Vectorization analysis not supported for Core i7 -4770 ?

[johnyb_]

Hi,

I am running Intel® VTune™ Amplifier XE 2015 Update 4, (build 410668) on Windows 8.1 with a Core i7-4770 processor

When I check "Analyze vectorization" in a custom analysis I get the message "Vectorization analysis is not supported for this microprocessor. See release note for more détails".

I want to run this on a program compiled in VS2012 with Intel Fortran Version 15.0.0122.11. My user account on this PC has administrator rights.

I can't find any explanation for this in the VTune Help.

Johny

P.S. This is the output of the Intel CPUID utility:

Intel(R) Processor Identification Utility
Version: 5.30.20150729
Time Stamp: 2015/08/14 09:43:09
Operating System: 6.2-9200-
Number of processors in system: 1
Current processor: #1
Active cores per processor: 4
Disabled cores per processor: 0
Processor Name: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz
Type: 0
Family: 6
Model: 3C
Stepping: 3
Revision: 19
Maximum CPUID Level: D
L1 Instruction Cache: 4 x 32 KB
L1 Data Cache: 4 x 32 KB
L2 Cache: 4 x 256 KB
L3 Cache: 8 MB
Packaging: LGA1150
Enhanced Intel SpeedStep(R) Technology: Yes
MMX(TM): Yes
Intel(R) SSE: Yes
Intel(R) SSE2: Yes
Intel(R) SSE3: Yes
Intel(R) SSE4: Yes
Intel(R) AES-NI: Yes
Intel(R) AVX: Yes
Enhanced Halt State: Yes
Execute Disable Bit: Yes
Intel(R) Hyper-Threading Technology: Yes
Intel(R) 64 Architecture: Yes
Intel(R) Virtualization Technology: Yes
Intel(R) VT-x with Extended Page Tables: Yes
System Graphics: Intel(R) HD Graphics 4600
Base TDP: 84
Expected Processor Frequency: 3.40 GHz
Reported Processor Frequency: 3.69 GHz
Expected System Bus Frequency: 100 MHz
Reported System Bus Frequency: 100 MHz
*************************************************************

[Dmitry_P_Intel1 · Accepted solution]

Hello Johny,

The knob was enabled for Ivy Bridge and Broadwell based processors that have FLOP performance counters. Your processor is Haswell based and since there are no proper counters there - it is not available.

Thanks & Regards, Dmitry

[johnyb_]

Thanks Dmitry,

is there any other way to estimate the floating point load ?

Johny

[Dmitry_P_Intel1]

I would recommend to use Intel Vector Advisor to look at vectorization efficiency. If you have subscribed to Intel Parallel Studio 2016 Beta then you have it. If no - very soon there will be 2016 Gold release and you will be able to do upgrade having an actionable license.

Thank you, Regards, Dmitry

[johnyb_]

Thanks again,

I'll wait then for the release of Parallel Studio 2016

Johny