Ok, I have seen various articles about the assiging RAM to IRQs via system.ini file.
Several say that no, it doesn't actually do anything while others says it does (Speedguide included).
What is and where can I find the definitive
answer?
Thanks,
Assigning RAM to IRQ - Work or Not?
There is none to date.
Group A says No, primarily because it isn't directly referenced in the microsoft docs.
Group B says yes, because of noticed effect, and because the lovebug exploit wasn't anywhere in the microsoft docs either...until it took down over ten thousand systems. Relying on MS documentation as the ultimate authority is NOT a good idea, because they are constantly finding new bugs and occaisionally features. The fact is, their code can be made to behave in ways they never intended. For example:
MinUserÐDiskSpace=4096 directly prohibts the swap file from accessing a certain amount of memory, in this case 4096K, or 4megs. In this way the system.ini file is directly controlling memory access of a device.
DEVICE=PATH\HIMEM.SYS /INT15=X
Where PATH is the physical location of the HIMEM.SYS file and /INT15=X tells Windows 95 to link INT15 (an area of memory reserved for miscellaneous systems services) to X, the amount of memory you wish to exclude.
In this config.sys file setting, we've again directly linked an interrupt with excepted memory.
I've done a bit of research on this and what I've found is the following. Unix languages DO allow for directly allocating resources to IRQ lines. So does the PnP specification, so does the programming a lot of device drivers are written in. Because Windows 95/98 are still built on the older DOS core (Yes, they are), they carry along traces of that core, including settings that have been basically unused since windows 3.1, but which still exist.
It's my working theory that what's happening is this is invoking the PCI-to-DRAM Buffer of the PnP specification. Remember that PnP is controlled by both Windows, AND the BIOS. Speedguide includes this tweak because quite a few people report success with it, and we'd rather present the option than not. Since even the sites that disagree with it agree that at worst it can do no harm, it's worth checking out.
BTW, oddly enough, many owners of GeForce cards report that this solved stuttering playback issues for them. I don't advocate it as a cure all, but it's an interesting tidbit. Because if true, then the setting IS doing SOMETHING, and it's a matter of tracking down exectly what. If it's doing nothing, than this placebo effect is somehow powerfully altering the perceptions of many users in a permanent way, which may be even more impressive
Regards,
-Bouncer-
------------------
"Yeah Baby, YEAH!!!"
Group A says No, primarily because it isn't directly referenced in the microsoft docs.
Group B says yes, because of noticed effect, and because the lovebug exploit wasn't anywhere in the microsoft docs either...until it took down over ten thousand systems. Relying on MS documentation as the ultimate authority is NOT a good idea, because they are constantly finding new bugs and occaisionally features. The fact is, their code can be made to behave in ways they never intended. For example:
MinUserÐDiskSpace=4096 directly prohibts the swap file from accessing a certain amount of memory, in this case 4096K, or 4megs. In this way the system.ini file is directly controlling memory access of a device.
DEVICE=PATH\HIMEM.SYS /INT15=X
Where PATH is the physical location of the HIMEM.SYS file and /INT15=X tells Windows 95 to link INT15 (an area of memory reserved for miscellaneous systems services) to X, the amount of memory you wish to exclude.
In this config.sys file setting, we've again directly linked an interrupt with excepted memory.
I've done a bit of research on this and what I've found is the following. Unix languages DO allow for directly allocating resources to IRQ lines. So does the PnP specification, so does the programming a lot of device drivers are written in. Because Windows 95/98 are still built on the older DOS core (Yes, they are), they carry along traces of that core, including settings that have been basically unused since windows 3.1, but which still exist.
It's my working theory that what's happening is this is invoking the PCI-to-DRAM Buffer of the PnP specification. Remember that PnP is controlled by both Windows, AND the BIOS. Speedguide includes this tweak because quite a few people report success with it, and we'd rather present the option than not. Since even the sites that disagree with it agree that at worst it can do no harm, it's worth checking out.
BTW, oddly enough, many owners of GeForce cards report that this solved stuttering playback issues for them. I don't advocate it as a cure all, but it's an interesting tidbit. Because if true, then the setting IS doing SOMETHING, and it's a matter of tracking down exectly what. If it's doing nothing, than this placebo effect is somehow powerfully altering the perceptions of many users in a permanent way, which may be even more impressive
Regards,
-Bouncer-
------------------
"Yeah Baby, YEAH!!!"