Kampfsau FFdecsa # ./optimizer.sh
### FFdecsa optimization helper/benchmark
### Version 9d
Warning: No valid path specified
Searching...
Found possible FFdecsa source at:
/home/laurent/vdr-plugin-sc-1.0.0+hg20110429/FFdecsa
...and source files looks valid.
proceed...
### CPU-INFO ###
System: x86_64
Auto detected arch: core2
Vendor-ID: GenuineIntel
CPU-Family: 6
CPU-Model: 15
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm lahf_lm dts tpr_shadow vnmi flexpriority
gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-8ubuntu4)
Using compilers "native" flags
### FFdeCSA TEST ###
Using compiler: g++
Flags: -march=native -fPIC -fexpensive-optimizations -fomit-frame-pointer -funroll-loops
Testing optimization levels 2 and 3
Level -O2:
PARALLEL_32_INT
- 204, 239, 239, 218, 184, 236, 235, 239, 230, 179
- 239 Mbit/s max.
PARALLEL_64_2INT
- 166, 211, 169, 216, 219, 218, 217, 217, 197, 165
- 219 Mbit/s max.
PARALLEL_64_LONG
- 279, 348, 353, 265, 259, 293, 317, 261, 314, 348
- 353 Mbit/s max.
PARALLEL_64_MMX
- 280, 275, 329, 368, 364, 370, 367, 374, 294, 277
- 374 Mbit/s max.
PARALLEL_128_2LONG
- 283, 307, 252, 257, 334, 340, 341, 338, 270, 254
- 341 Mbit/s max.
PARALLEL_128_2MMX
- 247, 327, 332, 257, 313, 322, 250, 250, 247, 253
- 332 Mbit/s max.
PARALLEL_128_SSE
- 376, 468, 479, 400, 393, 476, 476, 481, 460, 479
- 481 Mbit/s max.
PARALLEL_128_SSE2
- 405, 394, 404, 398, 458, 529, 520, 529, 529, 536
- 536 Mbit/s max.
Fastest PARALLEL_MODE = PARALLEL_128_SSE2 (536 Mbit/s)
Level -O3:
PARALLEL_32_INT
...failed!
PARALLEL_64_2INT
...failed!
PARALLEL_64_LONG
- 227, 270, 263, 205, 203, 201, 203, 260, 270, 270
- 270 Mbit/s max.
PARALLEL_64_MMX
- 228, 291, 291, 291, 232, 282, 289, 286, 287, 265
- 291 Mbit/s max.
PARALLEL_128_2LONG
- 262, 345, 347, 261, 261, 259, 262, 257, 345, 347
- 347 Mbit/s max.
PARALLEL_128_2MMX
- 271, 344, 334, 257, 257, 306, 278, 320, 336, 344
- 344 Mbit/s max.
PARALLEL_128_SSE
- 365, 434, 486, 371, 492, 407, 370, 388, 482, 380
- 492 Mbit/s max.
PARALLEL_128_SSE2
- 411, 416, 417, 418, 507, 553, 543, 550, 555, 549
- 555 Mbit/s max.
Fastest PARALLEL_MODE = PARALLEL_128_SSE2 (555 Mbit/s)
Best result with -O3 and PARALLEL_128_SSE2 at 555 Mbit/s
### VDR-SC Makefile FFdeCSA OPTS ###
CPUOPT ?= native
PARALLEL ?= PARALLEL_128_SSE2
CSAFLAGS ?= -O3 -fPIC -fexpensive-optimizations -fomit-frame-pointer -funroll-loops
### GENERIC FFdeCSA make OPTS ###
FLAGS="-O3 -march=native -fPIC -fexpensive-optimizations -fomit-frame-pointer -funroll-loops" PARALLEL_MODE=PARALLEL_128_S