In an x86-P6 based computer system, does PC BIOS map the memory prior to startup?

Does the onboard firmware (BIOS) map the memory on power up before OS lookup or Power On Self Test?

This is arbitrary since every BIOS is different.

Would the process vary based on firmware, microarchitecture, or any other manufacturing process?

I know about CPUID for memory map info and such, but I am asking specifically about the mapping process, not the information.

EXTRA: This would mean that if the firmware mapped the memory (and not manufacturers of the motherboard, i.e., unchangeable) that would mean that memory can be mapped manually by editing BIOS instructions, or "re-programming"/updating the chip.

In old computer systems, like the NES, Gameboy, Segas, etc., consoles did not use firmware (ok, some did, but most didn't back then), so all of the memory was set and permanent for specific purposes, and not changeable by firmware, if the firmware is what's mapping the memory map in the first place.

While burning your own BIOS would probably be a fun project, I'm pretty sure that at this point in the PC lifecycle, the BIOS/UEFI doesn't do much with the memory mapping. The memory controller and all of it's functions now live on-die with the CPU. I'm not sure you can get the control you are talking about at such a low level these days with an x86 based CPU, at least not modern ones.

Then again, I know almost nothing about the subject. Even my basic POST knowledge has started to fade away these days.

Any links or further reading to help us understand what you are trying to do?

You obviously don't know much about memory mapping, as apparent in your response, you state BIOS or UEFI/EFI firmware doesn't map memory and insist a memory controller does which, according to Wikipedia's sources claim, a memory controller:

Quote:

---

is a digital circuit which manages the flow of data going to and from the main memory.

---

Controlling the data (bits) going to and from main memory has no effect on mapped RAM in an address space circuit.

That would not be the process or either mapping the memory, or handling the mapped memory writes, as an address decoding circuit would likely be the control of flow, including the control unit itself.

You say you know nothing of the subject, but insist burned and etched firmware has no role, and that a memory controller, which refreshes DRAM with current and uses multiplexing/demultiplexing circuit logic, has some say.

For knowing "nothing" you sure do insist a lot of things.

At the risk of feeding the troll....

1. I never insist anything.
2. I admitted I know little to nothing about memory mapping in general. I just know that many of the functions that used to be handled by the motherboard and northbridge/southbridge have been migrated within the CPU die itself.

Oh, I get it. You think I'm a troll because I didn't candy coat your answers.

Just because one expresses distaste and disagreement with you doesn't make them a troll.

You could have simply stated, "I don't know the answer", or maybe not have bothered answering at all, if your answer was going to contain arbitrary info, or blatant speculation of the subject.

At least answer on topic, rather than just shift away and go about on your own business.

Quote:

---

Originally Posted by James

At the risk of feeding the troll....

1. I never insist anything.
2. I admitted I know little to nothing about memory mapping in general. I just know that many of the functions that used to be handled by the motherboard and northbridge/southbridge have been migrated within the CPU die itself.

---

LOL. I get the feeling based on his few posts this guy is asking questions on homework or quiz from school. And Wikipedia... my favorite citation to have student re-write....

I'm not trolling I'm serious.

I want to be an electronic engineer, IC designer and developer, and computer scientist/programmer.

I am sufficient at most C, C++, and Java basics, I have studied Assembly for x86, Motorola 68000, Ricoh 5A22, even Disassembly.

I am 100% serious, and hope to be successful and accomplished in all things related to computing.

Quote:

---

Originally Posted by Big Tunda

I'm not trolling I'm serious.

I want to be an electronic engineer, IC designer and developer, and computer scientist/programmer.

I am sufficient at most C, C++, and Java basics, I have studied Assembly for x86, Motorola 68000, Ricoh 5A22, even Disassembly.

I am 100% serious, and hope to be successful and accomplished in all things related to computing.

---

If that's the case, I wish you the best of luck!

Sadly, most of my hardware knowledge is of the macro variety. I can put together a datacenter and create large systems from off-the-shelf parts, but I can't design a chip itself. Sometimes I wish I could though. Also, maybe I'm overreaching here, but I don't know that we have many EE's or chip designers on the forums. Mostly general knowledge computer enthusiasts, a few good system engineers, etc. (See: Vertices and others). If you decide to stick around, I would be interested to hear how you progress.

Can you explain how the BIOS works with memory mapping? From what I understood, the BIOS was just in charge of getting the computer up and running, and then letting the OS communicate with the hardware.

I will be sticking around for now. And yes, as I said, I don't know if the firmware is what maps the memory out from RAM, or if it's an on-die controller, etc.

Unless the memory was "hard-wired", something must be able to change it.