Acceleration circuitry for posit operations

1. A method, comprising:

receiving, at a memory controller couplable to acceleration circuitry configured to perform arithmetic operations, logical operations, or both on bit strings formatted in a universal number (unum) or posit format, signaling indicative of performance of an operation to write a first bit string to a first buffer resident on the acceleration circuitry;

receiving, at the memory controller, signaling indicative of performance of an operation to write a second bit string to a second buffer resident on the acceleration circuitry;

transmitting, to the acceleration circuitry, signaling indicative of an arithmetic operation, a logical operation, or both, to be performed using the first bit string written to the first buffer and the second bit string written to the second buffer;

performing, via a plurality of computation circuits of the acceleration circuitry and according to the signaling received by the acceleration circuitry, sub-operations associated with the arithmetic operation, the logical operation, or both,

wherein each computation circuit is configured to perform at least one of the sub-operations; and

transmitting, to the memory controller, signaling indicative of a result of the arithmetic operation, the logical operation, or both.

2. The method of claim 1 , wherein the first bit string, the second bit string, or both is formatted in the unum or posit format.

3. The method of claim 1 , further comprising, prior to performing the arithmetic operation, the logical operation, or both, formatting, via the acceleration circuitry, the first bit string, the second bit string, or both in the unum or posit format.

4. The method of claim 1 , wherein formatting the first bit string, the second bit string, or both in the unum or posit format comprises converting, via the acceleration circuitry, the first bit string, the second bit string, or both from a floating point format to the unum or posit format.

5. The method of claim 1 , further comprising transmitting the signaling indicative of the result of the arithmetic operation, the logical operation, or both, from a third buffer resident on the acceleration circuitry, to the memory controller.

6. The method of claim 1 , wherein the acceleration circuitry is implemented on a field programable gate array (FPGA), and

wherein the method further comprises, accessing, via the memory controller and according to the signaling received by the memory controller, a memory resident on the FPGA to retrieve the first bit string, the second bit string, or both.

7. The method of claim 1 , further comprising accessing, via the memory controller and according to the signaling received by the memory controller, a main memory external to the acceleration circuitry to retrieve the first bit string, the second bit string, or both.

8. The method of claim 1 , further comprising transmitting, from a host couplable to the acceleration circuitry, the signaling indicative of the first bit string, the second bit string, or both to a memory couplable to the acceleration circuitry.

9. The method of claim 8 , further comprising transmitting the signaling indicative of the first bit string, the second bit string, or both from the host to the memory via a Peripheral Component Interconnect Express (PCIe) interface.

10. The method of claim 1 , further comprising transmitting, from a host couplable to the acceleration circuitry, the signaling indicative of the arithmetic operation, the logical operation, or both to a memory couplable to the acceleration circuitry.

11. The method of claim 10 , further comprising transmitting the signaling indicative of the arithmetic operation, the logical operation, or both from the host to the memory via a Peripheral Component Interconnect Express (PCIe) interface.

12. A method, comprising:

receiving, at a first buffer resident on acceleration circuitry configured to perform arithmetic operations, logical operations or both on bit strings formatted in a universal number (unum) or posit format, signaling indicative of a first bit string formatted in the unum or posit format;

receiving, at a second buffer resident on the acceleration circuitry, signaling indicative of a second bit string formatted in the unum or posit format;

transmitting, to processing circuitry resident on the acceleration circuitry, signaling indicative of an arithmetic operation, a logical operation, or both; and

performing, via a plurality of computation circuits of the processing circuitry, sub-operations associated with the arithmetic operation, the logical operation, or both using the first bit string and the second bit string,

wherein each computation circuit is configured to perform at least one of the sub-operations.

13. The method of claim 12 , further comprising transmitting, from the processing circuitry, a result of the arithmetic operation, the logical operation, or both to a third buffer resident on the acceleration circuitry.

14. The method of claim 12 , further comprising;

transmitting, from a memory couplable to the acceleration circuitry, the signaling indicative of the first bit string; and

transmitting, from the memory couplable to the acceleration circuitry, the signaling indicative of the second bit string.

15. The method of claim 12 , further comprising:

transmitting, from a first memory component couplable to the acceleration circuitry, the signaling indicative of the first bit string; and

transmitting, from a second memory component couplable to the acceleration circuitry, the signaling indicative of the second bit string,

wherein the first memory component is distinct from the second memory component.

16. The method of claim 12 , further comprising

parsing, at a processing device couplable to the acceleration circuitry, host-level instructions associated with the arithmetic operation, the logical operation, or both into machine-level instructions; and

transmitting, to the processing circuitry, signaling indicative of the machine-level instructions.

17. A system, comprising:

a host; and

a memory device couplable to the host and comprising a plurality of compute circuits,

wherein the memory device is configured to:

receive, from the host, signaling indicative of host-level instructions to perform arithmetic operations, logical operations, or both on bit strings formatted in universal number (unum) or posit format;

parse the signaling indicative of host-level instructions to perform the arithmetic operations, the logical operations, or both into signaling indicative of machine-level instructions to perform sub-operations associated with the arithmetic operations, the logical operations, or both; and

perform, according to the signaling indicative of the machine-level instructions, the sub-operations using the signaling indicative of the bit strings,

wherein each computation circuit is configured to perform at least one of the sub-operations.

18. The system of claim 17 , wherein the memory device is further configured to:

determine whether the machine-level instructions have been executed; and

responsive to determining that the machine-level instructions have been executed, transmit, to the host, signaling indicative of performance of the arithmetic operations, the logical operations, or both.

19. The system of claim 17 , wherein the memory device is further configured to:

receive, at a first buffer, signaling indicative of a first bit string formatted in the unum or posit format;

receive, at a second buffer, signaling indicative of a second bit string formatted in the unum or posit format; and

perform an arithmetic operation, a logical operation, or both using the first bit string and the second bit string.

20. The system of claim 17 , wherein the signaling indicative of the host-level instructions is indicative of the host-level instructions to perform a matrix multiplication operation, and

wherein each computation circuit is configured to, according to the signaling indicative of the machine-level instructions, perform a dot product operation of the matrix multiplication operation.