Another Sizable Performance Optimization To Benefit Network Code With Linux 5.17
Last week I wrote about a big TCP performance optimization having been queued up into net-next for Linux 5.17. That optimization can yield significant TCP throughput improvements especially with today's high-end 100Gb+ network hardware. There is now another separate juicy optimization to benefit the Linux network performance in the next kernel cycle.
This completely separate optimization but that will also benefit the Linux 5.17 network code path is optimizing the x86_64 csum_partial() function. This latest optimization also comes from Google's Eric Dumazet.
This new optimization is working its way to the Linux kernel by way of x86/core changes being queued there. It's a rewritten/optimized csum_partial() function for x86/x86_64. That function is widely used by the kernel's network code for computing TCP checksums.
This new optimized version of the "heavily used" csum_partial() function is showing a "big reduction" in the cost by 50~80% across a range of processors.
See this patch for more details. Look for this optimized code in Linux 5.17 early next year.
This completely separate optimization but that will also benefit the Linux 5.17 network code path is optimizing the x86_64 csum_partial() function. This latest optimization also comes from Google's Eric Dumazet.
This new optimization is working its way to the Linux kernel by way of x86/core changes being queued there. It's a rewritten/optimized csum_partial() function for x86/x86_64. That function is widely used by the kernel's network code for computing TCP checksums.
This new optimized version of the "heavily used" csum_partial() function is showing a "big reduction" in the cost by 50~80% across a range of processors.
See this patch for more details. Look for this optimized code in Linux 5.17 early next year.