# \[VOLK\] volk\_profile failures **URL:** \ **Category:** GNU Radio\ **Created:** [June 5, 2015, 11:44pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230 "2015-06-05T23:44:20Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![Detlef\_R](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/58f4c7/32.png) [@Detlef\_R](https://www.ruby-forum.com/u/Detlef_R)\ **Post date:** [June 5, 2015, 11:44pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/1 "2015-06-05T23:44:20Z")
I compiled GNU Radio 3.7.7.1 natively on an ODROID XU3-Lite with Ubuntu 14.04.2 LTS. I ran volk\_profile and got a lot of failures for various embedded architectures. Any ideas what could be happening? Could it be something with the compile flags I chose? volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_32fc\_s32f\_magnitude\_16i: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32f\_sqrt\_32f: fail on arch neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch a\_neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch neonvert volk\_32fc\_s32fc\_multiply\_32fc: fail on arch neon The full output of volk\_profile is here: [https://gist.github.com/nowls/37c78d89342053bae0ea](https://gist.github.com/nowls/37c78d89342053bae0ea) > Resulting volk\_config: > Some relevant configuration info taken from CMakeCache.txt (gcc/g++ version and C/CXX/ASM flags): CMAKE\_CXX\_COMPILER:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_CXX\_FLAGS\_DEBUG:STRING=-g CMAKE\_CXX\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_CXX\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_CXX\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_CXX\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_CXX\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_CXX\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_CXX\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_C\_COMPILER:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_C\_FLAGS\_DEBUG:STRING=-g CMAKE\_C\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_C\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_C\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_C\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_C\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_C\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_C\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_ASM\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_ASM\_FLAGS\_DEBUG:STRING=-g CMAKE\_ASM\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_ASM\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_ASM\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG Sean Nowlan ---
**Author:** ![NowlanS\_Sean](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/f05b48/32.png) [@NowlanS\_Sean](https://www.ruby-forum.com/u/NowlanS_Sean)\ **Post date:** [June 6, 2015, 12:49am UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/2 "2015-06-06T00:49:46Z")
Hi Sean, I didn’t read all of the debug info, but your symptoms are a result of ORC doing bad things. You can either uninstall ORC or upgrade to the latest release (0.4.23). The commit you really need is this one: [http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c](http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c) If you choose to upgrade ORC let me know how it goes. Cheers, Nathan ---
**Author:** ![NowlanS\_Sean](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/f05b48/32.png) [@NowlanS\_Sean](https://www.ruby-forum.com/u/NowlanS_Sean)\ **Post date:** [June 6, 2015, 6:48pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/3 "2015-06-06T18:48:08Z")
​Ok, I will install ORC 0.4.23 from source. Is ORC involved with all the other architectures that had failures? Sean * * * From: West, Nathan [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) Sent: Friday, June 5, 2015 6:48 PM To: Nowlan, Sean Cc: removed\_email\_address@domain.invalid Subject: Re: [Discuss-gnuradio] [VOLK] volk\_profile failures Hi Sean, I didn’t read all of the debug info, but your symptoms are a result of ORC doing bad things. You can either uninstall ORC or upgrade to the latest release (0.4.23). The commit you really need is this one: [http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c](http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c) If you choose to upgrade ORC let me know how it goes. Cheers, Nathan On Friday, June 5, 2015, Nowlan, Sean \ wrote: I compiled GNU Radio 3.7.7.1 natively on an ODROID XU3-Lite with Ubuntu 14.04.2 LTS. I ran volk\_profile and got a lot of failures for various embedded architectures. Any ideas what could be happening? Could it be something with the compile flags I chose? volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_32fc\_s32f\_magnitude\_16i: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32f\_sqrt\_32f: fail on arch neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch a\_neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch neonvert volk\_32fc\_s32fc\_multiply\_32fc: fail on arch neon The full output of volk\_profile is here: [https://gist.github.com/nowls/37c78d89342053bae0ea](https://gist.github.com/nowls/37c78d89342053bae0ea) > Resulting volk\_config: > Some relevant configuration info taken from CMakeCache.txt (gcc/g++ version and C/CXX/ASM flags): CMAKE\_CXX\_COMPILER:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_CXX\_FLAGS\_DEBUG:STRING=-g CMAKE\_CXX\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_CXX\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_CXX\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_CXX\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_CXX\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_CXX\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_CXX\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_C\_COMPILER:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_C\_FLAGS\_DEBUG:STRING=-g CMAKE\_C\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_C\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_C\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_C\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_C\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_C\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_C\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_ASM\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_ASM\_FLAGS\_DEBUG:STRING=-g CMAKE\_ASM\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_ASM\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_ASM\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG Sean Nowlan ---
