Instruction set architectures
have and use similar sound and operating instructions, such as add, xor, load, store and move.
The mapping you are looking for is found in the same place no matter what the two sets of instructions are. Get documents for one set of commands, get documents for another, and just find instructions. x86 loves to overload the mov instruction for downloads and storages, and also move and register it immediately to register movements, for the hand you use to load and store to load and store and move it for register registration and immediate registration.
intel is cisc, arm risc, so there will not be one card. very few commands will be displayed without visible finds, add rd = rd + rn and another register to register alu functions (for full-size registers, not fractional x86 registers).
As a result of not seeming, but not one to one, there are many real solutions for moving from one to another, so finding a single card does not make sense.
What you are looking for is either a static binary translation or a dynamic binary translation, static translation is easier if you have never done it. And do not translate from one machine code to another assembly or machine code. Translate from machine code to C, and then let the compiler compile it to the target during optimization, and also give you the choice to use different purposes. Another alternative is the instruction set simulator.
If you just want to rewrite the code manually, use two sources of guidance and targets. even if there was a card, you would still need to know two sets of instructions well to use this card.