similar to: Randomly selecting rows with unique values in two columns

The following IR implements the following nested loop: for (int i = start ; i < end ; ++i ) for (int p = 0 ; p < 4 ; ++p ) a[i*4+p] = b[i*4+p] + c[i*4+p]; define void @main(i64 %arg0, i64 %arg1, i1 %arg2, i64 %arg3, float* noalias %arg4, float* noalias %arg5, float* noalias %arg6) { entrypoint: br i1 %arg2, label %L0, label %L1 L0:

The loop vectorizer relies on cleanup passes to be run after it: from Transforms/IPO/PassManagerBuilder.cpp: // Add the various vectorization passes and relevant cleanup passes for // them since we are no longer in the middle of the main scalar pipeline. MPM.add(createLoopVectorizePass(DisableUnrollLoops)); MPM.add(createInstructionCombiningPass());

Hi, How does the requirement that phi instructions have one value per predecessor basic block interact with indirectbr instructions? For instance, take the following code: L1: br i1 %somevalue, label %L2, label %L3 L2: %ret1 = i8* blockaddress(@myfunction, %L5) br label %L4 L3: %ret2 = i8* blockaddress(@myfunction, %L6) br label %L4 L4: %ret = phi i8* [%ret1, L2], [%ret2, L3]

The instcombine pass cleans up a lot. Any idea why there are still shufflevector, insertelement, *and* bitcast (!!) etc. instructions left? The original loop is so clean, a textbook example I'd say. There is no need to shuffle anything.At least I don't see it. Frank vector.ph: ; preds = %L5 %broadcast.splatinsert1 = insertelement <4 x

Hello, I have improved quality for -l4 and -l5 high compression settings.These settings are now really better and very competitive, I now prefer them compared to x265 (HEVC). Update at: http://nhwcodec.blogspot.com/ I am also working on -l6 quality setting, I can simply increase quantization for it but I am not totally satisfied with this solution for now... Looking for more advanced processing

I am trying a setup where the one loop is rewritten as two loops. This avoids the 'rem' and 'div' instructions in the index calculation (which give the loop vectorizer a hard time). However, with this setup the loop vectorizer complains about a too small loop. LV: Checking a loop in "main" LV: Found a loop: L3 LV: Found a loop with a very small trip count. This loop

Hi David! Many thanks for your very encouraging and kind feedback!!! (Hope you don't mind that I made your post public on the Theora channel). I try to be the more sincere as I can in my posts.Some people suggested me to make demo pages to explain how the codec works step by step, and I realize that it would be good, because certainly it will show that the NHW codec is not that complex and

For what it's worth, I'm also experiencing this same issue. If there is interest I can provide some very simple reproducible test cases, but I was planning on moving to MCJIT this week anyway. -- View this message in context: http://llvm.1065342.n5.nabble.com/loop-vectorizer-JIT-AVX-segfaults-tp63089p63115.html Sent from the LLVM - Dev mailing list archive at Nabble.com.

In the case when coming from C it was probably the loop unroller and SLP vectorizer which vectorized the code. Potentially I could do the same in the IR. However, the loop body that is generated in the IR can get very large. Thus, the loop unroller will refuse to unroll the loop in a large number of (important) cases. Isn't there a way to convince the loop vectorizer that it should

On 11/12/2009, at 10:43 AM, Anton Korobeynikov wrote: > Hi, Chris > >> That is target independent code, so you should not put sparc specific changes there. It sounds like one of the sparc-specific target hooks is wrong. > Since sparc does not provide any hooks for operation of branches (e.g. > AnalyzeBranch and friends) it might be possible that generic codegen > code is

Hi, Chris > That is target independent code, so you should not put sparc specific changes there.  It sounds like one of the sparc-specific target hooks is wrong. Since sparc does not provide any hooks for operation of branches (e.g. AnalyzeBranch and friends) it might be possible that generic codegen code is broken in absence of these hooks. -- With best regards, Anton Korobeynikov Faculty

Hello Dave, Thank you very much for your support and your interest. Actually so few people took a look at the NHW Project, but with this version the NHW Project is becoming very interesting, it is royalty-free and very fast furthermore. Really if you could find time, I would be very interested in any opinion from the Xiph.org community! Cheers, Raphael 2018-04-28 21:40 GMT+02:00 Dave

On Feb 8, 2010, at 12:37 AM, Nathan Keynes wrote: > Firstly, the BNE/BA pair should be reduced to a BE (I assume this is > the responsibility of AnalyzeBranch and friends that you mention). Right. Implementing AnalyzeBranch will allow a bunch of block layout and branch optimizations to happen. > However I still wouldn't have expected that to result in the label > being

Hi to all, I am trying to run breakpoints() on a fairly large sample (>10.000 observations). The process is very slow, any idea on how to speed this up? I have tried the hpc="foreach" parameter, but this didn't work at all when I tried to run it on a smaller sample. breakpoints(x ~ x.l1 + x.l2 + X.l3 + x.l4 + x.l5 + x.l6 + x.l7 + x.l8 + y.l1 + y.l2 + y.l3 + y.l4 + y.l5 + y.l6

Thank you so much Sir!!! If you have time, do not hesitate to let us know what you would think of the NHW Project.Any remark, opinion would be very welcome! Many thanks again! Best regards, Raphael 2018-04-28 23:02 GMT+02:00 Dave Johnson Games <davefilms.us at gmail.com>: > Cool, I'll check it out. > > On Sat, Apr 28, 2018, 3:09 PM Raphael Canut <nhwcodec at gmail.com>

Hello All, I have a discontinuous dataset and I used spplot to plot the points. I was wondering whether it is possible to show the legend of the plot as color ramp. I saw that on continuous data. If anyone could tell me that would be great. Another thing I would like to know is how can we create the horizontal legend. By default, the legend is vertical. I was able to shift the legend inside

On 09/02/2010, at 3:57 AM, Chris Lattner wrote: > On Feb 8, 2010, at 12:37 AM, Nathan Keynes wrote: >> Firstly, the BNE/BA pair should be reduced to a BE (I assume this is the responsibility of AnalyzeBranch and friends that you mention). > > Right. Implementing AnalyzeBranch will allow a bunch of block layout and branch optimizations to happen. > >> However I still

The loop vectorizer is doing an amazing job so far. Most of the time. I just came across one function which led to unexpected behavior: On this function the loop vectorizer finds a 256 bit vector as the wides vector type for the x86-64 architecture. (!) This is strange, as it was always finding the correct size of 128 bit as the widest type. I isolated the IR of the function to check if this is

Hi, I’m writing this email behalf of my manager in Alibaba. Our team is working on a deep learning engine called MNN. We open sourced it this May. (Github link here: https://github.com/alibaba/MNN ). It has industry-leading blazing fast inference speed on mobile and it is tiny in size. In the internal experiments we ran, MNN outperforms other inference engines under almost all settings by

hi all, I have a plot to which i would like to add text labels. And i cant find a way...here is the code : enaD2<-idw(D2~1, loca=dva, newdata=grd) pts = list("sp.points", dva, pch = 20, cex=1.5, col = "darkred spplot(enaD2, "var1.pred",sp.layout=pts, main = "globina 60 cm", sub="D2",