**Author:** ![NowlanS\_Sean](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/f05b48/32.png) [@NowlanS\_Sean](https://www.ruby-forum.com/u/NowlanS_Sean)\ **Post date:** [June 6, 2015, 6:58pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/4 "2015-06-06T18:58:39Z")
Most likely. What we’ve seen before is that ORC didn’t preserve registers according to ARM specifications. This basically results in unpredictable data corruption. The commit I pointed to is the fix (vpush certain registers if used, then vpop after the routine). The visible corruption is seen by corrupting the test tolerance. If you look at ctest -V you should see that the "error"s are actually the same result, but since tolerance gets corrupted the test erroneously is false. The only QA failures should be on kernels that have an ORC protokernel, but this bug can affect non-ORC protokernels for that kernel. Nathan On Sat, Jun 6, 2015 at 12:47 PM, Nowlan, Sean [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) ---
**Author:** ![NowlanS\_Sean](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/f05b48/32.png) [@NowlanS\_Sean](https://www.ruby-forum.com/u/NowlanS_Sean)\ **Post date:** [June 7, 2015, 6:20pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/5 "2015-06-07T18:20:57Z")
On Sat, Jun 6, 2015 at 4:51 PM, Nowlan, Sean [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > CPPFLAGS=“-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 > -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations” > > UHD built fine w/ ORC 0.4.23. > > Sean Just disable ORC Altogether in VOLK. Most of the kernels have replacement NEON proto-kernels now, anyways. Pass -DENABLE\_ORC=False to cmake. Tom ---
**Author:** ![NowlanS\_Sean](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/f05b48/32.png) [@NowlanS\_Sean](https://www.ruby-forum.com/u/NowlanS_Sean)\ **Post date:** [June 6, 2015, 10:52pm UTC](https://www.ruby-forum.com/t/volk-volk-profile-failures/242230/6 "2015-06-06T22:52:54Z")
That makes sense. I installed ORC 0.4.23 and tried to recompile GR, but the build is failing on Volk. GCC segfaults a couple times before failing with an error on volk\_16ic\_magnitude\_16i\_a\_orc\_impl. Full output here: > ORC 0.4.23 was built with: CFLAGS=“-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations” CPPFLAGS=“-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations” UHD built fine w/ ORC 0.4.23. Sean * * * From: West, Nathan [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) Sent: Saturday, June 6, 2015 12:57 PM To: Nowlan, Sean Cc: removed\_email\_address@domain.invalid Subject: Re: [Discuss-gnuradio] [VOLK] volk\_profile failures Most likely. What we’ve seen before is that ORC didn’t preserve registers according to ARM specifications. This basically results in unpredictable data corruption. The commit I pointed to is the fix (vpush certain registers if used, then vpop after the routine). The visible corruption is seen by corrupting the test tolerance. If you look at ctest -V you should see that the "error"s are actually the same result, but since tolerance gets corrupted the test erroneously is false. The only QA failures should be on kernels that have an ORC protokernel, but this bug can affect non-ORC protokernels for that kernel. Nathan On Sat, Jun 6, 2015 at 12:47 PM, Nowlan, Sean \ wrote: ​Ok, I will install ORC 0.4.23 from source. Is ORC involved with all the other architectures that had failures? Sean * * * From: West, Nathan \ Sent: Friday, June 5, 2015 6:48 PM To: Nowlan, Sean Cc: removed\_email\_address@domain.invalid[mailto:removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) Subject: Re: [Discuss-gnuradio] [VOLK] volk\_profile failures Hi Sean, I didn’t read all of the debug info, but your symptoms are a result of ORC doing bad things. You can either uninstall ORC or upgrade to the latest release (0.4.23). The commit you really need is this one: [http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c](http://cgit.freedesktop.org/gstreamer/orc/commit/?id=acdaac31c648fd10f1bd0a49c4c33b483bb9c35c) If you choose to upgrade ORC let me know how it goes. Cheers, Nathan On Friday, June 5, 2015, Nowlan, Sean \ wrote: I compiled GNU Radio 3.7.7.1 natively on an ODROID XU3-Lite with Ubuntu 14.04.2 LTS. I ran volk\_profile and got a lot of failures for various embedded architectures. Any ideas what could be happening? Could it be something with the compile flags I chose? volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch neon volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_16ic\_s32f\_deinterleave\_32f\_x2: fail on arch u\_orc volk\_32fc\_s32f\_magnitude\_16i: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch a\_generic volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neon\_opttests volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch neonasm volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32fc\_x2\_multiply\_32fc: fail on arch u\_orc volk\_32f\_sqrt\_32f: fail on arch neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch a\_neon volk\_32f\_x3\_sum\_of\_poly\_32f: fail on arch neonvert volk\_32fc\_s32fc\_multiply\_32fc: fail on arch neon The full output of volk\_profile is here: [https://gist.github.com/nowls/37c78d89342053bae0ea](https://gist.github.com/nowls/37c78d89342053bae0ea) > Resulting volk\_config: > Some relevant configuration info taken from CMakeCache.txt (gcc/g++ version and C/CXX/ASM flags): CMAKE\_CXX\_COMPILER:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/g+±4.9 CMAKE\_CXX\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_CXX\_FLAGS\_DEBUG:STRING=-g CMAKE\_CXX\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_CXX\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_CXX\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_CXX\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_CXX\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_CXX\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_CXX\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_C\_COMPILER:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_COMPILER\_WITH\_PATH:FILEPATH=/usr/bin/gcc-4.9 CMAKE\_C\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_C\_FLAGS\_DEBUG:STRING=-g CMAKE\_C\_FLAGS\_DEBUGPARANOID:STRING=-Wall -Wextra -g -O0 CMAKE\_C\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_C\_FLAGS\_NOOPTWITHASM:STRING=-Wall -save-temps -g -O0 CMAKE\_C\_FLAGS\_O2WITHASM:STRING=-Wall -save-temps -g -O2 CMAKE\_C\_FLAGS\_O3WITHASM:STRING=-Wall -save-temps -g -O3 CMAKE\_C\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_C\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG CMAKE\_ASM\_FLAGS:STRING=-mcpu=cortex-a15.cortex-a7 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard -funsafe-math-optimizations CMAKE\_ASM\_FLAGS\_DEBUG:STRING=-g CMAKE\_ASM\_FLAGS\_MINSIZEREL:STRING=-Os -DNDEBUG CMAKE\_ASM\_FLAGS\_RELEASE:STRING=-O3 -DNDEBUG CMAKE\_ASM\_FLAGS\_RELWITHDEBINFO:STRING=-O2 -g -DNDEBUG Sean Nowlan