I desire to set my MAX cache allocation to 64 MByte, does anyone know how this can be done on a current kernel?

I am having a problem with programs being killed by the kernel when an out of memory occurs, and yet the cache allocation remains about 40% to 50% of my RAM. When the cache allocation is above 50%, the kernel will decrease cache and then it decreases programs. It appears to kill the largest running program first. This sacrafice from the kernel is not concerned if the data file is saved or if it is actively busy like printing.

I am using kernel 2.4.18 and it does not have the pagecache tuneable item in the VM system anymore. Sysctl does not list any item that would at least appear to replace it. I am configured for 384 MByte RAM and no swap space.

I have checked the 2.4.19 and 2.4.20 kernels and they do not offer the setting either. Any suggestions or thoughts on how the cache memory allocation can be managed would be greatly appreciated.


<Followup>

Even though 2.4.19 does not have the pagecache setting is does improve the VM management considerably. I ran across a thread discussing 2.4.19 swap management and how improved it was over 2.4.18. I was not able to duplicate on my notebook under 2.4.19, so I figured it was at least worth a try. I had other problems attempting 2.4.20 that I will have to delay that for a awhile until the issues are resolved. My initial tests showed I could open about three times the programs before any were terminated. The cache size was down to 94 MByte when that happened, about 100 MByte less than before. So at this point, time will tell.

Thanks to everybody that read the post and considered the problem.

</Followup>