search for: hardware_

...st not be optimized as integers - eg if two saturated fixed point constants would overflow in an addition operation, the result should as well be saturated. Doing this in a series of steps with intrinsics would be quite ridiculous as far as performance goes, because this is _supported by the DSP in hardware_ in a single instruction :-) What about perhaps introducing a new type, and allowing just minor extensions, while barring it from more heavy scattered code? I mean, the GVN pass works by just adding an enum entry and a switch case for each new instruction. But in other cases, there are more LOCs i...

...integers - eg if two saturated fixed point constants would overflow in > an addition operation, the result should as > well be saturated. Doing this in a series of steps with intrinsics would > be quite ridiculous as far as performance goes, > because this is _supported by the DSP in hardware_ in a single instruction > :-) Isn't it better to add instructions for saturated and/or fixed point arithmetic? ... At least that would be consistent with the design of having separate instructions for signed and unsigned add etc. - Morten

...> > two saturated fixed point constants would overflow in an addition > > operation, the result should as well be saturated. Doing this in a > > series of steps with intrinsics would be quite ridiculous as far as > > performance goes, because this is _supported by the DSP in hardware_ > > in a single instruction :-) > > > > What about perhaps introducing a new type, and allowing just minor > > extensions, while barring it from more heavy scattered code? I mean, > > the GVN pass works by just adding an enum entry and a switch case > > for eac...

...as integers - eg if > two saturated fixed point constants would overflow in an addition > operation, the result should as well be saturated. Doing this in a > series of steps with intrinsics would be quite ridiculous as far as > performance goes, because this is _supported by the DSP in hardware_ > in a single instruction :-) > > What about perhaps introducing a new type, and allowing just minor > extensions, while barring it from more heavy scattered code? I mean, > the GVN pass works by just adding an enum entry and a switch case for > each new instruction. But in othe...

...two saturated fixed point constants would overflow in an addition >>> operation, the result should as well be saturated. Doing this in a >>> series of steps with intrinsics would be quite ridiculous as far as >>> performance goes, because this is _supported by the DSP in hardware_ >>> in a single instruction :-) >>> >>> What about perhaps introducing a new type, and allowing just minor >>> extensions, while barring it from more heavy scattered code? I mean, >>> the GVN pass works by just adding an enum entry and a switch case &g...

...> patches on it. > > Yes, this sounds like the right direction. The targets obviously need to > provide this information. > I'd like to help review this as it'll be necessary to implement http://wg21.link/p0154r1 which is (likely) in C++17. It adds two values, constexpr std::hardware_{constructive,destructive}_interference_size, because in some configurations you don't have a precise cacheline size but rather a range based on what multiple architectures have implemented for the same ISA. I think you'll want something similar. -------------- next part -------------- An HT...

...fixed point constants would overflow in an addition > >>> operation, the result should as well be saturated. Doing this in a > >>> series of steps with intrinsics would be quite ridiculous as far as > >>> performance goes, because this is _supported by the DSP in hardware_ > >>> in a single instruction :-) > >>> > >>> What about perhaps introducing a new type, and allowing just minor > >>> extensions, while barring it from more heavy scattered code? I mean, > >>> the GVN pass works by just adding an enum e...

Hi, This is a proposal about implementing an analysis that calculates loop cost based on cache data. The primary motivation for implementing this is to write profitability measures for cache related optimizations like interchange, fusion, fission, pre-fetching and others. I have implemented a prototypical version at http://reviews.llvm.org/D21124. The patch basically creates groups of references