Memory chip connecting a system on a chip and an accelerator chip
Abstract
A memory chip (e.g., DRAM) connecting a SoC and an accelerator chip (e.g., an AI accelerator chip). A system including the memory chip and the accelerator chip. The system can include the SoC. The memory chip can include first memory cells to store and provide computation input data (e.g., AI computation input data) received from the SoC to be used by the accelerator chip as computation input (e.g., AI computation input). The memory chip can include second memory cells to store and provide first computation output data (e.g., AI computation output data) received from the accelerator chip to be retrieved by the SoC or reused by the accelerator chip as computation input. The memory chip can also include third memory cells to store second computation output data (e.g., non-AI computation output data) related to non-AI tasks received from the SoC to be retrieved by the SoC for non-AI tasks.
Claims
exact text as granted — not AI-modifiedWhat is claimed is:
1 . A system, comprising:
an accelerator chip; and a memory chip, comprising:
a first set of pins configured to connect to a system on a chip (SoC);
a first plurality of memory cells configured to receive and store computation input data received from the SoC, via the first set of pins, to be used by the accelerator chip as computation input.
2 . The system of claim 1 , wherein the accelerator chip is an artificial intelligence (AI) accelerator chip, and wherein the first plurality of memory cells is configured to store and provide AI computation input data received from the SoC, via the set of pins, to be used by the AI accelerator chip as AI computation input.
3 . The system of claim 1 , wherein the memory chip comprises:
a second set of pins configured to connect to the accelerator chip; and a second plurality of memory cells configured to store and provide first computation output data received from the accelerator chip, via the second set of pins, to at least one of be retrieved by the SoC or reused by the accelerator chip as computation input.
4 . The system of claim 3 , wherein the memory chip comprises a third plurality of memory cells configured to store second computation output data received from the SoC, via the first set of pins, to be retrieved by the SoC.
5 . The system of claim 4 , wherein the first, second, and third pluralities of memory cells comprise dynamic random-access memory (DRAM) cells.
6 . The system of claim 4 , wherein the first, second, and third pluralities of memory cells comprise non-volatile random-access memory (NVRAM) cells.
7 . The system of claim 6 , wherein the NVRAM cells comprise 3D XPoint memory cells.
8 . The system of claim 3 , wherein the accelerator chip comprises a vector processor that is configured to perform numerical calculations on vectors and matrices for the SoC using the first and second pluralities of memory cells as memory.
9 . The system of claim 8 , wherein the accelerator chip comprises an application-specific integrated circuit (ASIC) that comprises the vector processor and is configured to accelerate AI computations through the vector processor.
10 . The system of claim 8 , wherein the accelerator chip comprises field-programmable gate arrays (FPGA) that comprise the vector processor and are configured to accelerate AI computations through the vector processor.
11 . A system, comprising:
an accelerator chip; a system on a chip (SoC); and a memory chip, comprising:
a first set of pins configured to connect to the SoC;
a first plurality of memory cells configured to receive and store computation input data received from the SoC, via the first set of pins, to be used by the accelerator chip as computation input.
12 . The system of claim 11 , wherein the memory chip comprises:
a second set of pins configured to connect to the accelerator chip; and a second plurality of memory cells configured to store and provide first computation output data received from the accelerator chip, via the second set of pins, to be retrieved by the SoC or reused by the accelerator chip as first computation input.
13 . The system of claim 12 , wherein the SoC comprises a graphics processing unit (GPU), and wherein the accelerator chip is configured to perform and accelerate computations for the GPU using the first and second pluralities of memory cells as memory.
14 . The system of claim 13 , wherein the accelerator chip comprises a vector processor that is configured to perform numerical calculations on vectors and matrices for the GPU using the first and second pluralities of memory cells as memory.
15 . A memory chip, comprising:
a first set of pins configured to connect to a system on a chip (SoC); and a first plurality of memory cells configured to receive and store artificial intelligence (AI) computation input data received from the SoC, via the first set of pins, to be used by an AI accelerator chip as AI computation input.
16 . The memory chip of claim 15 , comprising:
a second set of pins configured to connect to the accelerator chip; and a second plurality of memory cells configured to store and provide AI computation output data received from the AI accelerator chip, via the second set of pins, to at least one of be retrieved by the SoC or reused by the AI accelerator chip as AI computation input.
17 . The memory chip of claim 16 , comprising a third plurality of memory cells configured to store non-AI data related to non-AI tasks received from the SoC, via the first set of pins, to be retrieved by the SoC for non-AI tasks.
18 . The memory chip of claim 17 , wherein the first, second, and third pluralities of memory cells comprise dynamic random-access memory (DRAM) cells.
19 . The memory chip of claim 17 , wherein the first, second, and third pluralities of memory cells comprise non-volatile random-access memory (NVRAM) cells.
20 . The memory chip of claim 19 , wherein the NVRAM cells comprise 3D XPoint memory cells.Join the waitlist — get patent alerts
Track US2022300437A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.