Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Denormalized numbers are not yer properly supported | Erik Schnetter | 2013-07-03 | 1 | -1/+1 |
| | |||||
* | Turn cpp warnings into TODO items | Erik Schnetter | 2013-07-03 | 1 | -3/+3 |
| | |||||
* | Choose FP conformance settings automatically depending on fast-math | Erik Schnetter | 2013-07-02 | 1 | -5/+22 |
| | |||||
* | Explain what VML_HAVE_FP_CONTRACT means | Erik Schnetter | 2013-06-20 | 1 | -1/+7 |
| | |||||
* | Explain why vec_builtin doesn't work | Erik Schnetter | 2013-06-17 | 1 | -0/+1 |
| | |||||
* | Introduce VML_HAVE_FP_CONTRACT (still unused) | Erik Schnetter | 2013-06-17 | 1 | -0/+1 |
| | |||||
* | Support NEON float32x4_t | Erik Schnetter | 2013-06-17 | 1 | -1/+1 |
| | |||||
* | Rename implementation files to indicate vector size | Erik Schnetter | 2013-06-17 | 1 | -13/+16 |
| | |||||
* | Correct NEON detection | Erik Schnetter | 2013-06-12 | 1 | -1/+1 |
| | |||||
* | Begin to implement ARM NEON | Erik Schnetter | 2013-06-12 | 1 | -0/+5 |
| | |||||
* | Suggest some additional vector architectures | Erik Schnetter | 2013-06-12 | 1 | -1/+4 |
| | |||||
* | Initial implementation of QPX (Blue Gene/Q) support | Erik Schnetter | 2013-06-04 | 1 | -0/+4 |
| | |||||
* | Implement VSX support | Erik Schnetter | 2013-06-02 | 1 | -6/+6 |
| | |||||
* | Put url in comment into angle brackets | Erik Schnetter | 2013-06-02 | 1 | -1/+1 |
| | |||||
* | Begin to add Altivec support (work in progress) | Erik Schnetter | 2013-04-20 | 1 | -0/+5 |
| | |||||
* | Correct handling Debian 6.0 | Erik Schnetter | 2013-03-23 | 1 | -2/+5 |
| | | | | Ensure that the tested macros are present before testing them. | ||||
* | Re-instate typeinfo work-around for Debian 6.0 | Erik Schnetter | 2013-03-22 | 1 | -0/+10 |
| | |||||
* | We handle signed zeros correctly | Erik Schnetter | 2013-03-21 | 1 | -1/+1 |
| | |||||
* | Introduce "builtin" vectors (still disabled) | Erik Schnetter | 2013-03-15 | 1 | -3/+7 |
| | |||||
* | Provide vector types with 8 and 16 bit precision | Erik Schnetter | 2013-02-19 | 1 | -9/+5 |
| | | | | Use non-functional dummy float types fp8 and fp16. | ||||
* | Clean up vector classes | Erik Schnetter | 2013-02-14 | 1 | -6/+15 |
| | | | | | | realvec<T,N>: machine-specific, optimised implementations realpseudovec<T,N>: scalarised implementation based on libm (for benchmark comparisons) realtestvec<T,N>: scalarised implementation without libm, using only vecmathlib's algorithms (for correctness testing) | ||||
* | Introduce VML_NODEBUG | Erik Schnetter | 2013-02-14 | 1 | -1/+8 |
| | |||||
* | Activate #defines handling IEEE compliance | Erik Schnetter | 2013-02-05 | 1 | -4/+4 |
| | |||||
* | Add flag VML_DEBUG | Erik Schnetter | 2013-02-04 | 1 | -4/+16 |
| | | | | Also define VML_ASSERT for error-checking | ||||
* | Add new template {bool|int|real}pseudovec that scalarizes all operations | Erik Schnetter | 2013-02-03 | 1 | -0/+2 |
| | |||||
* | Add fdim fmax fmin, fma, isfinite isinf isnan isnormal | Erik Schnetter | 2012-12-18 | 1 | -0/+5 |
| | |||||
* | Support more SSE and AVX vector types | Erik Schnetter | 2012-12-02 | 1 | -0/+10 |
| | |||||
* | Improve tests, correct errors found | Erik Schnetter | 2012-12-01 | 1 | -0/+1 |
| | |||||
* | Import initial version | Erik Schnetter | 2012-11-30 | 1 | -0/+11 |