Do I need to align to page boundaries?

Question

0.00/5 (No votes)

See more:

I have decided to reinvent the wheel for the millionth time and write my own memory pool. My only question is about page size boundaries.

Let's say GetSystemInfo() call tells me that the page size is 4096 bytes. Now, I want to preallocate a memory area of 1MB (could be smaller, or larger), and divide this area into 128 byte blocks. HeapAlloc()/VirtualAlloc() will have an overhead between 8 and 16 bytes I guess. Might be some more, I've read posts talking about 60 bytes.

Question is, do I need to pay attention to not to have one of my 128 byte blocks across page boundaries?

Do I simply allocate 1MB in one chunk and divide it into my block size?

Or should I allocate many blocks of, say, 4000 bytes (to take into account HeapAlloc() overhead), and sub-divide this 4000 bytes into 128 byte blocks (4000 / 128 = 31 blocks, 128 bytes each) and not use the remaining bytes at all (4000 - 31x128 = 32 bytes in this example)?

Posted 3-Feb-14 8:03am

tablespoon

Updated 3-Feb-14 8:04am

v2

Add a Solution

Comments

Sergey Alexandrovich Kryukov 3-Feb-14 15:10pm

Not quite clear. You should understand that if you do something in one chunk, sub-chunks may not make any sense.
Perhaps you need to put more detail in your explanation of the problem, starting with the explanation of the goal of this activity.
—SA

tablespoon 3-Feb-14 16:05pm

I want to implement a memory pool. I want to allocate a large chunk with a system call, like HeapAlloc() or malloc(). Then I want to write my own routines to allocate/free memory blocks from this large chunk, using my own memory manager.

Question: Do I need to pay attention to the Page Size of the system when I return these small memory chunks to the caller of my routines?

Example: My initial chunk is 8000 bytes. This chunk fits into two pages of 4KB each. I suballocate blocks of 128 bytes to the caller of my routines from this chunk. I start allocating from the beginning of the chunk. At some point, one of my blocks would reside in two pages because 8000 is not divisible by 128.

Is it a bad idea to have my blocks spread to two pages? Would that be bad for performance?

[no name] 5-Feb-14 12:02pm

The page size isn't 4k, it is 4096. 4096 and 8192 are divisible by 128.

tablespoon 6-Feb-14 17:15pm

bling, in the example I allocated 4000 bytes because of the HeapAlloc()/malloc() memory header. I thought if I allocated a little less than the page size, I'd guarantee that my block would entirely reside in one page, and wouldn't cross the boundaries. I was wrong.

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

tablespoon · Accepted Answer · 2014-02-06T11:21:00

I've discovered that what I really needed was _aligned_malloc(), or some formulation to ensure that I'd always have my blocks on page boundaries.

HeapAlloc()/malloc() doesn't guarantee that any allocation will fit in a page. According to the documentation, I have blocks alinged on 8 bytes/16 bytes on 32 bit/64 bit systems, and that's all the alignement I get.

If I want to ensure page alignement, I'll have to accept some memory remaining unused and unaccessible.