Differentiating cache reliability to reduce minimum on-die voltage
Systems and methods may provide for determining whether a memory access request is error-tolerant, and routing the memory access request to a reliable memory region if the memory access request is error-tolerant. Moreover, the memory access request may be routed to an unreliable memory region if the memory access request is error-tolerant. In one example, use of the unreliable memory region enables a reduction in the minimum operating voltage level for a die containing the reliable and unreliable memory regions.
1. A system comprising:
at least one processor core;
a local on-die memory including a reliable region and an unreliable region, wherein the reliable region has a failure probability below a given threshold for a set operational voltage and the unreliable region has a failure probability above the given threshold for the set operational voltage;
logic to,
determine whether a memory access request is error-tolerant,
route the memory access request to the reliable region if the memory access request is not error-tolerant, wherein the logic is to identify an instruction annotation in the memory access request to determine whether the memory access request is error-tolerant, or wherein the logic is to identify a physical memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant or wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, and
route the memory access request to the unreliable region if the memory access request is error-tolerant; and a power supply coupled to the at least one processor core and the local on-die memory.
2. The system of claim 1 , wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant and wherein the logic is to access a reliability bit in a page table.
3. The system of claim 1 , wherein the logic is to determine that the memory access request corresponds to a level one (L1) cache miss.
4. The system of claim 1 , further including a voltage regulator having an input coupled to the power supply and an output coupled to the at least one processor core and the local on-die memory.
5. The system of claim 4 , further including a die coupled to the at least one processor core and the local on-die memory, wherein the output of the voltage regulator establishes a common operational voltage for the die.
6. The system of claim 1 , wherein the unreliable region has more storage capacity than the reliable region.
7. The system of claim 1 , wherein the reliable region is to include instructions and data, and the unreliable region is to be dedicated to data.
8. An apparatus comprising:
logic, implemented in one or more of configurable logic or fixed-functionality logic hardware, to,
determine whether a memory access request is error-tolerant wherein the logic is to identify a physical memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the logic is to identify an instruction annotation in the memory access request to determine whether the memory access request is error-tolerant,
route the memory access request to a reliable memory region if the memory access request is not error-tolerant, and
route the memory access request to an unreliable memory region if the memory access request is error-tolerant, wherein the reliable memory region has a failure probability below a given threshold for a set operational voltage and the unreliable memory region has a failure probability above the given threshold for the set operational voltage.
9. The apparatus of claim 8 , wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant and wherein the logic is to access a reliability bit in a page table.
10. The apparatus of claim 8 , wherein the logic is to determine that the memory access request corresponds to a level one (L1) cache miss.
11. A non-transitory computer readable storage medium comprising a set of instructions which, if executed by a processor, cause a computer to:
determine whether a memory access request is error-tolerant wherein the instructions, if executed, cause a computer to identify a physical memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the instructions, if executed, cause a computer to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the instructions, if executed, cause a computer to identify an instruction annotation in the memory access request to determine whether the memory access request is error-tolerant;
route the memory access request to a reliable memory region if the memory access request is not error-tolerant; and
route the memory access request to an unreliable memory region if the memory access request is error-tolerant, wherein the reliable memory region has a failure probability below a given threshold for a set operational voltage and the unreliable memory region has a failure probability above the given threshold for the set operational voltage.
12. The medium of claim 11 , wherein the instructions, if executed, cause a computer to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant and wherein the instructions, if executed, cause a computer to access a reliability bit in a page table.
13. The medium of claim 11 , wherein the instructions, if executed, cause a computer to determine that the memory access request corresponds to a level one (L1) cache miss.
14. A computer implemented method comprising:
determining whether a memory access request is error-tolerant, wherein determining whether the memory access request is error-tolerant includes identifying a physical memory allocation for a variable associated with the memory access request, or wherein determining whether the memory access request is error-tolerant includes identifying a virtual memory allocation for a variable associated with the memory access request, or wherein determining whether the memory access request is error-tolerant includes identifying an instruction annotation in the memory access request;
routing the memory access request to a reliable memory region if the memory access request is not error-tolerant; and
routing the memory access request to an unreliable memory region if the memory access request is error-tolerant, wherein the reliable memory region has a failure probability below a given threshold for a set operational voltage and the unreliable memory region has a failure probability above the given threshold for the set operational voltage.
15. The method of claim 14 , wherein determining whether the memory access request is error-tolerant includes identifying a virtual memory allocation for a variable associated with the memory access request and wherein identifying the virtual memory allocation includes accessing a reliability bit in a page table.
16. The method of claim 14 , further including determining that the memory access request corresponds to a level one (L1) cache miss.
17. A system comprising:
at least one processor core;
a local on-die memory including a reliable region and an unreliable region, wherein the reliable region has a failure probability below a given threshold for a set operational voltage and the unreliable region has a failure probability above the given threshold for the set operational voltage; and
logic to,
determine whether a memory access request is error-tolerant wherein the logic is to identify a physical memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant, or wherein the logic is to identify an instruction annotation in the memory access request to determine whether the memory access request is error-tolerant,
route the memory access request to the reliable region if the memory access request is not error-tolerant, and
route the memory access request to the unreliable region if the memory access request is error-tolerant.
18. The system of claim 17 , wherein the logic is to identify a virtual memory allocation for a variable associated with the memory access request to determine whether the memory access request is error-tolerant and wherein the logic is to access a reliability bit in a page table.
19. The system of claim 17 , wherein the logic is to determine that the memory access request corresponds to a level one (L1) cache miss.
20. The system of claim 17 , further including a voltage regulator having an input coupled to the power supply and an output coupled to the first processor core, the second processor core and the local on-die memory.
21. The system of claim 20 , further including a die coupled to the first processor core, the second processor core and the local on-die memory, wherein the output of the voltage regulator establishes a common operational voltage for the die.
22. The system of claim 17 , wherein the unreliable region has more storage capacity than the reliable region.
23. The system of claim 17 , wherein the reliable region is to include instructions and data, and the unreliable region is to be dedicated to data.