Version 0 of Tcl invoke performance

See also Tcl IO performance, where we reached the

Interim conclusion The time spent in the original benchmark seems to be decomposable in

• 8% in startup and loop overhead (0.19/2.5)
• 20% in command call and setting a variable in the called command (0.50/2.50) Further experimentation with a C-coded command that just returns TCL_OK shows that the cost of setting the variable is almost negligible
• 72% doing the file access, reading and conversion within gets itself
• the cost of setting the variable in the called command seems to be below the measurable threshold for the larger dataset (read1a actually faster than read2 - the difference must be noise)

Further experiments

• on a core instrumented with code developed together with GPS
• running a tiny do-almost-nothing script
• using a command empty defined in C to just return Tcl_Ok;

provide the following timings for the most-exercised opcodes (measured in cpu ticks at 1.6MHz):

8.5a6 time: 2481129141, count: 20010476
----------------------------------------------------
op         %T       avgT         %ops         Nops
6       43.53    1079.24         5.00      1000825      INST_INVOKE_STK1
80      17.47     433.44         5.00      1000050      INST_LIST_INDEX
105      7.16      44.37        19.99      4000995      INST_START_CMD
29       7.05      87.45        10.00      2000096      INST_INCR_SCALAR1_IMM
10       6.26      31.07        24.99      5001552      INST_LOAD_SCALAR1
103      5.18     128.60         5.00      1000008      INST_LIST_INDEX_IMM
17       4.26      52.80        10.00      2000511      INST_STORE_SCALAR1
47       3.71      92.07         5.00      1000001      INST_LT
1        3.30      40.83        10.01      2002404      INST_PUSH1
3        2.00      49.66         5.00      1000640      INST_POP