
+    size_t PrephaseStages= InnerBlockSteps + 1;
+
     Timer_sorting.start();
-    for (size_t stage = 1; stage <= Stages; ++stage) {
+    prephase<<<Nbl, Nth, kernelMemSize>>>(dev_data, size, PrephaseStages, Stages);
+    cudaDeviceSynchronize();
+    for (size_t stage = PrephaseStages + 1; stage <= Stages; ++stage) {
         size_t step = stage - 1;
         for ( ; step > InnerBlockSteps; --step) {
             interBlockStep<<<Nbl, Nth>>>(dev_data, size, step, stage);
@@ -446,7 +542,7 @@ void bitonicSort(DataT& data) {
 
     Timer_memory.start();
     if (cudaMemcpy(data.data(), dev_data, size * sizeof(value_t), cudaMemcpyDeviceToHost) != cudaSuccess)
-        throw std::runtime_error("[CUDA] - Can not copy memory from device\n");
+        throw std::runtime_error("[CUDA] - Can not copy memory from device");
     cudaFree(dev_data);
     Timer_memory.stop();
 }

homework_3/src/config.h

@@ -65,8 +65,8 @@ using ArraySize_t = uint64_t;
  *  The values of the members are set from the command line.
  */
 struct config_t {
-	ArraySize_t	arraySize{DEFAULT_DATA_SIZE};   //!< The array size of the local data to sort.
+    ArraySize_t	arraySize{DEFAULT_DATA_SIZE};   //!< The array size of the local data to sort.
-	size_t      blockSize{THREADS_PER_BLOCK};	//!< The block size (threads per block) for the session.
+    size_t      blockSize{THREADS_PER_BLOCK};	//!< The block size (threads per block) for the session.
     bool        validation{false};              //!< Request a full validation at the end, performed by process rank 0.
     size_t      perf{1};                        //!< Enable performance timing measurements and prints. Repeat
                                                 //!< the sorting <perf> times to do so.

homework_3/src/main.cpp

@@ -8,6 +8,7 @@
  */
 
 #include <exception>
+#include <stdexcept>
 #include <iostream>
 #include <algorithm>
 #include <random>
@@ -140,14 +141,14 @@ bool get_options(int argc, char* argv[]){
 
     // Check configuration requirements
     if (config.blockSize % device.warpSize)
-    	throw std::runtime_error("[Config] - Number of threads per block is not an exact multiple of warp size\n");
+    	throw std::runtime_error("[Config] - Number of threads per block is not an exact multiple of warp size");
     if (config.arraySize < 2*config.blockSize)
     	throw std::runtime_error("[Config] - Unsupported array size (smaller than "
-                                 + std::to_string(SizeToThreadsRatio*config.blockSize) + ")\n");
+                                 + std::to_string(SizeToThreadsRatio*config.blockSize) + ")");
     if (device.totalGlobalMem < config.arraySize * sizeof(Value_t))
     	throw std::runtime_error("[CUDA] - Unsupported array size: "
                                  + std::to_string(config.arraySize * sizeof(Value_t))
-                                 + " (larger than GPU's: " + std::to_string(device.totalGlobalMem) + ")\n");
+                                 + " (larger than GPU's: " + std::to_string(device.totalGlobalMem) + ")");
 
     return status;
 }
@@ -197,6 +198,7 @@ int main(int argc, char* argv[]) try {
     // Init everything
     init(&argc, &argv);
 
+    logger << "Code version:   " << 'V' << STR(CODE_VERSION) << logger.endl;
     logger << "Array size:     " << config.arraySize << " (Q=" << static_cast<size_t>(log2(config.arraySize))<< ")" << logger.endl;
     logger << "Repeated sorts: " << config.perf << logger.endl;
     logger << "GPU:            " << device.name << logger.endl;
@@ -213,7 +215,7 @@ int main(int argc, char* argv[]) try {
         logger << " Done." << logger.endl;
 
         // Run distributed sort
-        logger << "Start sorting ...    ";
+        logger << "Start sorting    ... ";
         Timer_total.start();
         bitonicSort(Data);
         Timer_total.stop();