https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/New-Squaring-Method-Comparison-with-Karatsuba-and-GMP-Enhanced/m-p/1684027#M37080 <P>Hello Intel Developer Community,<BR /><BR />I’m excited to share my new squaring algorithm, which I’ve been developing for the last 4 years. It provides significant improvements in terms of performance and efficiency when compared to the widely used Karatsuba method.<BR /><BR />In the first phase, I’ve compared my raw method (RPF) against Karatsuba’s algorithm for squaring numbers. The results show how my method outperforms Karatsuba in terms of execution time and scaling, particularly as the input size increases.<BR /><BR />In the second phase, I’ve also enhanced both RPF and Karatsuba with GMP (GNU Multiple Precision Arithmetic Library) optimizations. When I apply GMP functions, the performance of both methods significantly improves, but my method still maintains a superior edge.<BR /><BR />I’ve attached some charts comparing the execution times and performance metrics for both raw and GMP-enhanced versions of the algorithms. I’d love to get some feedback or thoughts from the community, especially from those involved in high-performance computation or numerical algorithms.<BR /><BR />Looking forward to your thoughts and discussions.<BR /><BR />Best Regards,<BR />Krishil Rohit Sheth</P><P>&lt;personal information removed&gt;</P> Sat, 19 Apr 2025 12:56:01 GMT Krishil 2025-04-19T12:56:01Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/New-Squaring-Method-Comparison-with-Karatsuba-and-GMP-Enhanced/m-p/1684027#M37080 <P>Hello Intel Developer Community,<BR /><BR />I’m excited to share my new squaring algorithm, which I’ve been developing for the last 4 years. It provides significant improvements in terms of performance and efficiency when compared to the widely used Karatsuba method.<BR /><BR />In the first phase, I’ve compared my raw method (RPF) against Karatsuba’s algorithm for squaring numbers. The results show how my method outperforms Karatsuba in terms of execution time and scaling, particularly as the input size increases.<BR /><BR />In the second phase, I’ve also enhanced both RPF and Karatsuba with GMP (GNU Multiple Precision Arithmetic Library) optimizations. When I apply GMP functions, the performance of both methods significantly improves, but my method still maintains a superior edge.<BR /><BR />I’ve attached some charts comparing the execution times and performance metrics for both raw and GMP-enhanced versions of the algorithms. I’d love to get some feedback or thoughts from the community, especially from those involved in high-performance computation or numerical algorithms.<BR /><BR />Looking forward to your thoughts and discussions.<BR /><BR />Best Regards,<BR />Krishil Rohit Sheth</P><P>&lt;personal information removed&gt;</P> Sat, 19 Apr 2025 12:56:01 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/New-Squaring-Method-Comparison-with-Karatsuba-and-GMP-Enhanced/m-p/1684027#M37080 Krishil 2025-04-19T12:56:01Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/New-Squaring-Method-Comparison-with-Karatsuba-and-GMP-Enhanced/m-p/1686658#M37103 <P>Thank you for sharing this information. Please be aware that GMP is no longer part of the oneMKL domain. You might find more relevant feedback from the GMP community.</P><P><BR /></P><P>Regards, </P><P>Alex</P><BR /> Wed, 30 Apr 2025 10:41:11 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/New-Squaring-Method-Comparison-with-Karatsuba-and-GMP-Enhanced/m-p/1686658#M37103 Aleksandra_K 2025-04-30T10:41:11Z