Thanks for following up. I think you have solved my problem. The reason I wasn't able to find the MEMORY.DMP file is that it did not create one. And the reason it didn't create one is given in the first link you provided.
Namely, the page file must be located on the boot drive for the memory dump to work, and I currently have my page file on another drive.
So I'll move the page file and wait for the BSOD to happen again.
The difficult we do right away...
...the impossible takes slightly longer.
Go to the system advanced and in start up and recovery set the machine to produce a memory dump, you can chose full or mini. Mini is OK ish, in which case its written to system\minidumps as a dmp file with the date/time in the name. If you chose full dump its written to \system\memory.dmp and overwrites any existing one.
To look at the contents use windbg, and open the file.
Then run !analyze -v from the command line (bottom of windbg window) and it will do an autoanalyse. Be careful, these aren't always accurate and often blame the wrong component.
Now you need to debug the issue. Commands like !poolused can hep you find drivers that are using excessive memory, !thread and .thread are very useful to switch to any thread in the system so you can look at each threads stack and wait times.
In fact what you can do with windbg is immense, its got a massive and powerful command set. There is a doc that comes with windbg on basic debugging practices, it is very worth reading.
Even though you wrote that you keep paging file in the RAM disk, one question is, do you have enough available RAM for the RAMDisk? The performance would be catastrophic if the OS is paging anyway.
Also I don't believe that it would make sense to use RAM for paging file. The purpose of the paging file is basically to provide additional space if RAM is exhausted. I believe it would be better to let the OS use it for file caching etc.
Sorry if I am not clear, but I am using 32 bit OS, so you can access only about 3GB of RAM for OS. Since I already have 16GB of memory installed which is not being utilized, I made a 8GB RAMDisk using the unutilized RAM which I use for paging file.
I have a pretty much brand new system with 16GB DDR4 memory on which I have Win
7 32bit OS.
That almost equals my setup, but with DDR3 and a laptop CPU
Kiran Satish wrote:
I made a RAMDisk (using Dataram RAMDisk) of 8GB on which I keep windows paging
Ehr.. aight, we have a paging file because PC's have a limited amount of memory, and lots of harddisk space. The only reason to put something in "virtual memory" is because it is too expensive to put it in real memory (aka, hardware RAM).
Putting the page-file on a RAM-disk defeats the purpose of having a page-file. I also have an SSD, and did not believe there should be a page-file with 16Gb worth o' memory. I have had recommendations here to do otherwise and am happy to have followed the advice.
Kiran Satish wrote:
Having created my own ram-disk[^], I would say that performance is limited in a number of ways. One of those is the limitation of the speed of your memory; virtual memory is slower than blocks of data pinned into physical memory. Did that make my RAM-disk faster? No, it is limited by the driver itself, bottlenecking at 48 Mb/s, which is actually rather impressive.
Works nice if the RAM-disk leaves at least 6 Gb of free working memory; it is fast in terms of access, making up for the lack in terms of transfer, and thus doing nice when manipulating lots of small files. Like compiling and building a project
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
it's probably a stupid question but when packets are sent to a router that is for example at 192.168.1.1 on the LAN with a destination 192.168.1.4
the router searches a lookup table for the MAC associated with 192.168.1.4 and sends the frame just there and doesnt forward it to all connected devices
so does that count as a switch ?
well it wouldn't be the router searching the ARP table. The sending client would do the lookup in its own ARP table. If it didn't have an address for x.x.1.4, it would do an ARP broadcast to get it.
I don't think the router has a role in this situation. Why? because if the default route is 192.168.1.1 then it is probably already on that subnet. The client is going to realize the destination is on the same subnet using the subnet mask.
If the router had an integrated switch, the switch would be involved in forwarding the frame. Anyways my $0.02
no I'm not talking about lan communication
suppose you send a packet to codeproject from your pc.
its trip will be (with all address change of course) 192.168.1.4->192.168.1.1(def.gate.)->10 more machines->codeproject
say your external ip is 100.100.100.100
when its reply gets at 100.100.100.100 your router is in charge of sending it to x.x.x.4...I'm talking about this bit
My knowledge is a bit rusty, but routing and switching occur at different protocol levels. Routing is on the network layer (like IP) and switching on the link layer (e.g.: Ethernet)
I guess you are talking about a router with multiple LAN ports. I don't think there are any more unmanaged switches around, where packets are unconditionally sent out on all ports; the switch knows what devices are on which port, and will only transmit on the appropriate one (it will have to use address resolution if it doesn't know the MAC address, but this only happens once, or when the device is moved to another port)
If the destination IP address is not on the local network, the router will naturally forward it to the WAN side.
I am setting up tablet for our customer.
It has 3 network interfaces: WiFi, LET/4G/Phone network and a slow but (allegedly) reliable IP radio network.
I want to make sure the Radio network is ONLY used by my application (and NO OTher App use it for "internet" acceess)
I try to call (as admin) the ROUTE (DOS) command with ROUTE DELETE AND ROUTE ADD.
Our project seems to be heading in the need of having a small computer to mediate between the UI control/data presentation Windows PC (workstation) and an OEM vendor's device.
The system must support at least 1 USB 2.0 interface (to the device) and some other interface (>= 100BaseT ?) to the workstation. The vendor provides an x86 DLL to encapsulate communication with their device so this will need to run 32-bit Windows Embedded Compact (for the real-time capability as well).
Smaller physical size is better, as is lower cost.
Does anyone have any experiences and/or recommendations (and/or who/what to avoid!) they're willing to share about such systems?
"Fairy tales do not tell children the dragons exist. Children already know that dragons exist. Fairy tales tell children the dragons can be killed."
- G.K. Chesterton
Last Visit: 31-Dec-99 18:00 Last Update: 2-Aug-21 6:46