| ??? 09/02/03 10:08 Read: times |
#53856 - RE: hyperkernel for NT/2000 Responding to: ???'s previous message |
This is a very interesting product, however their documentation goes on to say that neither they nor any OS can actually guarantee a certain level of interrupt latency. The problem seems to be the "SMI interrupt" available on x86 processors which they say preempts and disables all other interrupts and is handled entirely in hardware. The length of time taken to service the interrupt is determined by the motherboard manufacturer.
So, it looks as though it is actually impossible for any OS to be truly deterministic on an x86 PC (assuming that deterministic interrupt latency is what we're measuring the 'real timeness' of an OS in this week). Clearly the solution would be to port windows to the 8051. |
| Topic | Author | Date |
| is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: hyperkernel for NT/2000 | 01/01/70 00:00 | |
| RE: hyperkernel for NT/2000 | 01/01/70 00:00 | |
| RE: hyperkernel for NT/2000 | 01/01/70 00:00 | |
RE: hyperkernel for NT/2000 | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: is it not possible to run real time ? | 01/01/70 00:00 | |
| RE: Real time - Raj Nambiar | 01/01/70 00:00 |



