## ncu_6144x5120.ncu-rep
served_M8    void exl3_gemm_kernel<4, 0, 2, 16, 32, 128, 4, 3> dur=  24.26us dram= 66.8% (  655GB/s rd=  15.9MB) sm= 46.2% issue= 46.2% tensor= 13.4% alu= 46.5% fma= 18.4% occ= 33.3/  33% l2hit= 25.6% regs=107 smem= 91.0K grid=128x512 bankcf=14318 stalls=barrier:2.0,wait:1.3,math_pipe_throttle:1.2,selected:1.0
mma_M8       void ag_mma_kernel<4, 2, 1, 1, 1>(const void *, const unsign dur=  24.74us dram= 65.6% (  643GB/s rd=  15.9MB) sm= 36.9% issue= 36.9% tensor=  6.7% alu= 37.7% fma= 15.7% occ= 58.8/  67% l2hit= 33.9% regs=64 smem=  9.0K grid=480x256 bankcf=261120 stalls=long_scoreboard:9.9,wait:1.6,not_selected:1.3,selected:1.0
up_gemv_M8   void exl3_gemv_kernel<4, 0, 2, 1, 0, 0> dur=  24.45us dram= 66.2% (  649GB/s rd=  15.9MB) sm= 28.0% issue= 28.0% tensor= 12.7% alu= 23.4% fma= 13.6% occ= 40.6/  67% l2hit= 46.5% regs=64 smem= 17.0K grid=160x512 bankcf=61440 stalls=long_scoreboard:8.3,barrier:2.8,wait:1.6,selected:1.0
simt_M8      void ag_gemv_kernel<Exl3GemvLoader<4, 2, 8>, 4, 0, 8, 4>(con dur=  36.58us dram= 44.3% (  435GB/s rd=  15.9MB) sm= 80.6% issue= 56.2% tensor=  0.0% alu= 19.4% fma= 34.0% occ= 50.6/  67% l2hit= 28.0% regs=64 smem= 13.0K grid=640x256 bankcf=0 stalls=mio_throttle:2.7,not_selected:2.1,barrier:1.7,selected:1.0
simt_M8      void ag_gemv_reduce_kernel<Exl3GemvLoader<4, 2, 8>, 8> dur=   5.28us dram= 52.3% (  511GB/s rd=   2.7MB) sm=  4.6% issue=  4.6% tensor=  0.0% alu=  4.4% fma=  2.2% occ= 35.5/  83% l2hit= 25.1% regs=44 smem=  5.0K grid=320x256 bankcf=0 stalls=long_scoreboard:46.6,imc_miss:8.5,barrier:7.5,wait:2.2
i8_M8        void arbi_serve_exl3_a_prep::a_prep_kernel<1, 4, 1, 1, arbi_ dur=   8.13us dram=  1.6% (   16GB/s rd=   0.1MB) sm=  1.6% issue=  1.6% tensor=  0.0% alu=  6.5% fma=  2.9% occ= 13.3/ 100% l2hit= 80.7% regs=40 smem=  1.0K grid=192x128 bankcf=0 stalls=imc_miss:6.2,long_scoreboard:2.4,short_scoreboard:1.9,wait:1.4
i8_M8        void exl3_i8_gemm_kernel<4, 2, 16, 128, 128, 1, 8, 2, 8, 0,  dur=  31.62us dram= 51.1% (  501GB/s rd=  15.9MB) sm= 13.1% issue= 13.1% tensor= 14.9% alu= 37.5% fma= 19.1% occ= 16.6/  67% l2hit= 21.2% regs=58 smem= 21.0K grid=40x256 bankcf=0 stalls=long_scoreboard:1.0,selected:1.0,wait:0.8,not_selected:0.3
served_M16   void exl3_gemm_kernel<4, 0, 2, 16, 32, 128, 4, 3> dur=  25.66us dram= 63.5% (  623GB/s rd=  16.0MB) sm= 43.9% issue= 43.9% tensor= 12.9% alu= 45.2% fma= 17.9% occ= 33.0/  33% l2hit= 39.9% regs=107 smem= 91.0K grid=128x512 bankcf=28813 stalls=barrier:2.1,wait:1.3,math_pipe_throttle:1.2,selected:1.0
mma_M16      void ag_mma_prestage_kernel<2, 1>(const void *, const __half dur=   2.34us dram=  9.6% (   93GB/s rd=   0.2MB) sm=  5.7% issue=  5.7% tensor=  0.0% alu= 10.9% fma=  5.2% occ= 16.8/ 100% l2hit= 59.4% regs=24 smem=  1.0K grid=96x256 bankcf=0 stalls=long_scoreboard:5.4,imc_miss:3.5,wait:1.4,selected:1.0
