If you're going to play the "what I said was strictly semanticaly right, don't really care about the implications" card, let me remind you your initial statement And I don't know where you trying to go to. memory-mapped data file is the same thing as file mapping view

The central question we have been discussing is whether virtual memory hard page faults are possible without a paging file, and the answer still is "no". Cheers, - jahman.

The central question we have been discussing is whether virtual memory hard page faults are possible without a paging file, and the answer still is "no". Cheers, - jahman.

Wrong jahman. You'll get a hard page fault once a file mapped page that was dropped is requested again. That was exactly was I meant with all the file mapping view thing. Not only you will get a hard fault, it will happen precisely because there is no page file available to the virtual memory manager to page out something else that was more suitable to be paged outSo hard faults can happen without a page file, and more importantly, disabling the page file can induce hard page faults on itself

Wrong jahman. You'll get a hard page fault once a file mapped page that was dropped is requested again. That was exactly was I meant with all the file mapping view thing. Not only you will get a hard fault, it will happen precisely because there is no page file available to the virtual memory manager to page out something else that was more suitable to be paged outSo hard faults can happen without a page file, and more importantly, disabling the page file can induce hard page faults on itself

Read my lips: No paging file, no hard page faults for virtual memory. Period. End of story. You keep insisting abour memory-mapped files: Memory mapped files are something different and separate from virtual memory and they definitely do not use a page swap file. <--- You need to understand this! A memory-mapped file is its own page file!) Cheers, - jahman.

Read my lips: No paging file, no hard page faults for virtual memory. Period. End of story. You keep insisting abour memory-mapped files: Memory mapped files are something different and separate from virtual memory and they definitely do not use a page swap file. Cheers, - jahman.

Separate thing from virtual memory? and where are mapped files then? in the Windows file limbo? lolNo, they don't use the page file, they point directly to the HDD location where the file sits. That's exactly why disabling the page file can't prevent the VMM from discarding those pages from RAM and pointing them back to disk without a page file

Separate thing from virtual memory? and where are mapped files then? in the Windows file limbo? lolNo, they don't use the page file, they point directly to the HDD location where the file sits. That's exactly why disabling the page file can't prevent the VMM from discarding those pages from RAM and pointing them back to disk without a page file

You are confused: With memory-mapped files there is no separate paging file involved because the paging file is the data file. They are one and the same thing. Cheers, - jahman.

You are confused: With memory-mapped files there is no separate paging file involved because the paging file is the data file. They are one and the same thing. Cheers, - jahman.

I'm not confused jahman. I never said there is a "separate page file". I said that the VMM can drop a page's reference to RAM in VM to free up RAM, and point that page back to the HDD where the file sits. something like C:/Windows/myDLL.dll

If you take the time to read the link I posted, you'll see that memory mapped files do not count for the commit limit, which is the total virtual memory available to the system based on its configuration of RAM + page file(s) size.So if it's in VM but it doesn't count for the commit limit (meaning that it's not either in RAM nor in the page file(s)) where are those VM pages pointing to then? yes, you guessed it... the HDD

I'm not confused jahman. I never said there is a "separate page file". I said that the VMM can drop a page's reference to RAM in VM to free up RAM, and point that page back to the HDD where the file sits. something like C:/Windows/myDLL.dll

Yes! What you're seeing here is the VMM is polymorphic: It is being used for VM and for memory-mapped files. But with memory-mapped files we're no longer talking about virtual memory and the page swap file that is needed for virtual meory to work, and the fact that you will only get page hard faults related to virtual memory when that page swap file is present, which was the original question. So to recap, "no memory page file, no virtual memory hard page faults". Cheers, - jahman.

Yes! What you're seeing here is the VMM is polymorphic: It is being used for VM and for memory-mapped files. But with memory-mapped files we're no longer talking about virtual memory and the page swap file that is needed for virtual meory to work, and the fact that you will only get page hard faults related to virtual memory when that page swap file is present, which was the original question. So to recap, "no memory page file, no virtual memory hard page faults". Cheers, - jahman.

Now you're making things up, hahaha. How does the OS access the memory mapped files if it's not via virtual memory?what's that polimorphic VMM? if you're not going to read the link I posted, for pete's sake, at least read yours! http://en.wikipedia.org/wiki/Memory-mapped_file

A memory-mapped file is a segment of virtual memory which has been assigned

Now you're making things up, hahaha. How does the OS access the memory mapped files if it's not via virtual memory?what's that polimorphic VMM? if you're not going to read the link I posted, for pete's sake, at least read yours! http://en.wikipedia....ory-mapped_file

Dazz, the complete qute should be "A memory-mapped file is a segment of virtual memory which has been assigned a direct byte-for-byte correlation with some portion of a file or file-like resource ", in other words the MMF is not using the VM page file, OK? (And the discussion was always about the benefits of using/not using a memory page file.) Cheers, - jahman.

Dazz, the complete qute should be "A memory-mapped file is a segment of virtual memory which has been assigned a direct byte-for-byte correlation with some portion of a file or file-like resource ", in other words the MMF is not using the VM page file, OK? (And the discussion was always about the benefits of using/not using a memory page file.) Cheers, - jahman.

wow, you can't be serious. a direct byte-for-byte correlation with some portion of a file or file-like resource that's a generic way of saying a pointer to secondary storage, aka HDD, and a portion of a file, not necessarily an entire file (because the virtual memory works with pages, not files)I highlighted the first part of the definition where it clearly relates to virtual memory. If you are going to interpret that final part to exclude virtual memory then I'm done with this

the MMF is not using the VM page file

*facepalm*

wow, you can't be serious. a direct byte-for-byte correlation with some portion of a file or file-like resource that's a generic way of saying a pointer to secondary storage, aka HDD, and a portion of a file, not necessarily an entire file (because the virtual memory works with pages, not files)I highlighted the first part of the definition where it clearly relates to virtual memory. If you are going to interpret that final part to exclude virtual memory then I'm done with this *facepalm*

Yes, its pointing to the memory-mapped file, of course! But it is *not* pointing to the VM paging file! You are still confusing both concepts: You can have MMF without VM and you can have VM without MMF and you can have both VM and MMF. They are i-n-d-e-p-e-n-d-e-n-t of each other.And the VM Manager is mentioned because the VMM performs both tasks because they are very similar, but the tasks themselves are separate. Cheers, - jahman.

You completely fail to understand what I'm saying (or deliberately refuse to). You insist in the page file, and keep on saying that I need the page file for memory mapped files.Care to elaborate on how you can have MMF outside the VM scope?that's a challenge

You completely fail to understand what I'm saying (or deliberately refuse to). You insist in the page file, and keep on saying that I need the page file for memory mapped files.Care to elaborate on how you can have MMF outside the VM scope?that's a challenge

No Dazz, you are misunderstanding me (and please tone down your aggressiveness if you want this exchange to continue.)Now re-read my posts and you will see I never say that you need a page file for memory-mapped files. I'm actually saying exactly the opposite! I do insist on the page file as the whole point of this discussion because this exchange started with whether or not you could have virtual memory hard page faults without a memory page file, to which the answer still is "no". Cheers, - jahman.

because this exchange started with whether or not you could have virtual memory hard page faults without a memory page file, to which the answer still is "no". Cheers, - jahman.

Explain this please. I already argumented why that's wrong. You can have hard page faults. And hard page faults are only possible in the virtual memory scope. If you know any other way it could happen, please elaborateBy challenge I didn't mean it in an aggressive way, sorry about that, I got lost in translationI mean that it's going to be challenging for you to separate VM from MMF