Hi all,
I have had non-stop CPU spikes and freezes related to denormal numbers
on my Pentium 4 1.6GHz laptop ever since I switch to Linux, going from
kernel 2.4.18 up to 2.6.4.
The main applications which seem to be affected are Freqtweak, Jamin and
Pure Data, along with a handful of LADSPA
Universal Audio is doing that for years with their UAD-1 card.
It's a videoboard without videoconnector with the programmable mpact2
multimedia chip.
Nothing to worry about.
Universal Audios UAD card is in fact a GPU (MPACT2 chip), so using a GPU is
not that new.
Anyway, call it GPU, DSP, CPU or whatever else. They are chips supposed to run
whatever
algorithm you load up on them.
There is nothing new that could be patented.
The problem is apparent on all P4s.
Also the DAZ (Denormals are Zero) and FTZ (Flush to zero) flags are available
on any processor which has a SSE2 unit implemented, e.g. all P$ and AMD64s.
Saying that, they are effective only when using the SSE(2) unit for
calculations.
You con try to use SSE
Hi Ralf,
Thanks for your reply!
Googling around, I found this on the Denormals are zero flag:
http://aulos.calarts.edu/pipermail/music-dsp/2001-May/010182.html
But it doesn't explain how these flags are set!
And of course write a small proggi that sets the FTZ and DAZ flags.
Yes, this is the part
Ralf Beck wrote:
You con try to use SSE by adding the CC options -mfpmath=sse and
-march=pentium4 when compiling.
This is for apps, correct? Not the kernel? cat /proc/cpuinfo lists my
flags as:
fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36
clflush dts acpi mmx fxsr sse
This is for apps, correct? Not the kernel? cat /proc/cpuinfo lists
my flags as:
fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36
clflush dts acpi mmx fxsr sse sse2 ss ht tm
which indicates that SSE is already in place. yes?
which indicates that your cpu supports sse
Yes, this is the part that I am unclear on. Can you explain briefly
how these flags are set? Is it during kernel-compilation or is it
something done later on? Is it a one-time option, or is it reset after
every reboot?
intel processor identification and the CPUID instruction
t
--
JACK RELEASE 0.98.16
JACK is a low-latency audio server, written primarily for the GNU/Linux
operating system. It can connect a number of different applications to
an audio device, as well as allowing them to share audio between
themselves. Its clients can run in their own processes (ie. as normal
The release number is actually 0.99.0.
Jack's website is at http://jackit.sf.net.
Taybin
From: Jens M Andreasen [EMAIL PROTECTED]
Here is an early paper:
http://www.cs.washington.edu/homes/oskin/thompson-micro2002.pdf
It is about general-purpose-computing of which audio-processing is a
subset. (BTW: Check out the aging references on the last page.)
I found nothing on the audio
From: Steve Harris [EMAIL PROTECTED]
http://plugin.org.uk/libgdither/
You have so many software there that I keep loosing the track.
Would you please list all the software in the main page so that
they can be found easily?
Also, I have succesfully downloaded older version of swh-plugins
because
The following proggi does the job:
#include xmmintrin.h
#define _MM_DENORM_ZERO_ON0x0040
main()
{
// enable flush to zero
_mm_setcsr(_MM_FLUSH_ZERO_ON | _MM_MASK_UNDERFLOW | _mm_getcsr());
// enable denormals are zero
_mm_setcsr(_MM_DENORM_ZERO_ON | _mm_getcsr());
}
Note: you will need
On Saturday 18 Sep 2004 23:17, Dmitry Baikov wrote:
Drumatic-VST also segfaults.
I use Drumatic as a DSSI plugin through the DSSI-VST bridge
(dssi-vst.so, see http://www.sf.net/projects/dssi,
http://dssi.sf.net/). It works very well for me.
I'd love to see a native DSSI drum synth, as
On sön, 2004-09-19 at 19:59, Juhana Sadeharju wrote:
From: Jens M Andreasen [EMAIL PROTECTED]
Here is an early paper:
http://www.cs.washington.edu/homes/oskin/thompson-micro2002.pdf
It is about general-purpose-computing of which audio-processing is a
subset. (BTW: Check out the aging
15 matches
Mail list logo