| US 7,500,244 B2 | ||
| Adaptive algorithm for selecting a virtualization algorithm in virtual machine environments | ||
| Gehad Galal, Rancho Cordova, Calif. (US); and Randolph Campbell, Folsom, Calif. (US) | ||
| Assigned to Intel Corporation, Santa Clara, Calif. (US) | ||
| Filed on Jun. 30, 2004, as Appl. No. 10/881,784. | ||
| Prior Publication US 2006/0005199 A1, Jan. 05, 2006 | ||
| Int. Cl. G06F 9/455 (2006.01); G06F 9/46 (2006.01) | ||
| U.S. Cl. 718—108 [718/1] | 63 Claims |
| 43. A computer-implemented method comprising:
calculating an exit-enter time (EET) to exit and enter a context and a save-restore time (SRT) to save and restore a machine
state;
executing a selective algorithm that saves and restores the machine state on selected context changes and accumulating statistics
for an expected value for EET overhead and an expected value for SRT overhead while executing the selective algorithm;
computing a first cost of the selective algorithm using the expected value for EET overhead and the expected value for SRT
overhead;
executing a second selective algorithm that selectively saves and restores the machine state when there is a change of context;
accumulating statistics on an expected value for EET per context change plus an expected value for SRT per context change
for the second selective algorithm;
computing a second cost of the second selective algorithm as the expected value for EET per context change plus the expected
value for SRT per context change;
computing a third cost of an unconditional algorithm that always saves and restores the machine state on context changes as
two times SRT; and
selecting one of the selective algorithm, the second selective algorithm, or the unconditional algorithm having the least
cost as a virtualization algorithm to virtualize context changes.
|