[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [microblaze-uclinux] News and updates



John,
I saw the following statement in one of your earlier mails to the mailing list.

"The HW acceleration is for single precision only - using doubles
(including literal constants like 2.0), will force calls to the software
libraries.  Since uClibc's math library does everything with doubles,
you won't see a lot of speedup on things like trig functions. However
you will see instant 20-100X speedup on raw floating point arithmetic in
your apps."


Does that mean all the double precision floating point will be
converted to single precision by the library to calculate the result.?
I am running ntpd which has lots of double. So putting a FPU in the
microblaze, does it increase the speed.?

Thanx
- Prasad


On 10/18/05, John Williams <jwilliams@xxxxxxxxxxxxxx> wrote:
Hi everyone,

Just an update on some changes recently commited to the CVS repository,
that should be good news for just about everyone.

Summary:

- Hardware floating point for libs/apps, if enabled in the CPU
- Xilinx ethernet driver now supports run-time setting of MAC address
- Hardware pattern compare support for libs/apps, if enabled in the CPU
- Minor fixups and speedups in the kernel.

Details below.

Have fun,

John



Drivers:
--------
Added suport for the SIOCHWADDR ioctl call in the xilinx EMAC driver.
You can now use ifconfig to set the MAC address:

$ ifconfig eth0 down
$ ifconfig eth0 hw ether 00:aa:bb:cc:dd:ee:ff
$ ifconfig eth0 up

Libs/Applications:
------------------
Hardware floating point now supported.  Thanks to the magic of
auto-config, it's all automatic.  If you are using MicroBlaze v4.00, and
you have the FPU enabled, then your auto-config settings will force the
uClinux libs and apps to be compiled with hardware FP support.

The HW acceleration is for single precision only - using doubles
(including literal constants like 2.0), will force calls to the software
libraries.  Since uClibc's math library does everything with doubles,
you won't see a lot of speedup on things like trig functions. However
you will see instant 20-100X speedup on raw floating point arithmetic in
your apps.

Hardware Pattern Comparison instructions (PCMP). This is now also
supported, through auto-config.  If it's enabled in your HW then the
kernel, libs and apps will be built with support for the instruction.

Kernel:
-------
Fixups in the kernel make files to ensure that we correctly link against
the appropriate GCC libraries.  You need to be using the latest (gcc-3.4
based) toolchain to take advantage.  This gives a nice speedup in the
kernel, if you have HW mul/div/shift enabled.

Committed interrupt controller fixups to CVS, as discussed recently.

Minor bugfix in handling of system calls with 4 parameters.

Support for kernel eXecute In Place (from Flash) for uclinux-auto and
ml401 platforms.

Other minor fixups.
--
Dr John Williams, Research Fellow,
Embedded Systems Group / Reconfigurable Computing
School of ITEE, The University of Queensland, Brisbane, Australia
(p) +61 7 33652185  (f) +61 7 33654999 (m) +61 403969243

___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/



___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/