Open vladkraievskiy opened 8 years ago
ha7ilm
commented
8 years ago Seems like csdr fails:
Illegal instruction (core dumped)
Illegal instruction (core dumped)Maybe it is a problem with compiler flags. Could you provide more information about your hardware (PC/embedded)?
Could you also post the output of the following:
cat /proc/cpuinfo...and the output of:
cd csdr
make
vladkraievskiy
commented
8 years ago pavel@pavel-System-Product-Name:~$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 16
model : 6
model name : AMD Athlon(tm) II X2 250 Processor
stepping : 2
microcode : 0x100009f
cpu MHz : 1800.000
cache size : 1024 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 2
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 5
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock nrip_save vmmcall
bugs : tlb_mmatch apic_c1e fxsave_leak
bogomips : 5999.82
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 48 bits physical, 48 bits virtual
power management: ts ttp tm stc 100mhzsteps hwpstate
processor : 1
vendor_id : AuthenticAMD
cpu family : 16
model : 6
model name : AMD Athlon(tm) II X2 250 Processor
stepping : 2
microcode : 0x100009f
cpu MHz : 800.000
cache size : 1024 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
apicid : 1
initial apicid : 1
fpu : yes
fpu_exception : yes
cpuid level : 5
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock nrip_save vmmcall
bugs : tlb_mmatch apic_c1e fxsave_leak
bogomips : 5999.82
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 48 bits physical, 48 bits virtual
power management: ts ttp tm stc 100mhzsteps hwpstate
pavel@pavel-System-Product-Name:~$ cd csdr/
pavel@pavel-System-Product-Name:~$ sudo make
rm -f dumpvect*.vect
NOTE: you may have to manually edit Makefile to optimize for your CPU (especially if you compile on ARM, please edit PARAMS_NEON).
Auto-detected optimization parameters: -msse -msse2 -msse4 -mfpmath=sse
c99 -O3 -ffast-math -fdump-tree-vect-details -dumpbase dumpvect -msse -msse2 -msse4 -mfpmath=sse fft_fftw.c libcsdr_wrapper.c -g -lm -lrt -lfftw3f -DUSE_FFTW -DLIBCSDR_GPL -DUSE_IMA_ADPCM -Wno-unused-result -fpic -shared -o libcsdr.so
./parsevect dumpvect*.vect
Auto-vectorization built into gcc can increase the execution speed of algorithms with automatic
generation of SIMD instructions if the CPU is capable.
We parse the output of the vectorizer to analyze which loops could be optimized (thus speeded up) this way.
Warning! The result may be different on different CPU architectures...
Colors:
- can't be vectorized
- successfully vectorized
- not intended to be vectorized (not important)
libcsdr.c:137: note: LOOP VECTORIZED. firdes_lowpass_f: normalize pass 2
libcsdr.c:133: note: LOOP VECTORIZED. firdes_lowpass_f: normalize pass 1
libcsdr.c:125: note: not vectorized: number of iterations cannot be computed. firdes_lowpass_f: calculate taps
libcsdr.c:155: note: not vectorized: multiple nested loops. firdes_bandpass_c
libcsdr.c:161: note: not vectorized: number of iterations cannot be computed. firdes_bandpass_c
libcsdr.c:160: note: not vectorized: number of iterations cannot be computed. firdes_bandpass_c
libcsdr.c:192: note: not vectorized: multiple nested loops. shift_math_cc
libcsdr.c:203: note: not vectorized: number of iterations cannot be computed. shift_math_cc: normalize phase
libcsdr.c:202: note: not vectorized: number of iterations cannot be computed. shift_math_cc: normalize phase
libcsdr.c:216: note: not vectorized: relevant stmt not supported: _18 = __builtin_sinf (_17);
libcsdr.c:236: note: not vectorized: multiple nested loops. shift_math_cc
libcsdr.c:261: note: not vectorized: number of iterations cannot be computed. shift_math_cc: normalize phase
libcsdr.c:260: note: not vectorized: number of iterations cannot be computed. shift_math_cc: normalize phase
libcsdr.c:279: note: not vectorized: multiple nested loops. fir_decimate_cc: i loop
libcsdr.c:281: note: LOOP VECTORIZED. fir_decimate_cc: q loop
libcsdr.c:279: note: LOOP VECTORIZED. fir_decimate_cc: i loop
libcsdr.c:300: note: not vectorized: inner-loop count not invariant.
