Logic Circuit Assembler (LCA) / Logic Gate Assembler (LGA)
Java OpenCL Logic Circuit Simulator for simulating and debugging fully pipelined binary gate logic. Includes visual designer that also converts OpenCL C code to binary micro-fpga gate logic.
- Not designed to be sequential operation execution similar to assembly code, but rather be a continuous execution circuit definition language with core-width x pipeline-depth input work item dimensions.
- System architecture is based on 1-cycle latency FPGA gates and large SRAM block with three full-length sram block mimo OR-multiplexers to read/write int32 argument and store pointer values directly for each gate.
- Any external communication to the logic gate system is through SRAM direct read/write, from such as PCIe, USB, SD-card, HBM or DDR5 memory bridge controllers.
- Each separately programmable/assignable micro-FPGA gate runs internally at a multiplier speed of the main circuit clock speed to enable one clock cycle per gate operation.
- Programmer/IDE is responsible of assigning correct output pointer values for each gate considering multiple OR-multiplexed value store collision.
Language text/binary syntax:
Code format: 4-int32 blocks of [OPERATION ARGUMENT-PTR1 ARGUMENT-PTR2: STORE-PTR1].
Value format: 32-bit shared integer and floating point values (int32 and fp32).
Run format: new-old value store-update each gate once per clock cycle.
Operation list:
0=BUF: delay buffer arg1
1=NOT: bitwise invert arg1
2=AND: bitwise arg1 and arg2
3=OR: bitwise arg1 or arg2
4=XOR: bitwise arg1 xor arg2
5=NAND: bitwise arg1 nand arg2
6=NOR: bitwise arg1 nor arg2
7=XNOR: bitwise arg1 xnor arg2
8=SHL: bitwise arg1 shift left by arg2
9=SHR: bitwise arg1 shift right by arg2
10=NEGi: negate arg1
11=SUMi: arg1 sum arg2
12=SUBi: arg1 minus arg2
13=MULi: arg1 multiply arg2
14=DIVi: arg1 division by arg2
15=COS: cos arg1 (float)
16=SIN: sin arg1 (float)
17=TAN: tan arg1 (float)
18=ACOS: acos arg1 (float)
19=ASIN: asin arg1 (float)
20=ATAN: atan arg1 (float)
21=LOG: log arg1 (float)
22=EXP: exp arg1 (float)
23=POW: arg1 power to arg2 (float)
24=SQRT: sqrt arg1 (float)
25=NROOT: nth-root arg1 (float)
26=NULL: NULL value (0)
27=ITOF: convert arg1 to fp32 value
28=FTOI: convert arg1 to int32 value
29=MGET: get arg1 pointer value
30=MSTO: store arg1 value to arg2 pointer
31=IFBUF: delay buffer arg1 if arg2 is 1, otherwise NULL (0)
32=NEG: negate arg1 (float)
33=SUM: arg1 sum arg2 (float)
34=SUB: arg1 minus arg2 (float)
35=MUL: arg1 multiply arg2 (float)
36=DIV: arg1 division by arg2 (float)
User defined function blocks (memory-height x pipeline-depth):
#<NAME> <ARG1-ptr-in> <ARG2-ptr-in> ... <ARGn-ptr-in>: <STO1-ptr-out> ... <STOn-ptr-out>
<code blocks>
###
Comments:
//comment