Skip to content

Instantly share code, notes, and snippets.

@cyyself
Last active July 18, 2024 03:52
Show Gist options
  • Save cyyself/01cb0c6158c0139bb47235a5ba671a54 to your computer and use it in GitHub Desktop.
Save cyyself/01cb0c6158c0139bb47235a5ba671a54 to your computer and use it in GitHub Desktop.

XiangShan: 136f64975e337409416bed849b58026e7085d758

Verilator: 0a9b31bb30fd604ff68854290cc5cc90e7dfd172

Build:

make emu MFC=1 EMU_THREADS=8 NUM_CORES=1 PGO_WORKLOAD=`realpath ready-to-run/coremark-2-iteration.bin` PGO_EMU_ARGS=--no-diff LLVM_PROFDATA=llvm-profdata -j `nproc`

Hardware: Intel Core i9 13900K

Run: numactl --physcpubind=0-15 perf stat -e L1-dcache-loads -e L1-dcache-load-misses -e LLC-load-misses -e L1-icache-load-misses -e branch-instructions -e branch-misses -e cycles -e instructions sh -c './build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null'

llvm-mod-e63ee389e8b7edcdff981070585ea7e752b40d73:

 Performance counter stats for 'sh -c ./build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null':

     <not counted>      cpu_atom/L1-dcache-loads/                                               (0.00%)
   454,710,431,927      cpu_core/L1-dcache-loads/                                               (87.50%)
   <not supported>      cpu_atom/L1-dcache-load-misses/                                       
    15,220,909,809      cpu_core/L1-dcache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/LLC-load-misses/                                               (0.00%)
           555,797      cpu_core/LLC-load-misses/                                               (87.50%)
     <not counted>      cpu_atom/L1-icache-load-misses/                                         (0.00%)
   121,624,224,296      cpu_core/L1-icache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/branch-instructions/                                           (0.00%)
   150,335,058,323      cpu_core/branch-instructions/                                           (87.50%)
     <not counted>      cpu_atom/branch-misses/                                                 (0.00%)
     8,668,828,278      cpu_core/branch-misses/                                                 (87.50%)
     <not counted>      cpu_atom/cycles/                                                        (0.00%)
 3,117,663,903,436      cpu_core/cycles/                                                        (87.50%)
     <not counted>      cpu_atom/instructions/                                                  (0.00%)
 1,623,276,069,859      cpu_core/instructions/                                                  (87.50%)

      78.061620370 seconds time elapsed

     624.038322000 seconds user
       0.019998000 seconds sys

llvm-809855b56f06dd7182685f88fbbc64111df9339a:

 Performance counter stats for 'sh -c ./build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null':

     <not counted>      cpu_atom/L1-dcache-loads/                                               (0.00%)
   469,726,618,942      cpu_core/L1-dcache-loads/                                               (87.50%)
   <not supported>      cpu_atom/L1-dcache-load-misses/                                       
    15,538,301,530      cpu_core/L1-dcache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/LLC-load-misses/                                               (0.00%)
           394,060      cpu_core/LLC-load-misses/                                               (87.50%)
     <not counted>      cpu_atom/L1-icache-load-misses/                                         (0.00%)
   120,546,599,257      cpu_core/L1-icache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/branch-instructions/                                           (0.00%)
   124,636,184,648      cpu_core/branch-instructions/                                           (87.50%)
     <not counted>      cpu_atom/branch-misses/                                                 (0.00%)
     1,082,610,311      cpu_core/branch-misses/                                                 (87.51%)
     <not counted>      cpu_atom/cycles/                                                        (0.00%)
 1,697,731,463,511      cpu_core/cycles/                                                        (87.51%)
     <not counted>      cpu_atom/instructions/                                                  (0.00%)
 1,640,896,750,339      cpu_core/instructions/                                                  (87.51%)

      39.566765608 seconds time elapsed

     316.053698000 seconds user
       0.031998000 seconds sys

llvm-603c286334b07f568d39f6706c848f576914f323:

 Performance counter stats for 'sh -c ./build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null':

