73bf552cd6
There are two related optimizations that happen now. (These currently only apply for gathers where the mask is known to be all on, and to gathers that are accessing 32-bit sized elements, but both of these may be generalized in the future.) First, for any single gather, we are now more flexible in mapping it to individual memory operations. Previously, we would only either map it to a general gather (one scalar load per SIMD lane), or an unaligned vector load (if the program instances could be determined to be accessing a sequential set of locations in memory.) Now, we are able to break gathers into scalar, 2-wide (i.e. 64-bit), 4-wide, or 8-wide loads. Further, we now generate code that shuffles these loads around. Doing fewer, larger loads in this manner, when possible, can be more efficient. Second, we can coalesce memory accesses across multiple gathers. If we have a series of gathers without any memory writes in the middle, then we try to analyze their reads collectively and choose an efficient set of loads for them. Not only does this help if different gathers reuse values from the same location in memory, but it's specifically helpful when data with AOS layout is being accessed; in this case, we're often able to generate wide vector loads and appropriate shuffles automatically.
==============================
Intel(r) SPMD Program Compiler
==============================
``ispc`` is a compiler for a variant of the C programming language, with
extensions for `single program, multiple data
<http://en.wikipedia.org/wiki/SPMD>`_ programming. Under the SPMD model,
the programmer writes a program that generally appears to be a regular
serial program, though the execution model is actually that a number of
*program instances* execute in parallel on the hardware.
Overview
--------
``ispc`` compiles a C-based SPMD programming language to run on the SIMD
units of CPUs; it frequently provides a 3x or more speedup on CPUs with
4-wide vector SSE units and 5x-6x on CPUs with 8-wide AVX vector units,
without any of the difficulty of writing intrinsics code. Parallelization
across multiple cores is also supported by ``ispc``, making it
possible to write programs that achieve performance improvement that scales
by both number of cores and vector unit size.
There are a few key principles in the design of ``ispc``:
* To build a small set of extensions to the C language that
would deliver excellent performance to performance-oriented
programmers who want to run SPMD programs on the CPU.
* To provide a thin abstraction layer between the programmer
and the hardware--in particular, to have an execution and
data model where the programmer can cleanly reason about the
mapping of their source program to compiled assembly language
and the underlying hardware.
* To make it possible to harness the computational power of SIMD
vector units without the extremely low-programmer-productivity
activity of directly writing intrinsics.
* To explore opportunities from close coupling between C/C++
application code and SPMD ``ispc`` code running on the
same processor--to have lightweight function calls between
the two languages and to share data directly via pointers without
copying or reformatting.
``ispc`` is an open source compiler with the BSD license. It uses the
remarkable `LLVM Compiler Infrastructure <http://llvm.org>`_ for back-end
code generation and optimization and is `hosted on
github <http://github.com/ispc/ispc/>`_. It supports Windows, Mac, and
Linux, with both x86 and x86-64 targets. It currently supports the SSE2,
SSE4, AVX1, and AVX2 instruction sets.
Features
--------
``ispc`` provides a number of key features to developers:
* Familiarity as an extension of the C programming
language: ``ispc`` supports familiar C syntax and
programming idioms, while adding the ability to write SPMD
programs.
* High-quality SIMD code generation: the performance
of code generated by ``ispc`` is often close to that of
hand-written intrinsics code.
* Ease of adoption with existing software
systems: functions written in ``ispc`` directly
interoperate with application functions written in C/C++ and
with application data structures.
* Portability across over a decade of CPU
generations: ``ispc`` has targets for SSE2, SSE4, AVX
(and soon, AVX2).
* Portability across operating systems: Microsoft
Windows, Mac OS X, and Linux are all supported
by ``ispc``.
* Debugging with standard tools: ``ispc``
programs can be debugged with standard debuggers (OS X and
Linux only).
Additional Resources
--------------------
Prebuilt ``ispc`` binaries for Windows, OS X and Linux can be downloaded
from the `ispc downloads page <http://ispc.github.com/downloads.html>`_.
See also additional
`documentation <http://ispc.github.com/documentation.html>`_ and additional
`performance information <http://ispc.github.com/perf.html>`_.