GPU_PTX 3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs "Mgl-d-Rs-Wg Rf-w-g Po-d-Rg-Ws Rf-w-s Ctrl-d-s-Ws Rf-w-s" {} P0 | P1 | P2 ; ld.cg.s32 r0, [r1] | ld.cg.s32 r0, [r1] | ld.cg.s32 r0, [r1] ; membar.gl | mov.s32 r2, 1 | and.b32 r2, r0 2147483648 ; mov.s32 r2, 1 | st.cg.s32 [r3], r2 | setp.eq.b32 p2, r2, 0 ; st.cg.s32 [r3], r2 | | @p2 bra LC00 ; | | mov.s64 r4, 3494 ; | | LC00: ; | | mov.s32 r3, 1 ; | | st.cg.s32 [r4], r3 ; exists (0:r0=1 /\ 1:r0=1 /\ 2:r0=1) (device (kernel (cta (warp P0) (warp P1) (warp P2)))) x:local; y:global; z:local Generated assembler //START _litmus_P2 ld.cg.s32 sr02, [sr12] ; and.b32 sr22, sr02, 2147483648 ; setp.eq.b32 sp22, sr22, 0 ; @sp22 bra LC00 ; mov.s64 sr42, 3494 ; LC00: mov.s32 sr32, 1 ; st.cg.s32 [sr42], sr32 ; //START _litmus_P1 ld.cg.s32 sr01, [sr11] ; mov.s32 sr21, 1 ; st.cg.s32 [sr31], sr21 ; //START _litmus_P0 ld.cg.s32 sr00, [sr10] ; membar.gl ; mov.s32 sr20, 1 ; st.cg.s32 [sr30], sr20 ; Binary '/tmp/dire79b59.tmp/3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs-0.exe' successfully loaded Specification clusters: 3 Specification: [[(698, 3, 'R4'), (696, 2, 'R12')], [(754, 3, 'R4'), (751, 2, 'R10')], [(812, 3, 'R4'), (810, 2, 'R12')]] Cluster 0: OK Cluster 1: OK Cluster 2: OK !!SUCCESS!! Test 3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs Allowed Histogram (4 states) 36817 :>0:r0=0; 1:r0=0; 2:r0=0; 7038 :>0:r0=1; 1:r0=0; 2:r0=0; 5997 :>0:r0=0; 1:r0=0; 2:r0=1; 148 :>0:r0=1; 1:r0=0; 2:r0=1; No Witnesses Positive: 0, Negative: 50000 Condition exists (0:r0=1 /\ 1:r0=1 /\ 2:r0=1) is NOT validated Observation 3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs Never 0 50000 Time 3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs 13.72 -- device info -- Name: GeForce GTX TITAN Major revision number: 3 Minor revision number: 5 Total global memory: 6442123264 Total shared memory per block: 49152 Number of multiprocessors: 14 Machine: formal Revision 1383, version 5.01 GPU_PTX experimental backend Command line: ./litmus -mach GTXTitan.cfg -a 8 -gpu-verf-comp true big_bucket/3.LB+membar.gl+po+ctrl-p0:p1:p2-xsygzs-0.litmus -s 50000 Parameters #ifndef SIZE_OF_TEST #define SIZE_OF_TEST 50000 #endif #ifndef NUMBER_OF_RUN #define NUMBER_OF_RUN 1 #endif /* cc options: -arch=sm_35 */ /* barrier: user */ /* speedcheck: no */ /* gpu_general_bc: true */ /* gpu_mem_stress: true */ /* gpu-rand-threads: true */ /* gpu_cuda: true */ /* gpu_gen_addr: true */ /* gpu_cvt_global_to_gen: true */ /* gpu_cvt_shared_to_gen: true */ /* gpu_max_ctas: 14 */ /* gpu_min_ctas: 10 */ /* gpu_max_threads: 512 */ /* gpu_min_threads: 128 */