mma_M16      void ag_mma_kernel<4, 2, 2, 1, 1>(const void *, const unsign dur=  24.74us dram= 65.9% (  647GB/s rd=  16.0MB) sm= 28.8% issue= 28.8% tensor= 14.6% alu= 31.5% fma= 12.7% occ= 59.5/  67% l2hit= 51.1% regs=64 smem= 17.0K grid=480x256 bankcf=95581 stalls=long_scoreboard:11.6,wait:1.6,math_pipe_throttle:1.4,barrier:1.1
i8_M16       void arbi_serve_exl3_a_prep::a_prep_kernel<1, 4, 1, 1, arbi_ dur=   2.62us dram=  9.0% (   86GB/s rd=   0.2MB) sm=  7.1% issue=  7.1% tensor=  0.0% alu=  7.7% fma=  3.8% occ= 12.1/ 100% l2hit= 72.8% regs=40 smem=  1.0K grid=192x128 bankcf=0 stalls=long_scoreboard:3.7,imc_miss:3.2,wait:1.3,short_scoreboard:1.3
i8_M16       void exl3_i8_gemm_kernel<4, 2, 16, 128, 128, 1, 8, 2, 8, 0,  dur=  31.14us dram= 51.9% (  509GB/s rd=  15.9MB) sm= 13.4% issue= 13.4% tensor= 15.2% alu= 38.3% fma= 19.6% occ= 16.3/  67% l2hit= 21.8% regs=58 smem= 21.0K grid=40x256 bankcf=0 stalls=long_scoreboard:1.1,selected:1.0,wait:0.8,short_scoreboard:0.3
served_M32   void exl3_gemm_kernel<4, 0, 2, 32, 32, 128, 4, 3> dur=  31.49us dram= 52.3% (  514GB/s rd=  16.2MB) sm= 38.8% issue= 38.8% tensor= 20.9% alu= 38.6% fma= 15.2% occ= 33.3/  33% l2hit= 56.0% regs=112 smem= 91.0K grid=128x512 bankcf=138871 stalls=barrier:2.7,math_pipe_throttle:2.1,wait:1.2,selected:1.0
mma_M32      void ag_mma_prestage_kernel<4, 1>(const void *, const __half dur=   2.53us dram= 16.8% (  164GB/s rd=   0.4MB) sm= 10.2% issue= 10.2% tensor=  0.0% alu= 14.5% fma=  6.8% occ= 24.1/ 100% l2hit= 58.3% regs=24 smem=  1.0K grid=192x256 bankcf=0 stalls=long_scoreboard:6.6,imc_miss:2.9,wait:1.4,selected:1.0
mma_M32      void ag_mma_kernel<4, 2, 4, 1, 1>(const void *, const unsign dur=  31.52us dram= 52.4% (  514GB/s rd=  16.2MB) sm= 25.1% issue= 25.1% tensor= 24.2% alu= 27.8% fma= 11.1% occ= 40.4/  50% l2hit= 66.9% regs=64 smem= 33.0K grid=480x256 bankcf=221004 stalls=long_scoreboard:5.6,math_pipe_throttle:2.9,wait:1.6,barrier:1.0
i8_M32       void arbi_serve_exl3_a_prep::a_prep_kernel<1, 4, 1, 1, arbi_ dur=   2.85us dram= 15.2% (  148GB/s rd=   0.4MB) sm= 12.8% issue= 12.8% tensor=  0.0% alu= 14.1% fma=  6.9% occ= 23.1/ 100% l2hit= 70.5% regs=40 smem=  1.0K grid=384x128 bankcf=0 stalls=long_scoreboard:4.4,imc_miss:2.9,wait:1.4,short_scoreboard:1.1
i8_M32       void exl3_i8_gemm_kernel<4, 2, 16, 128, 128, 1, 8, 2, 8, 0,  dur=  31.74us dram= 51.2% (  503GB/s rd=  16.0MB) sm= 26.6% issue= 26.6% tensor= 15.0% alu= 37.9% fma= 19.4% occ= 17.2/  67% l2hit= 60.4% regs=58 smem= 21.0K grid=80x256 bankcf=0 stalls=long_scoreboard:1.2,selected:1.0,wait:0.8,short_scoreboard:0.3