     <not counted>      cpu_atom/L1-dcache-loads/                                               (0.00%)
   457,302,855,831      cpu_core/L1-dcache-loads/                                               (87.50%)
   <not supported>      cpu_atom/L1-dcache-load-misses/                                       
    14,879,466,301      cpu_core/L1-dcache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/LLC-load-misses/                                               (0.00%)
         1,073,389      cpu_core/LLC-load-misses/                                               (87.50%)
     <not counted>      cpu_atom/L1-icache-load-misses/                                         (0.00%)
   122,062,056,955      cpu_core/L1-icache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/branch-instructions/                                           (0.00%)
   160,629,490,386      cpu_core/branch-instructions/                                           (87.50%)
     <not counted>      cpu_atom/branch-misses/                                                 (0.00%)
    10,492,102,956      cpu_core/branch-misses/                                                 (87.50%)
     <not counted>      cpu_atom/cycles/                                                        (0.00%)
 3,544,457,281,488      cpu_core/cycles/                                                        (87.50%)
     <not counted>      cpu_atom/instructions/                                                  (0.00%)
 1,626,453,829,111      cpu_core/instructions/                                                  (87.50%)

      80.681841158 seconds time elapsed

     644.997805000 seconds user
       0.043997000 seconds sys

llvm-mod-73d693ca2332c6b265153b97dcb0bc1896820b1e

 Performance counter stats for 'sh -c ./build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null':

     <not counted>      cpu_atom/L1-dcache-loads/                                               (0.00%)
   463,763,498,705      cpu_core/L1-dcache-loads/                                               (87.50%)
   <not supported>      cpu_atom/L1-dcache-load-misses/                                       
    15,203,567,702      cpu_core/L1-dcache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/LLC-load-misses/                                               (0.00%)
           856,470      cpu_core/LLC-load-misses/                                               (87.50%)
     <not counted>      cpu_atom/L1-icache-load-misses/                                         (0.00%)
   118,554,991,629      cpu_core/L1-icache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/branch-instructions/                                           (0.00%)
   119,772,615,935      cpu_core/branch-instructions/                                           (87.50%)
     <not counted>      cpu_atom/branch-misses/                                                 (0.00%)
     1,044,652,584      cpu_core/branch-misses/                                                 (87.50%)
     <not counted>      cpu_atom/cycles/                                                        (0.00%)
 1,648,766,988,422      cpu_core/cycles/                                                        (87.51%)
     <not counted>      cpu_atom/instructions/                                                  (0.00%)
 1,575,416,736,769      cpu_core/instructions/                                                  (87.51%)

      37.557306681 seconds time elapsed

     299.974017000 seconds user
       0.027997000 seconds sys

llvm-fe6c24000f2d7316899d4ec4c12273892326ed47

 Performance counter stats for 'sh -c ./build/emu -i ready-to-run/coremark-2-iteration.bin --no-diff 2>/dev/null':

     <not counted>      cpu_atom/L1-dcache-loads/                                               (0.00%)
   449,932,804,737      cpu_core/L1-dcache-loads/                                               (87.49%)
   <not supported>      cpu_atom/L1-dcache-load-misses/
    15,161,689,652      cpu_core/L1-dcache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/LLC-load-misses/                                               (0.00%)
           659,000      cpu_core/LLC-load-misses/                                               (87.50%)
     <not counted>      cpu_atom/L1-icache-load-misses/                                         (0.00%)
   122,921,964,627      cpu_core/L1-icache-load-misses/                                         (87.50%)
     <not counted>      cpu_atom/branch-instructions/                                           (0.00%)
   166,382,809,161      cpu_core/branch-instructions/                                           (87.50%)
     <not counted>      cpu_atom/branch-misses/                                                 (0.00%)
     3,348,296,167      cpu_core/branch-misses/                                                 (87.50%)
     <not counted>      cpu_atom/cycles/                                                        (0.00%)
 2,275,690,285,047      cpu_core/cycles/                                                        (87.50%)
     <not counted>      cpu_atom/instructions/                                                  (0.00%)
 1,601,347,645,672      cpu_core/instructions/                                                  (87.50%)

      53.020458750 seconds time elapsed

     423.706806000 seconds user
       0.027997000 seconds sy
@cyyself
Copy link
Author

cyyself commented Jul 17, 2024

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment