Messages in this thread | | | Date | Tue, 23 Feb 1999 06:47:35 -0600 (CST) | From | "Robert B. Hamilton" <> | Subject | Re: announce (experimental, i386 only): `nanokernel' #2 (PPSkit) |
| |
On Tue, 23 Feb 1999, Ulrich Windl wrote: > The 2^28 was my attempt (after little thinking) to avoid the > overflow, but it still overflowed, so I disabled it. The problem with > the limited number of bits is that you either have high precision or > wide range, bot not both.
OK, right. The precision is still very good anyway.
> You haven't changed the calibrate_tsc routine. Does it mean it works > for you?
Yep. For the record it reports a 267280004 Hz processor for a K6-266 The printk quotients are micro=16069168, nano=1004323704, both consistent.
>So maybe my CPU (100MHz Pentium) is too slow. As indicated > in the comment I had thought it should do from 64 MHz on (thus > including a Pentium 75). I don't have the code here now, but have I > missed something obvious (like the direction of the shift)?
The overflow was probably not in the calibration at all, but in time_init() just after the calibration, where cpu_hz was calculated. With the 4 shifts in the wrong direction it gives a factor of 256 too high. So 256*100*10^8 is much larger than 2^32.
Not so obvious though, since the difference is that the shifts are to make a change in the representation instead of a change in units.
-- Robert
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |