Jérôme Richard
Jérôme Richard
Hello, I just tried the plugin but it does not work with a relatively recent version of sublime text on my machine. The sublime text console show the following error:...
Hello, There is a build issue in the file `source/libhoard.cpp:119`: ``` source\libhoard.cpp(119): error C2065: 'flatten': undeclared identifier source\libhoard.cpp(119): error C2146: syntax error: missing ';' before identifier 'xxmalloc' source\libhoard.cpp(119): error C3646:...
Hello, I benchmarked the simple following dgemm call using 4096x4096 matrices (thus n=4096 and a, b and c are matrices) on a IBM LC922 machine with 2 POWER-9 processors (of...
The library hangs on on the parsing of one simple integer. Here is the code to reproduce the issue: ```fortran subroutine test_parsing() type(fson_value), pointer :: jsonValue jsonValue => fson_parse(str='42') !...
The library accept invalid json strings. Here is the code to reproduce the issue: ```fortran subroutine test_parsing() type(fson_value), pointer :: jsonValue jsonValue => fson_parse(str=' [ - 4 y 2 ]...
Hello, This PR strongly improves the relative precision of `asin` close to 0. As mentioned in #2236, the error was previously huge: > `asin` appears not to support subnormal numbers...
Hello, There is apparently a bug happening in the tests of the PR #2958 causing ISPC to suddenly crash with the following message: > /home/richard/Bureau/projects/ispc/src/module.cpp(1478): FATAL ERROR: Resulting module verification...
Hello, This PR add a fast and precise `cbrt` function for floats (there is also a double-precision implementation calling the standard library like for other functions). This fix the issue...
Hello, This PR fix few issues in the 32-bit implementation of `exp` and `log` (for the default ISPC math mode). Most of the issues have been mentioned [here](https://github.com/ispc/ispc/issues/2236#issuecomment-2267633771). More specifically,...
Hello, I found out that the code generated by ISPC is clearly sub-optimal when a (`varying`) condition scan variables that are of type `bool` as opposed to `uint8`. It seems...