[BACK]Return to 53.html CVS log [TXT][DIR] Up to [local] / www

Diff for /www/53.html between version 1.91 and 1.92

version 1.91, 2019/05/27 22:55:18 version 1.92, 2019/06/01 23:12:46
Line 225 
Line 225 
     <ul>      <ul>
     <li>Switch default disk I/O sort algorithm to NSCAN for improved fairness and response times.      <li>Switch default disk I/O sort algorithm to NSCAN for improved fairness and response times.
     <li>Softdep speedup improvements by the revert of a previously necessary workaround to prevent kva starvation.      <li>Softdep speedup improvements by the revert of a previously necessary workaround to prevent kva starvation.
     <li>Large performance and reliability improvements in      <li>Large performance and reliability improvements in
 <a href="https://man.openbsd.org/make.1">make(1)</a>,  <a href="https://man.openbsd.org/make.1">make(1)</a>,
 especially in parallel mode. make no longer does any busy waiting, it handles  especially in parallel mode. make no longer does any busy waiting, it handles
 concurrent targets correctly, and displays more accurate error messages.  concurrent targets correctly, and displays more accurate error messages.
Line 393 
Line 393 
     </ul>      </ul>
 <p>  <p>
   
 <li>Over 7,800 ports, major performance and stability improvements in  <li>Over 7,800 ports, major performance and stability improvements in
 the package build process  the package build process
     <ul>      <ul>
     <li>The parallel ports builder is more efficient. The main improvement is      <li>The parallel ports builder is more efficient. The main improvement is
 that dpb consumes much less cpu on busy boxes, but there are lots of small  that dpb consumes much less cpu on busy boxes, but there are lots of small
 optimizations that amount to a large performance increase:  optimizations that amount to a large performance increase:
 dpb can now build selected large ports using parallel make, and it  dpb can now build selected large ports using parallel make, and it
 has a notion of affinity, so that ports failing on a cluster will be  has a notion of affinity, so that ports failing on a cluster will be
 preferentially restarted on the same machine.  preferentially restarted on the same machine.
     </ul>      </ul>
 <p>  <p>
Legend:
Removed from v.1.91  
changed lines
  Added in v.1.92