On Mon, 27 Mar 2000, David Konerding wrote:
> OK thanks fro Doug Ledford and Oliver Xymoron, I was able to finally build
> a small test KNI app and confirm that it did indeed run on my P-III only
> when the necessary PIII/KNI patch was installed.
>
> This leads to me to two more questions:
>
> 1) Anybody know how to get gdb to dump KNI registers? The documentation
> from Intel says that the KNI registers are independent of the normal FP
> registers, it looks like gdb would need some extra code
> to grab the state of KNI and dump it
This currently isn't working, and you are right that it would require
extra code.
> 2) Does anybody know if I can use the "Four 32-bit single-precision (SP),
> floating-point numbers" described in the Intel manuals as two 64-bit double
> precision floating point numbers, and peform the KNI operations
> on those instead?
No. The registers are basically hardwired as either a 128bit register for
bitwise operations (aka, xorps, andps, etc) or as 4 32bit sp floating
point registers. How they are encoded is determined by the instruction of
course, and last I knew, there are no instructions to use them as 2 64 bit
entities.
-- Doug Ledford <dledford@redhat.com> Opinions expressed are my own, but they should be everybody's.- 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/
This archive was generated by hypermail 2b29 : Fri Mar 31 2000 - 21:00:21 EST