Process for providing submodel performance in a computer processing unit
Abstract
A simple and accurate processor derating method includes: sampling a real-time counter/clock too obtain an initial time value T 1 ; resetting an Icnt Counter; incrementing the Icnt Counter to reflect the processing of each instruction; comparing the count in the Icnt Counter to a predetermined count IcntMax and if the count in the Icnt Counter is at least IcntMax, then sampling the RTC to obtain a second time T 2 . T 1 is then subtracted from T 2 to obtain a time difference DT which is multiplied by ((1−1/DF)−1) to obtain a Degradation Delay DD period, DF being a constant having a value which is the desired submodel performance with respect to full performance. The Degradation Delay is instituted, the RTC is sampled from time to time to obtain a test third time T 3 . When a test T 3 minus T 2 is not less than DD, then T 1 is set to T 3 . Then, the procedure is repeated for a next group of instructions. Optionally, further accuracy can be achieved by treating “wait-type” and/or “RTC-access-type” instructions specially and also by calculating a DDExtra period value which is used to adjust the next DD.
Claims
exact text as granted — not AI-modified1 . A process for limiting performance of a processor to a selected submodel capability comprising the steps of:
a. Sample a Real Time Clock (RTC) to obtain a first time Value T 1 ; b. Reset an institution count (Icnt) Counter; c. Process instructions while incrementing the lent counter to reflect the processing of each instruction; d. Compare the count in the Icnt Counter to a predetermined Counter Instruction Count Maximum (IcntMax); and
i. If the count in the lcnt Counter is less than IcntMax, then return to step c; and
ii. If the counter in said lcnt Counter is at least IcntMax, then proceed to step e;
e. Sample the RTC to obtain a second time T 2 ; f. Subtract T 1 from T 2 to obtain a time difference DT; g. Multiply DT by ((1−1/DF)−1) to obtain a Degradation Delay DD period. Degradation Factor DF being a constant having a value which is the desired submodel performance with respect to full performance; h. Delay; During step h, sample the RTC to obtain it test third time T 3 ;
i. If test third time T 3 -minus T 2 is less than DD, then continue step h; and
ii. If test third time T 3 minus T 2 is not less than DD, then proceed to step j;
j. Set T 1 equal to T 3 as used in step i.ii; and k. Go to step b.
2 . A process for limiting performance of a processor to a selected submodel capability comprising the steps of:
a. Sample a Real Time Clock(RTC) to obtain a first time value T 1 ; b. Reset an Instruction Count(Icnt) Counter; c. Process instructions while incrementing the Icnt to reflect the processing of each instruction; d. Compare the Count in the Icnt Counter to a predetermined count Instruction Count Maximum(IcntMax); and p 2 i. If the count in the Icnt Counter is less than IcntMax, then return to step c; and p 2 ii. If the count in said Icnt Counter is at least IcntMax, then proceed to step e; e. Sample the RTC to obtain a second time T 2 ; f. Subtract T 1 from T 2 to obtain a time difference DT; g. Multiply DT by ((1−1/DF)−1) to obtain a Degradation Delay DD period. Degradation Factor DF being a constant having a value which is the desired submodel performance with respect to full performance; h. Delay; i. During step h, sample the RTC to obtain a test third time T 3 ;
i. If test third time T 3 minus T 2 is less than DD, then continue step h: and
ii. If test third time T 3 minus T 2 is not less than DD, then proceed to step j;
j. Obtain a period value, DDExtra, by subtracting T 2 and DD from T 3 as used in step i.ii; k. Set T 1 equal to T 3 as used in step i.ii; and l. Go to step b.
3 . A process for limiting performance of a processor to a selected submodel capability comprising the steps of:
a. Sample a Real Time Clock(RTC) to obtain a first time value T 1 ; b. Reset an Instruction Count(Icnt) counter; c. Determine if a current instruction to be process is a “wait-type ” having an externally-terminable internal wait;
i. If the current instruction is not a “wait-type”, proceed to step d; and
ii. If the current instruction is a “wait-type”, go to step f.
d. Process instruction while incrementing the Icnt Counter to reflect the processing of each instruction; e. Compare the counter in the Icnt Counter to a predetermined count Instruction Count Maximum(IcntMax); and
i. If the count in the Icnt Counter is less than IcntMax, then return to step d; and
ii. If the count in said Icnt Counter is at least IcntMax, then proceed to step f;
f. Sample the RTC to obtain a second time T 2 ; g. Subtract T 1 from T 2 to obtain a time difference DT; h. Multiply DT by ((1−1/DF)−1) to obtain a Degradation Delay DD period. Degradation Factor DF being a constant having a value which is the desired Submodel performance with respect to full performance; i. Delay; j. During step i,
i. Determine if the current instuction is a “wait-type”;
ii. If the Current instruction is not a wait “type”, go to step j.iv;
iii. If the current instruction is a i “wait-type”, determine if the current instruction internal wait is done
1 . if the current instruction internal wait is not done, go to step jv.iv; and
2 . If the current internal wait is done, go to step k.
iv. sample the RTC to obtain a test third time T 3 ;
v. If test third T 3 minus T 2 is less than DD, then continue step i; and
vi. If test third time T 3 minus T 2 is not less than DD, then proceed to step k;
k. Set T 1 equal to T 3 as used in step i.ii; and Go to step b.
4 . (canceled)
5 . The process of claim 1 in which the value of IcntMax is at least 100.
6 . The process of claim 2 in which the value of IcntMax is at least 100.
7 . The process of claim 3 in which the value of IcntMax is at least 100.
8 . (canceled)Join the waitlist — get patent alerts
Track US2007157050A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.