Community
cancel
Showing results for 
Search instead for 
Did you mean: 
idata
Community Manager
453 Views

NCSDK2 without AVX

I've just downloaded the new NCSDK 2 with the instructions from https://software.intel.com/en-us/neural-compute-stick/get-started and the install failed because the CPU on which I need to run it doesn't have AVX support (Intel Atom). That being said, are there ways to still use the sdk without that flag and if so, how should I proceed ? Is it possible to recompile some stuff for my specific machine ?

Tags (1)
0 Kudos
10 Replies
idata
Community Manager
97 Views

Hi @Oppiie

 

What are the specs of the system you're trying to install on?

 

Best Regards,

 

Sahira
idata
Community Manager
97 Views

Hi @Sahira_at_Intel,

 

cat cpuinfo

 

 

processor : 0

 

vendor_id : GenuineIntel

 

cpu family : 6

 

model : 76

 

model name : Intel(R) Atom(TM) x5-Z8350 CPU @ 1.44GHz

 

stepping : 4

 

microcode : 0x40c

 

cpu MHz : 480.000

 

cache size : 1024 KB

 

physical id : 0

 

siblings : 4

 

core id : 0

 

cpu cores : 4

 

apicid : 0

 

initial apicid : 0

 

fpu : yes

 

fpu_exception : yes

 

cpuid level : 11

 

wp : yes

 

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 rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology tsc_reliable nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm sse4_1 sse4_2 movbe popcnt tsc_deadline_timer aes rdrand lahf_lm 3dnowprefetch epb pti tpr_shadow vnmi flexpriority ept vpid tsc_adjust smep erms dtherm ida arat

 

bugs : cpu_meltdown spectre_v1 spectre_v2

 

bogomips : 2880.00

 

clflush size : 64

 

cache_alignment : 64

 

address sizes : 36 bits physical, 48 bits virtual

 

power management:

 

 

If you need other specific info, please let me know which info and I'll provide it. Thanks

idata
Community Manager
97 Views

Hi @Oppiie

 

Thanks for providing that info. What exactly is the error message you were getting?

 

Best Regards,

 

Sahira
idata
Community Manager
97 Views

Hi,

 

The source is at https://github.com/movidius/ncsdk/blob/ncsdk2/install.sh

 

Lines 39 to 45.

 

Thanks

idata
Community Manager
97 Views

Hi @Oppiie

 

Tensorflow 1.6 and later started using AVX instructions in their prebuild binaries. So CPUs with no AVX instructions like the Atom processor would have to use Tensorflow 1.5 or below or build from source. I hope this was helpful!

 

Best Regards,

 

Sahira
idata
Community Manager
97 Views

Hi @Sahira_at_Intel

 

Ok so if I build TF 1.12 for my machine (without AVX flag), I can simply remove that check from Install.sh and it should all work as intended ?

 

This is a lot of work since I have to cross-compile TF from Windows for Ubuntu 16.04, so I'd like to have some sort of confirmation that it should work past that. ^^'

 

Thank you,

 

Oppiie
idata
Community Manager
97 Views

Hi @Oppiie

 

I would recommend using the latest Tensorflow version that doesn't use AVX instructions. TF 1.12 is a little old and there have been some errors with running versions that old. Otherwise I think it should work.

 

Sincerely,

 

Sahira
idata
Community Manager
97 Views

Hi @Sahira_at_Intel,

 

I meant the r1.12 branch of Tensorflow, which is much more recent than r1.5 (the latest being r1.13). Is that ok ?

 

Thanks

idata
Community Manager
97 Views

Hello @Sahira_at_Intel ,

 

I managed to compile Tensorflow r1.12 from source and removed the AVX check from the Install.sh for the Movidius stick. It seems to be working fine and I ran inference using the NCS.

 

Thanks,

 

Mat
idata
Community Manager
97 Views

Hi @Oppiie

 

Glad to see it's working!

 

Best Regards,

 

Sahira
Reply