Developer Reference
Intel® oneAPI Math Kernel Library Vector Mathematics Performance and Accuracy Data
                    
                        ID
                        772989
                    
                
                
                    Date
                    12/04/2020
                
                
                    Public
                
            LGamma
Computes the natural logarithm of the absolute value of gamma function for vector elements.
Performance:
| Intel® Xeon® Processor E5-2680 | Intel® Xeon® Processor E5-2699 v3 | Intel® Xeon® Processor E5-2699 v4 | Intel® Xeon® Gold 6148 Processor | ||
|---|---|---|---|---|---|
| single | HA | 40.40 | 44.70 | 35.15 | 41.79 | 
| LA | 40.39 | 44.69 | 35.15 | 41.78 | |
| EP | 40.39 | 44.69 | 35.14 | 41.72 | |
| double | HA | 86.17 | 103.57 | 81.53 | 76.29 | 
| LA | 86.17 | 103.48 | 81.53 | 76.29 | |
| EP | 86.15 | 103.48 | 81.53 | 76.29 | |
      Notes: 
    
 
    - Units - CPE (Clocks per Element)
- Performance was measured on "working" interval (double precision: [ 0.00001, 1000.0 ]; single precision: [ 0.00001, 1000.0 ]) 
- Data - vectors of 1000 elements with random generated numbers. 
- All data was collected with 64-bit libraries. 
Accuracy:
| Intel® Xeon® Processor E5-2680 | Intel® Xeon® Processor E5-2699 v3 | Intel® Xeon® Processor E5-2699 v4 | Intel® Xeon® Gold 6148 Processor | ||
|---|---|---|---|---|---|
| single | HA | 0.50 | 0.50 | 0.50 | 0.50 | 
| LA | 0.50 | 0.50 | 0.50 | 0.50 | |
| EP | 0.50 | 0.50 | 0.50 | 0.50 | |
| double | HA | 0.50 | 0.50 | 0.50 | 0.50 | 
| LA | 0.50 | 0.50 | 0.50 | 0.50 | |
| EP | 0.50 | 0.50 | 0.50 | 0.50 | |
      NOTE: 
    
 Units - ULP (Units in the Last Place). 
   Performance Graphs:




 Parent topic: Special