libcsdr.c:306: note: LOOP VECTORIZED. rational_resampler_ff (inner loop)
libcsdr.c:360: note: LOOP VECTORIZED. fir_one_pass_ff
libcsdr.c:381: note: not vectorized: multiple nested loops. fractional_decimator_ff
libcsdr.c:360: note: LOOP VECTORIZED. fir_one_pass_ff
libcsdr.c:360: note: LOOP VECTORIZED. fir_one_pass_ff
libcsdr.c:360: note: LOOP VECTORIZED. fir_one_pass_ff
libcsdr.c:427: note: LOOP VECTORIZED. apply_fir_fft_cc: add overlap
libcsdr.c:421: note: LOOP VECTORIZED. apply_fir_fft_cc: normalize by fft_size
libcsdr.c:409: note: LOOP VECTORIZED. apply_fir_fft_cc: multiplication
libcsdr.c:453: note: LOOP VECTORIZED. amdemod: sqrt
libcsdr.c:448: note: LOOP VECTORIZED. amdemod: i*i+q*q
libcsdr.c:472: note: LOOP VECTORIZED. amdemod_estimator
libcsdr.c:495: note: not vectorized, possible dependence between data-refs *_23 and *_16 dcblock_f
libcsdr.c:519: note: LOOP VECTORIZED. fastdcblock_ff: remove DC component
libcsdr.c:511: note: LOOP VECTORIZED. fastdcblock_ff: calculate block average
libcsdr.c:557: note: LOOP VECTORIZED. fastagc_ff: apply gain
libcsdr.c:542: note: LOOP VECTORIZED. fastagc_ff: peak search
libcsdr.c:592: note: not vectorized: unsupported use in stmt. fmdemod_atan_novect
libcsdr.c:610: note: not vectorized: complicated access pattern. fmdemod_quadri_novect_cf
libcsdr.c:648: note: LOOP VECTORIZED. fmdemod_quadri_cf: output division
libcsdr.c:644: note: LOOP VECTORIZED. fmdemod_quadri_cf: output denomiator
libcsdr.c:640: note: LOOP VECTORIZED. fmdemod_quadri_cf: output numerator
libcsdr.c:635: note: LOOP VECTORIZED. fmdemod_quadri_cf: di
libcsdr.c:629: note: LOOP VECTORIZED. fmdemod_quadri_cf: dq
libcsdr.c:677: note: not vectorized, possible dependence between data-refs *_29 and *_24 deemphasis_wfm_ff
libcsdr.c:707: note: LOOP VECTORIZED. deemphasis_nfm_ff: inner loop
libcsdr.c:715: note: LOOP VECTORIZED. limit_ff
libcsdr.c:724: note: LOOP VECTORIZED. gain_ff
libcsdr.c:743: note: not vectorized: control flow in loop. log2n
libcsdr.c:758: note: not vectorized: control flow in loop. next_pow2
libcsdr.c:766: note: not vectorized: loop contains function calls or data references that cannot be analyzed apply_window_c
libcsdr.c:777: note: not vectorized: loop contains function calls or data references that cannot be analyzed apply_window_f
libcsdr.c:790: note: LOOP VECTORIZED. logpower_cf: pass 3
libcsdr.c:788: note: not vectorized: relevant stmt not supported: _33 = __builtin_log10f (_32); logpower_cf: pass 2
libcsdr.c:786: note: LOOP VECTORIZED. logpower_cf: pass 1
libcsdr.c:805: note: LOOP VECTORIZED. convert_u8_f
libcsdr.c:810: note: LOOP VECTORIZED. convert_s8_f
libcsdr.c:815: note: LOOP VECTORIZED. convert_i16_f
libcsdr.c:820: note: LOOP VECTORIZED. convert_f_u8
libcsdr.c:827: note: LOOP VECTORIZED. convert_f_s8
libcsdr.c:837: note: LOOP VECTORIZED. convert_f_i16
libcsdr.c:844: note: LOOP VECTORIZED. trivial_vectorize: should pass :-)
libcsdr_gpl.c:50: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: normalize starting_phase
libcsdr_gpl.c:49: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: normalize starting_phase
libcsdr_gpl.c:37: note: not vectorized: unsupported use in stmt. shift_addition_cc: work
libcsdr_gpl.c:77: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: work
libcsdr_gpl.c:77: note: not vectorized: Bad inner loop. shift_addition_cc: work
libcsdr_gpl.c:84: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: normalize phase
libcsdr_gpl.c:131: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: normalize starting_phase
libcsdr_gpl.c:130: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: normalize starting_phase
libcsdr_gpl.c:115: note: not vectorized: number of iterations cannot be computed. shift_addition_cc: work
libcsdr_gpl.c:171: note: not vectorized: control flow in loop. agc_ff
ima_adpcm.c:157: note: not vectorized: not suitable for gather load step_42 = _stepSizeTable[state$index_36];
ima_adpcm.c:168: note: not vectorized: not suitable for gather load step_34 = _stepSizeTable[state$index_30];
c99 -O3 -ffast-math -fdump-tree-vect-details -dumpbase dumpvect -msse -msse2 -msse4 -mfpmath=sse csdr.c -g -lm -lrt -lfftw3f -DUSE_FFTW -DLIBCSDR_GPL -DUSE_IMA_ADPCM -L. -lcsdr -Wno-unused-result -o csdr
bogstandard
commented
8 years ago I'm also facing similar issues on my AMD processor machine. After the localhost is declared open the next line is "illegal instruction".
