On Thu, Dec 17, 2009 at 1:59 PM, Gabriel Michael Black
gbl...@eecs.umich.edu wrote:
Yeah, picking the microops was a big concern for me when I was
starting out. If you remember, I closely based what we have now on
that patent I found for what looks like an older, 32 bit version of
AMD's
On Wed, Dec 16, 2009 at 4:33 PM, Gabriel Michael Black
gbl...@eecs.umich.edu wrote:
Quoting Vince Weaver vi...@csl.cornell.edu:
On Wed, 16 Dec 2009, Steve Reinhardt wrote:
On Sun, Dec 13, 2009 at 8:57 PM, Vince Weaver vi...@csl.cornell.edu wrote:
I did finish running and verifying spec2k on
I've been thinking about this since reading your email, and it occurs
to me the microops may be loads, ops, stores, or opstores and still
roughly fall into a RISC style architecture. Stores have to wait
around in the store queue anyway, so they could wait for their data to
be generated by
On Sun, Dec 13, 2009 at 8:57 PM, Vince Weaver vi...@csl.cornell.edu wrote:
I did finish running and verifying spec2k on x86_64 (it took longer than
it should have due to an unfortunate power-outage on our cluster). The
benchmarks all finished, and the retired instruction count matches actual
On Wed, 16 Dec 2009, Steve Reinhardt wrote:
On Sun, Dec 13, 2009 at 8:57 PM, Vince Weaver vi...@csl.cornell.edu wrote:
I did finish running and verifying spec2k on x86_64 (it took longer than
it should have due to an unfortunate power-outage on our cluster). The
benchmarks all finished,
Quoting Vince Weaver vi...@csl.cornell.edu:
On Wed, 16 Dec 2009, Steve Reinhardt wrote:
On Sun, Dec 13, 2009 at 8:57 PM, Vince Weaver vi...@csl.cornell.edu wrote:
I did finish running and verifying spec2k on x86_64 (it took longer than
it should have due to an unfortunate power-outage on
Looking through my email, I found this patch which I don't think ever
got committed. Do you know of any others like this, Vince? I found this
one specifically because I'm getting ready to commit my changes that
cleans up the ext flags, and this would need to be adjusted slightly.
The change is
On Sun, 13 Dec 2009, Gabe Black wrote:
Looking through my email, I found this patch which I don't think ever
got committed. Do you know of any others like this, Vince? I found this
one specifically because I'm getting ready to commit my changes that
cleans up the ext flags, and this would
On Fri, 6 Nov 2009, Gabe Black wrote:
You seem to be missing the majority of the PC relative version... Would
you like to fix that up, or should I?
The blow version properly has the PC relative version, which has been
tested.
Vince
# HG changeset patch
# User Vince Weaver
Vince Weaver wrote:
On Fri, 6 Nov 2009, Gabe Black wrote:
You seem to be missing the majority of the PC relative version... Would
you like to fix that up, or should I?
I'll fix it, I can't believe I missed that. I need to make sure my tests
handle PC relative.
when I fix
I debated adding an hadd microop or adding a flag that changed the
behavior of maddf, but in the end I didn't do either since I didn't have
a ready way to test any implementation of hadd. Between those two I'd
probably go with the hadd microop since maddf might end up overly
complicated and
You seem to be missing the majority of the PC relative version... Would
you like to fix that up, or should I?
Gabe
Gabe Black wrote:
I debated adding an hadd microop or adding a flag that changed the
behavior of maddf, but in the end I didn't do either since I didn't have
a ready way to
The patch below adds haddps support.
It is quite complicated, I'm not sure if there is a better way to access
32-bit chunks of the xmm registers.
attached is a small test program that tests the instruction.
Vince
# HG changeset patch
# User Vince Weaver vi...@csl.cornell.edu
# Date
13 matches
Mail list logo