IP Library Granted Patent US 12,416,965
Granted Patent B2
US 12,416,965 · App. 17/644,677 · Granted Sep 16, 2025

Distribution of frequency budget in a processor

Inventors: Madhusudan Chidambaram (Bangalore, IN); Vishwesh Rudramuni (Karnataka, IN); Shmuel Zobel (Hinanit, IL); Alexander Gendler (Kiriat Motzkin, IL)
Assignee: Intel Corporation
G06F1/324
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,416,965
App. No.
17/644,677
Granted
Sep 16, 2025
Kind
B2
Abstract

In an embodiment, a processor may include a core domain comprising a plurality of processing cores, an uncore domain comprising an internal network, and a processor power management circuit. The processor power management circuit may be to: receive scalability hint values from the processing cores; determine a total core gain and a total uncore gain based at least in part on the scalability hint values; and distribute a frequency budget between the core domain and the uncore domain based at least in part on the total core gain and the total uncore gain. Other embodiments are described and claimed.

Claims (76)

1. A processor comprising:

a core domain comprising a plurality of processing cores;

an uncore domain comprising an internal network; and

a processor power management circuit to:

receive a plurality of scalability hint values from the plurality of processing cores;

determine a total core gain and a total uncore gain based at least in part on the plurality of scalability hint values; and

distribute a frequency budget between the core domain and the uncore domain based at least in part on the total core gain and the total uncore gain.

2. The processor of claim 1 , wherein each processing core comprises:

a core power management circuit to:

determine a plurality of metrics that indicate a current workload of the processing core; and

execute, using the plurality of metrics as inputs, a prediction algorithm to generate the plurality of scalability hint values.

3. The processor of claim 2 , wherein the plurality of metrics comprises:

instructions per cycle,

count of inactive cycles,

counts of queue occupancy for multiple thresholds,

count of prefetches in a cycle,

count of cache hits, and

count of cache misses.

4. The processor of claim 1 , wherein the internal network comprises a mesh network including a plurality of mesh stops.

5. The processor of claim 1 , wherein the plurality of scalability hint values comprises:

an instructions per second (IPS) value,

a core scalability (CS) value,

an uncore scalability (US) value, and

a quiesced core (QC) indication.

6. The processor of claim 5 , wherein the processor power management circuit is to:

for each of the plurality of processing cores, calculate a core gain value by multiplying the core scalability value, the IPS value, and a bin ratio; and

for each of the plurality of processing cores, calculate an uncore gain value by multiplying the uncore scalability value, the IPS value, and the bin ratio.

7. The processor of claim 6 , wherein the processor power management circuit is to:

calculate the total core gain as a sum of the core gain values for the plurality of processing cores; and

calculate the total uncore gain as a sum of the uncore gain values for the plurality of processing cores.

8. The processor of claim 6 , wherein the processor power management circuit is to:

generate, based at least on the total core gain and the total uncore gain, a workload hint to indicate a workload profile of the processing core; and

determine an uncore frequency based at least on the generated workload hint.

9. The processor of claim 5 , wherein the processor power management circuit is to:

determine a per core frequency budget based at least in part on the core scalability values for each of the plurality of processing cores.

10. A method comprising:

determining, for each processing core of a processor, a plurality of metrics indicating a current workload of the processing core, wherein the processor comprises a core domain and an uncore domain, and wherein the core domain comprises a plurality of processing cores;

executing by a power management circuit of the processor, using the plurality of metrics as inputs, a prediction algorithm to generate a plurality of scalability hint values;

determining, by the power management circuit, a total core gain and a total uncore gain based at least in part on the plurality of scalability hint values; and

distributing, by the power management circuit, a frequency budget between the core domain and the uncore domain based at least in part on the total core gain and the total uncore gain.

11. The method of claim 10 , wherein the plurality of scalability hint values comprises:

an instructions per second (IPS) value,

a core scalability (CS) value,

an uncore scalability (US) value, and

a quiesced core (QC) indication.