I am running the very same OS and software as my Intel machine and I only get this error on the AMD Processor machine.
Are you able to say what would be the correct flags to have on the csdr makefile for an AMD Processor?
ha7ilm
commented
8 years ago A temporary solution for this problem would be removing the $(call cpufeature,sse4,-msse4) from the Makefile of csdr:
PARAMS_SSE = $(call cpufeature,sse,-msse) $(call cpufeature,sse2,-msse2) $(call cpufeature,sse3,-msse3) $(call cpufeature,sse4_1,-msse4.1) $(call cpufeature,sse4_2,-msse4.2) -mfpmath=sse (the problematic line here without the forced SSE4 support)
...and then recompiling and reinstalling csdr (with make && sudo make install).
Does it solve the problem for you?
US1GHQ
commented
8 years ago For me this solution is helped, Thanks
vladkraievskiy
commented
8 years ago Thanks, that solved a problem for me too.
cfenell
commented
5 years ago Hi, I am looking forward to testing this! However I also encounter csdr compilation errors and a program crash on my laptop, Intel Core i3-4030U. This is the cpuinfo and csdr make output: cpuinfo.txt makeout.txt
openwebrx.py prints out these errors: free(): double free detected in tcache 2 Aborted csdr squelch_and_smeter_cc: initial squelch level is 0 Signal caught, exiting! Short write, samples lost, exiting! [openwebrx-spectrum] client spectrum queue full, closing it. [openwebrx-httpd] client 127.0.0.1#6131c19b85d63f8acf1067cb05deeacb :: client being closed. [openwebrx-httpd:ws] client closed by other thread [openwebrx-dsp-plugin:csdr] try_delete_pipes() :: [Errno 2] No such file or directory: '/tmp/openwebrx_pipe_139709838640912_bpf_pipe' [openwebrx-dsp-plugin:csdr] try_delete_pipes() :: [Errno 2] No such file or directory: '/tmp/openwebrx_pipe_139709838640912_shift_pipe' [openwebrx-dsp-plugin:csdr] try_delete_pipes() :: [Errno 2] No such file or directory: '/tmp/openwebrx_pipe_139709838640912_squelch_pipe' [openwebrx-dsp-plugin:csdr] try_delete_pipes() :: [Errno 2] No such file or directory: '/tmp/openwebrx_pipe_139709838640912_smeter_pipe' [openwebrx-httpd] client cannot be closed: <class 'main.ClientNotFoundException'>
Hope this helps additional debugging. Are there any suggested Makefile settings to try for my CPU?
Best 73s de Fredrik SM2YHP
ha7ilm
commented
5 years ago ha7ilm
commented
5 years ago @cfenell I've proposed a fix for your problem. Try to update to the latest csdr, and please give feedback at simonyiszk/csdr#40 if it worked.
I'm using Ubuntu 14.04 LTS. For installation used this guide. When I open chrome, it states
WebSocket has closed unexpectedly. Please reload the page.. What I can to do? I have already attempted the nc -p\ncat edits to configure_webrx.py. Thanks.