Releases: krrishnarraj/clpeak
Releases · krrishnarraj/clpeak
1.1.3
What's Changed
- add a few more parameters by @isowson in #93
- Add 24-bit integer compute to RX 5500XT results in #88
- [cmake] use GNUInstallDirs to get canonical paths by @tgurr in #57
- Add log files for additional devices by @nchristensen in #95
- Fix compute integer unit reported in xml files by @rjodinchr in #114
- add profiling for char & short integer types by @rjodinchr in #115
- Adding results for Tesla P40 by @KatyushaScarlet in #119
- Adding results for MTT S80 by @KatyushaScarlet in #120
- Use stricter host buffer alignment (64B) required by modern CPUs. by @pioto1225 in #121
New Contributors
- @isowson made their first contribution in #93
- @tgurr made their first contribution in #57
- @nchristensen made their first contribution in #95
- @rjodinchr made their first contribution in #114
- @KatyushaScarlet made their first contribution in #119
- @pioto1225 made their first contribution in #121
Full Changelog: 1.1.2...1.1.3
Contributors
tgurr, nchristensen, and 4 other contributors
Assets 2
Release 1.1.2
- Snap related fixes
Full Changelog: 1.1.1...1.1.2
Release 1.1.1
What's Changed
- Add non-blocking read/write transfers by @infromantyk in #68
- Add RX 5500XT results by @atomsymbol in #69
- Added 24bit fast MAD Integer Compute test by @rigred in #70
- Updated results with various Intel, AMD and NVidia hardware by @jjkeijser in #75
- Create m1.log by @xchern in #77
- Add lib64 EGL libGLES_mali.so by @paranlee in #80
- Apple M1 benchmarking results w/ CLVK by @woachk in #82
- Fix icd build issue by @krrishnarraj in #87
New Contributors
- @infromantyk made their first contribution in #68
- @rigred made their first contribution in #70
- @jjkeijser made their first contribution in #75
- @xchern made their first contribution in #77
- @paranlee made their first contribution in #80
- @woachk made their first contribution in #82
Full Changelog: 1.1.0...1.1.1
Contributors
krrishnarraj, rigred, and 5 other contributors
Assets 2
Release 1.1.0
- All the kernels are now de-unrolled
Prior to this, all mad or fetch operations were manually unrolled which created register pressure on low-end devices. Now the operations are put in a for loop and left for the compiler to unroll - General improvements
1.1.0 rc2
1.1.0-rc2 Bump rc
1.1.0 rc1
1.1.0-rc1 [src] auto format cpp files
Release for OpenCL 1.2
Merge pull request #16 from ekondis/master Added results for AMD R7-260X GPU.