12. The method of claim 11 , further comprising:

for each of the plurality of processing cores, calculating a core gain value by multiplying the core scalability value, the IPS value, and a bin ratio; and

for each of the plurality of processing cores, calculating an uncore gain value by multiplying the uncore scalability value, the IPS value, and the bin ratio.

13. The method of claim 12 , further comprising:

calculating the total core gain as a sum of the core gain values for the plurality of processing cores; and

calculating the total uncore gain as a sum of the uncore gain values for the plurality of processing cores.

14. The method of claim 11 , further comprising:

determining a per core frequency budget based at least in part on the core scalability values for each of the plurality of processing cores.

15. The method of claim 10 , further comprising:

generating, based at least on the total core gain and the total uncore gain, a workload hint to indicate a workload profile of the processing core; and

determining an uncore frequency based at least on the generated workload hint.

16. The method of claim 10 , further comprising:

computing a ratio of the total core gain to the total uncore gain;

determining a budget distribution based on the ratio of the total core gain to the total uncore gain; and

distributing a frequency budget based on the determined budget distribution.

17. A system comprising:

a processor comprising:

a core domain comprising a plurality of processing cores;

an uncore domain comprising an internal network; and

a processor power management circuit to:

receive a plurality of scalability hint values from the plurality of processing cores;

determine a total core gain and a total uncore gain based at least in part on the plurality of scalability hint values; and

distribute a frequency budget between the core domain and the uncore domain based at least in part on the total core gain and the total uncore gain; and

a memory coupled to the processor.

18. The system of claim 17 , wherein the plurality of scalability hint values comprises an instructions per second (IPS) value, a core scalability (CS) value, an uncore scalability (US) value, and a quiesced core (QC) indication.

19. The system of claim 18 , wherein the processor power management circuit is to:

for each of the plurality of processing cores, calculate a core gain value by multiplying the core scalability value, the IPS value, and a bin ratio; and

for each of the plurality of processing cores, calculate an uncore gain value by multiplying the uncore scalability value, the IPS value, and the bin ratio.

20. The system of claim 19 , wherein the processor power management circuit is to:

calculate the total core gain as a sum of the core gain values for the plurality of processing cores; and

calculate the total uncore gain as a sum of the uncore gain values for the plurality of processing cores.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 16, 2021
From: CHIDAMBARAM, MADHUSUDAN; RUDRAMUNI, VISHWESH M.; ZOBEL, SHMUEL; GENDLER, ALEXANDER
To: INTEL CORPORATION
Reel/Frame 058409/0389 →
Continuity (1)
Related Publication 20230195199A1 · Jun 22, 2023
References Cited (17)
US 6205533B1 · Margolus · 2001 [cited by applicant]
US 10678724B1 · ChoFleming et al. · 2020 [cited by applicant]
US 10795853B2 · Nassif et al. · 2020 [cited by applicant]
US 10891240B2 · Mathew et al. · 2021 [cited by applicant]
US 11086816B2 · Fleming et al. · 2021 [cited by applicant]
US 11200186B2 · Fleming et al. · 2021 [cited by applicant]
US 20130111236A1 · Ananthakrishnan · 2013 [cited by examiner]
US 20140089688A1 · Man · 2014 [cited by examiner]
US 20140181538A1 · Shrall · 2014 [cited by examiner]
US 20150277538A1 · Yasin et al. · 2015 [cited by applicant]
US 20170031412A1 · Gendler et al. · 2017 [cited by applicant]
US 20220026972A1 · Shah · 2022 [cited by examiner]
US 20220100247A1 · Garg · 2022 [cited by examiner]
EP 2796961A2 · 2014 [cited by applicant]
EP 3901767A1 · 2021 [cited by applicant]
European Patent Office, Extended European Search Report for Appl. No. 22203049.6 dated May 4, 2023 (10 pages). [cited by applicant]
Gupta et al., U.S. Appl. No. 17/448,600 entitled Gating of a Mesh Clock Signal in a Processor filed Sep. 23, 2021 (145 pages). [cited by applicant]