Re: NETTYPE question

tienp@wholefoods.com writes: > I have a question regarding the NETTYPE parameter. Which one of the > following is more efficient? IDS 7.x? Probably the first > And does each configure communication > parameters for 256 users? No, the config is for 256 connections (read more below) > In other words, does the first NETTYPE > parameter configure 4 poll threads of class CPU each capable of handling > 64 users? Yes (and no) > > NETTYPE ipcshm,4,64,CPU > NETTYPE ipcshm,2,128,CPU From "Performace Guide" (9.2): "Tip: For best performance, Informix recommends that you allocate enough semaphores for double the number of ipcshm connections that you expect. Informix also recommends that you use the NETTYPE parameter to configure database server poll threads for this doubled number of connections. For a description of poll threads, refer to your Administrator s Guide. For information on configuring poll threads, refer to NETTYPE on page 3-16." and "The optimum number of connections per poll thread is approximately 300 for uniprocessor computers and 350 for multiprocessor computers. However, a poll thread can support 1024 or perhaps more connections." And acording to Art S. Kagel: (http://x25.deja.com/[ST_rn=ps]/getdoc.xp?AN=580083161&search=thread&CONTEXT=955472946.513277978&HIT_CONTEXT=955472847.514195458&HIT_NUM=1&hitnum=0) "If you place the shared memory connections in CPU VPs (and use all 3 CPU VPs) then since the CPU VPs cannot block, since they have work to do, they will periodically poll the shared memory blocks for connections and new data and will be able to handle the requests quickly. If all the CPU VPs are listening then requests do not have to wait for CPU VP#1, which is the busiest always, to become free to schedule the new request." Which boils down to: NETTYPE ipcshm,6,80,CPU > I have 6 CPUvps configured for this system. If responding, please copy > email address (tienp@wholefoods.com). Thanks. But I _guess_ this is hardware/OS dependant: On Intel (Linux, NT) and HP (HP_UX) there's some issues with too (in my eyes) few memory-registers, so you'd probably want to keep the number of shared-memory segments down... If someone has done some testing here, I'd _love_ some numbers;) Thomas