Re:Russia’s Elbrus 8CB Microarchitecture: 8-Core VLIW on TSM...
by AlEinstein on Tuesday June 02, 2020 @08:32AM
(#23390260)
The floating point results are suspiciously good compared to the rest of the results. I wonder if they have some good VLIW instructions for floating point or it’s just an error in the spreadsheet.
[ Parent
| on HN ]
Re:Russia’s Elbrus 8CB Microarchitecture: 8-Core VLIW on TSM...
by Symmetry on Tuesday June 02, 2020 @10:04AM
(#23391094)
VLIWs tend to do well for programs with predictable-at-compile-time memory access patterns which tends to correlate with floating point heavy code.
Re:Russia’s Elbrus 8CB Microarchitecture: 8-Core VLIW on TSM...
by brandmeyer on Tuesday June 02, 2020 @09:21AM
(#23390687)
Linpack is a floating-point benchmark focused on linear algebra operations. Coremark has some floating-point-heavy benchmarks in its suite, although a common criticism is that the footprint is too small to exercise the memory system.
So yeah, the "Maximum MFLOPS" row looks suspicious.