Little questions about GT.M
When I run two GTM> + GTM> process on single computer/host – they have two different “Global Buffers” or one unite buffer in memory? There are one copy ^Global in memory or two identical copies ^Global + ^Global for each process?
And this identical question - when we change “access_method” from BG to MM – There are one copy ^Global in memory or two identical copies ^Global + ^Global for each process?
May I run remote m-routine in host2 from GTM> in host1 and receive answer to host1?
64-bit OS supported with big memory?
1. There is only 1 set of global buffers.
2. MM is not supported on UNIX/Linux - only on OpenVMS.
3. You can use GT.CM to connect a GT.M process on one machine to a server on another.
4. GT.M is reported to run on 64 bit GNU/Linux as a 32-bit process.
As concept - may GT.M read from hard disk one ^Global with 10-20-30-100 Gigabyte of size into server RAM memory to support hundreds random $Order read users processes – for disk safe using? Sorry for my English :)
The minimum BLOCK_SIZE is 512 bytes -
The maximum BLOCK_SIZE is 65,024 bytes.
The minimum for GLOBAL_BUFFER_COUNT is 64 blocks.
The maximum for GLOBAL_BUFFER_COUNT is normally 65,536 blocks
65024*65536 = 3,96 Gigabyte max buffer size - it is limit 32 bit OS or limit GT.M “buffer engine idea” too?
If mainly OS - may I transfer/compile source “32-linux GT.M” to “64-linux GT.M” :) ?