similar to: [LLVMdev] Anyone is building a DSP-C frontend?

On Wed, 31 Aug 2005, Tzu-Chien Chiu wrote: > fixed-point number could be stored in LLVM first class integer types. > i cannot see the problem now. but to be type-safe, there should be a > first class 'fixed'. There is no need. Lowering is fine, in the same way that enums or typedefs are currently lowered to llvm integer types. > some llvm extensions required to mapping

Each register is a 4-component (namely, r, g, b, a) vector register. They are actually defined as llvm packed [4xfloat]. The instruction: add_sat r0.a, r1_bias.xxyy, r3_x2.zzzz Explaination: '.a' is a writemask. only the specified component will be update '.xxyy' and '.zzzz' are swizzle masks, specify the component permutation, simliar to the Intel SSE permutation

Actually the problems that Tzu-Chien Chiu are encountering are similar to what should be done for generating SSE code in the X86 backend and also other SIMD instruction sets. I think LLVM neeeds to add instructions for permuting components, extracting and injecting elements in packed types. If the architecture has instructions which can do permutations for each instruction (for example

It's not clear what is actually wrong from your original message, I think you need to give some more information as to what you are doing: Example source, what target GPU, compiler error messages or other evidence of "it's wrong" (llvm IR, disassembly, etc) ... -- Mats On 8 April 2016 at 09:55, Liu Xin via llvm-dev <llvm-dev at lists.llvm.org> wrote: > I built it

On Thu, 31 Aug 2006, Kenneth Hoste wrote: > Bummer. I think I'll contact the NAG support for more info on this. Can you > show me the content of your Makefile.nagfortran? It is identical to yours. > Also, it is possible to tell make only to compile benchmark X? How can I > enforce this? Go into the directory for that benchmark, then run 'make' or whatever. -Chris --

Hi, guys, I am comparing the performance of the default scheduler (seems to be the one that minimizes register pressure) with no scheduler (-pre-RA-sched=none), and I got these numbers. The ratio is low_reg_pressure/none, that is, the lower the number, the better the performance with low register pressure: CFP2000/177.mesa/177.mesa 1.00 CFP2000/179.art/179.art

On Wed, 30 Aug 2006, Kenneth Hoste wrote: > I've been trying (on and off) to compile the _full_ SPEC CPU2000 benchmark > suite to LLVM bytecode. The biggest problem > I'm facing is the Fortran benchmarks, for which some partial support is > already available it seems (using f2c). ok. > Unfortunately the f2c compiler only allows the translation of Fortran-77 >

On 31 Aug 2006, at 19:13, Chris Lattner wrote: > On Thu, 31 Aug 2006, Kenneth Hoste wrote: > >> When I adjust the settings in Makefile.nagfortran as follows, I'm >> able to get bytecode file for lucas, galgel and facerec, but make >> still quits with an error (after generating >> bytecode files for 7 (out of 26) benchmarks. >> >> Also, the file

Hello LLVM-people, I've been trying (on and off) to compile the _full_ SPEC CPU2000 benchmark suite to LLVM bytecode. The biggest problem I'm facing is the Fortran benchmarks, for which some partial support is already available it seems (using f2c). Unfortunately the f2c compiler only allows the translation of Fortran-77 programs to C code (which is then compiled using llvm-gcc).

Anyone know what to do about this: make[4]: Leaving directory `/proj/llvm/llvm-test-1/External/SPEC/CFP2000/173.applu' make[4]: Entering directory `/proj/llvm/llvm-test-1/External/SPEC/CFP2000/178.galgel' /usr/bin/f95 -w -S -O2 /opt/spec/CPU2000v1.3.1/benchspec//CFP2000/178.galgel/src/modules.f90 -o modules.c -fixed -kind=byte -dcfuns -dusty f95: unrecognized option '-kind=byte'

On 31 Aug 2006, at 23:46, Chris Lattner wrote: > On Thu, 31 Aug 2006, Kenneth Hoste wrote: >> Bummer. I think I'll contact the NAG support for more info on >> this. Can you >> show me the content of your Makefile.nagfortran? > > It is identical to yours. > >> Also, it is possible to tell make only to compile benchmark X? How >> can I >>

Hello folks, The makefile for 177.mesa says that for a small problem size, it will get 100 frames. But in the spec sources I have, the test folder only contains numbers for 10 frames: $ speccpu2000/benchspec/CFP2000/177.mesa/data $ wc -l test/input/numbers 10 test/input/numbers Generating 100 frames causes undefined behaviour because the program is doing unchecked fscanf on that

I have a doubt. This is an excerpt of the raw report I get after running Spec benchmarks through llvm-test.I am trying to calculate the program execution time.Does the output result in bold corresponds to "lli time" in Makefile.spec ? I am not interested in llc, jit or cbe.I simply need the normal bytecode and native code execution times after running my pass over them.I have modified

On Dec 23, 2009, at 6:26 PM, Julien Lerouge wrote: > Hello folks, > > The makefile for 177.mesa says that for a small problem size, it will > get 100 frames. But in the spec sources I have, the test folder only > contains numbers for 10 frames: > > $ speccpu2000/benchspec/CFP2000/177.mesa/data $ wc -l test/input/ > numbers > 10 test/input/numbers > >

Hi Chris, > > I use NAG with llvm-gcc4. What sort of errors do you get? Did you > configure llvm-test with the appropriate flags to find it? > Yes, I did. llvm-test is configured as follows: (in /work/LLVM/1.8/ llvm/project/llvm-test): ./configure --with-spec2000=/work/SPEC_CPU2000_1.3_src/benchspec -- without-f2c --with-f95-bin=/work/NAG_f95/bin --with-f95-lib=/work/

Hello, I have been trying to compile spec benchmark with llvm.I am trying the 2 methods suggested but have not succeeded yet . 1) Setting the flags -Wl,-native-cbe or -Wl,-native flags to llvm-gcc. This results in successful compilation with llvm BUT it doesn't create .bc files (?)which I need for my experimentation.Any other significant modification/addition missing ?? 2) Using llvm-test :

Hi, DSP-C is an extension to ISO C to support specific hardware features in DSP, e.g. fixed-point typer, memory spaces and circular pointers. It's generic for different DSP's. - http://www.dsp-c.org How difficult is it to extending the llvm-gcc for this extension? Any document like "hitchhiker's guide to extend the GCC fronted" ? (just like the guide for adding

On 01 Sep 2006, at 10:05, Kenneth Hoste wrote: > >> >>> Also, it is possible to tell make only to compile benchmark X? How >>> can I >>> enforce this? >> >> Go into the directory for that benchmark, then run 'make' or >> whatever. > I tried tom compile each of the SPEC CPU2000 benchmarks using the make command is each respective

Hi, Below you can see the updated benchmark results for the new SLP-vectorizer. As you can see, there is a small number of compile time regressions, a single major runtime *regression, and many performance gains. There is a tiny increase in code size: 30k for the whole test-suite. Based on the numbers below I would like to enable the SLP-vectorizer by default for -O3. Please let me know if you

Hi, LLVM’s SLP-vectorizer is a new pass that combines similar independent instructions in a straight-line code. It is currently not enabled by default, and people who want to experiment with it can use the clang command line flag “-fslp-vectorize”. I ran LLVM’s test suite with and without the SLP vectorizer on a Sandybridge mac (using SSE4, w/o AVX). Based on my performance measurements