Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/9/07, Stefan van der Walt <[EMAIL PROTECTED]> wrote: > > It doesn't really matter where the memory allocation occurs, does it? > As far as I understand, the underlying fftw function has some flag to > indicate when the data is aligned. If so, we could expose that flag > in Python, and do som

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/9/07, Charles R Harris <[EMAIL PROTECTED]> wrote: > > > On 8/9/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Charles R Harris wrote: > > > > > > Ah, you want it in C. > > What would be the use to get SIMD aligned arrays in python ? > > If I wanted a fairly specialized routine and didn't

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On Thu, Aug 09, 2007 at 04:52:38PM +0900, David Cournapeau wrote: > Charles R Harris wrote: > > > > Well, what you want might be very easy to do in python, we just need > > to check the default alignments for doubles and floats for some of the > > other compilers, architectures, and OS's out ther

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/9/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Charles R Harris wrote: > > > > Ah, you want it in C. > What would be the use to get SIMD aligned arrays in python ? If I wanted a fairly specialized routine and didn't want to touch the guts of numpy, I would pass the aligned array to a

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Charles R Harris wrote: > > Ah, you want it in C. What would be the use to get SIMD aligned arrays in python ? David ___ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/9/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Charles R Harris wrote: > > > > Well, what you want might be very easy to do in python, we just need > > to check the default alignments for doubles and floats for some of the > > other compilers, architectures, and OS's out there. On the o

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/9/07, Charles R Harris <[EMAIL PROTECTED]> wrote: > > > > On 8/8/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > > > Charles R Harris wrote: > > > Anne, > > > > > > On 8/8/07, *Anne Archibald* <[EMAIL PROTECTED] > > > > wrote: > > > > > > On 08/08/2007, Charles R Harris <[EMAIL PROTECT

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Charles R Harris wrote: > > Well, what you want might be very easy to do in python, we just need > to check the default alignments for doubles and floats for some of the > other compilers, architectures, and OS's out there. On the other hand, > you might not be able to request a c malloc that is

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/8/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Charles R Harris wrote: > > Anne, > > > > On 8/8/07, *Anne Archibald* <[EMAIL PROTECTED] > > > wrote: > > > > On 08/08/2007, Charles R Harris <[EMAIL PROTECTED] > > > wrote: > >

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Charles R Harris wrote: > Anne, > > On 8/8/07, *Anne Archibald* <[EMAIL PROTECTED] > > wrote: > > On 08/08/2007, Charles R Harris <[EMAIL PROTECTED] > > wrote: > > > > > > On 8/8/07, Anne Archibald <[EMAIL PROTECTED] >

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/8/07, Matthieu Brucher <[EMAIL PROTECTED]> wrote: > > My 64 bit linux on Intel aligns arrays, whatever the data type, on 16 byte > > boundaries. It might be interesting to see what happens with the Intel and > > MSVC comipilers, but I expect similar results. > > > > According to the doc on the

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

> > My 64 bit linux on Intel aligns arrays, whatever the data type, on 16 byte > boundaries. It might be interesting to see what happens with the Intel and > MSVC comipilers, but I expect similar results. > According to the doc on the msdn, the data should be 16-bits aligned. Matthieu ___

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Anne, On 8/8/07, Anne Archibald <[EMAIL PROTECTED]> wrote: > > On 08/08/2007, Charles R Harris <[EMAIL PROTECTED]> wrote: > > > > > > On 8/8/07, Anne Archibald <[EMAIL PROTECTED]> wrote: > > > Oh. Well, it's not *terrible*; it gets you an aligned array. But you > > > have to allocate the original

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 08/08/2007, Charles R Harris <[EMAIL PROTECTED]> wrote: > > > On 8/8/07, Anne Archibald <[EMAIL PROTECTED]> wrote: > > Oh. Well, it's not *terrible*; it gets you an aligned array. But you > > have to allocate the original array as a 1D byte array (to allow for > > arbitrary realignments) and the

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/8/07, Anne Archibald <[EMAIL PROTECTED]> wrote: > > On 08/08/2007, Stefan van der Walt <[EMAIL PROTECTED]> wrote: > > On Tue, Aug 07, 2007 at 01:33:24AM -0400, Anne Archibald wrote: > > > Well, it can be done in Python: just allocate a too-big ndarray and > > > take a slice that's the right sh

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 08/08/2007, Stefan van der Walt <[EMAIL PROTECTED]> wrote: > On Tue, Aug 07, 2007 at 01:33:24AM -0400, Anne Archibald wrote: > > Well, it can be done in Python: just allocate a too-big ndarray and > > take a slice that's the right shape and has the right alignment. But > > this sucks. > > Could

