Changes between Version 16 and Version 17 of DataParallel/BenchmarkStatus
- Timestamp:
- 03/04/09 17:42:46 (4 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
DataParallel/BenchmarkStatus
v16 v17 41 41 || DotP, primitives || 100M elements || 937/937 || 934/934 || 474/474 || 238/238 || 120/120 || 65/65 || 38/38 || 28/28 || 42 42 || DotP, vectorised || 100M elements || 937/937 || 942/942 || 471/471 || 240/240 || 118/118 || 65/65 || 43/43 || 29/29 || 43 || DotP, ref Haskell || 100M elements || – || || || || || || ||||43 || DotP, ref Haskell || 100M elements || – || 934 || 467 || 238 || 117 || 61 || 65 || 36 || 44 44 || DotP, ref C || 100M elements || – || || || || || || || || || 45 45 || SMVM, primitives || ?? elems, density ?? || || || || || || || || || … … 50 50 ==== Observations regarding DotP ==== 51 51 52 The benchmark scales nicely up to the maximum number of hardware threads. Memory latency is largely covered by excess parallelism. 52 The benchmark scales nicely up to the maximum number of hardware threads. Memory latency is largely covered by excess parallelism. It is unclear why the Haskell reference implementation "ref Haskell" falls of at 32 and 64 threads. 53 53 54 54 ----