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On Tue, Aug 07, 2007 at 01:33:24AM -0400, Anne Archibald wrote: > Well, it can be done in Python: just allocate a too-big ndarray and > take a slice that's the right shape and has the right alignment. But > this sucks. Could you explain to me why is this such a bad idea? Stéfan __

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/6/07, Anne Archibald <[EMAIL PROTECTED]> wrote: > > On 06/08/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > > Well, when I proposed the SIMD extension, I was willing to implement the > > proposal, and this was for a simple goal: enabling better integration > > with many numeric libraries

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Anne Archibald wrote: > Well, it can be done in Python: just allocate a too-big ndarray and > take a slice that's the right shape and has the right alignment. But > this sucks. Stephen G. Johnson posted code earlier in this thread that > provides a portable aligned-memory allocator - it handles the

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

> > For platforms without posix_memalign, I don't see how to > implement a memory allocator with an arbitrary alignment (more > precisely, I don't see how to free it if I cannot assume a fixed > alignement: how do I know where the "real" pointer is ?). Visual Studio seems to offer a counter part

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 07/08/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > Anne, you said previously that it was easy to allocate buffers for a > given alignment at runtime. Could you point me to a document which > explains how ? For platforms without posix_memalign, I don't see how to > implement a memory alloca

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Anne Archibald wrote: > > I have to agree. I can hardly volunteer David for anything, and I > don't have time to implement this myself, but I think a custom > allocator is a rather special-purpose tool; if one were to implement > one, I think the way to go would be to implement a subclass of ndarra

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 06/08/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > Well, when I proposed the SIMD extension, I was willing to implement the > proposal, and this was for a simple goal: enabling better integration > with many numeric libraries which need SIMD alignment. > > As nice as a custom allocator mig

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Lisandro Dalcin wrote: > On 8/3/07, David Cournapeau <[EMAIL PROTECTED]> wrote: >> Here is what I can think of: >> - adding an API to know whether a given PyArrayObject has its data >> buffer 16 bytes aligned, and requesting a 16 bytes aligned >> PyArrayObject. Something like NPY_ALIGNED, ba

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/3/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > Here is what I can think of: > - adding an API to know whether a given PyArrayObject has its data > buffer 16 bytes aligned, and requesting a 16 bytes aligned > PyArrayObject. Something like NPY_ALIGNED, basically. > - forcing da

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On Aug 4, 3:24 am, "Anne Archibald" <[EMAIL PROTECTED]> wrote: > It seems to me two things are needed: > > * A mechanism for requesting numpy arrays with buffers aligned to an > arbitrary power-of-two size (basically just using posix_memalign or > some horrible hack on platforms that don't have it

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 04/08/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Here's a hack that google turned up: I'd avoid hacks in favour of posix_memalign (which allows arbitrary degrees of alignment. For one thing, freeing becomes a headache (you can't free a pointer you've jiggered!). > - Check whethe

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

> > > Here's a hack that google turned up: > > (1) Use static variables instead of dynamic (stack) variables > (2) Use in-line assembly code that explicitly aligns data > (3) In C code, use "*malloc*" to explicitly allocate variables > > Here is Intel's example of (2): > >

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/3/07, Charles R Harris <[EMAIL PROTECTED]> wrote: > > > > On 8/3/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > > > Andrew Straw wrote: > > > Dear David, > > > > > > Both ideas, particularly the 2nd, would be excellent additions to > > numpy. > > > I often use the Intel IPP (Integrated Pe

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

On 8/3/07, David Cournapeau <[EMAIL PROTECTED]> wrote: > > Andrew Straw wrote: > > Dear David, > > > > Both ideas, particularly the 2nd, would be excellent additions to numpy. > > I often use the Intel IPP (Integrated Performance Primitives) Library > > together with numpy, but I have to do all my

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Andrew Straw wrote: > Dear David, > > Both ideas, particularly the 2nd, would be excellent additions to numpy. > I often use the Intel IPP (Integrated Performance Primitives) Library > together with numpy, but I have to do all my memory allocation with the > IPP to ensure fastest operation. I th

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Dear David, Both ideas, particularly the 2nd, would be excellent additions to numpy. I often use the Intel IPP (Integrated Performance Primitives) Library together with numpy, but I have to do all my memory allocation with the IPP to ensure fastest operation. I then create numpy views of the da

Re: [Numpy-discussion] numpy arrays, data allocation and SIMD alignement

Hi, >Following an ongoing discussion with S. Johnson, one of the developer > of fftw3, I would be interested in what people think about adding > infrastructure in numpy related to SIMD alignement (that is 16 bytes > alignement for SSE/ALTIVEC, I don't know anything about other archs). > The pr